UNIVERSITY OF MINNESOTA

This is to certify that I have examined this copy of a doctoral thesis by

Molly Harney Minkkinen

and found that it is complete and satisfactory in all respects, and that any and all revisions required by the final examining committee have been made.

Name of Faculty Adviser

Signature of Faculty Adviser

Date

GRADUATE SCHOOL

1

DOES THE CONCENTRATION OF CORTISOL IN MOTHER’S BREAST

MILK HAVE A RELATIONSHIP WITH THE BEHAVIOR OF THE INFANT

A THESIS

SUBMITTED TO THE FACULTY OF THE GRADUATE

SCHOOL OF THE UNIVERSITY OF MINNESOTA

BY

MOLLY HARNEY MINKKINEN

IN PATIAL FULLFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

FRANCES LAWRENZ, ADVISER

AUGUST, 2004

2

DEDICATION

To the mothers who so willingly shared such an intimate part of motherhood with me.

To the babies who shared their precious milk with me.

3

i

ACKNOWLEDGEMENT

I would like to thank the 22 mothers who volunteered to participate in this study. Without their donation of time, energy and breast milk this study would not have been possible. I would also like to thank the early childhood program administrators who helped me by identifying potential study participants.

I would like to thank Dr. Robert Lloyd for offering technical support and Dr. Michael Hartoonian and Dr. Harlan Hanson for serving on my committee.

I would like to thank Dr. Jean Stevenson for spending hours of her personal time editing my work and offering kind and supportive guidance.

I would like to thank Dr. Frances Lawrenz for her thoughtful guidance, sense of humor, and constant support throughout this project.

This study would not have without the support of my family. I would like to thank George and Marie Minkkinen for their care and support along the way. Because they were always there Mia was in loving hands while I was working on this project. Thank you to Brent and Kylee for their support. Kylee was always willing to read and re-read my work offering both technical and creative support to my writing.

A special thank you to my two daughters. To Marijah, who always offered a word of encouragement when I needed it most. To Mia, who had more patience than her years would have predicted.

Most of all I would like to thank my husband Phil. Your support never wavered, your selflessness was unending. But above all you believed in me. Thank you.

4

ii

ABSTRACT

Molly Harney Minkkinen 296 words

The purpose of this study was to determine whether the concentration of cortisol

in maternal breast milk had an impact on infant behavior beyond the influence of the

mother’s stress related behaviors. Nineteen mother infant breast feeding pairs were

recruited to participate in this study from two Midwestern states. Maternal stress levels,

cortisol production in breast milk, maternal socio-economic levels, and infant behavior

were assessed. Maternal stress was assessed using the State-Trait Anxiety Inventory,

cortisol in breast milk was analyzed using and ELIZA (enzyme immunoassay)

instrument, maternal SES was assessed with a tool designed by the researcher, and infant

behavior was assessed using the Rothbart Infant Behavior Questionarre-Revised. Data

were analyzed using Spearman Correlation Coefficients and linear regression analysis.

The Spearman Correlation Coefficient was used to determine the direction and degree of

linear relationship between variables. The linear regression was used to determine the

significance of the independent variables in relation to the dependent variable.

The results of the analysis of variables determined that the women who participated in

this study had middle to high family income levels and most had a college education.

The mothers also scored low on the State and Trait Anxiety test compared to the

normative sample which suggests that the women who participated in this study exhibited

lower than average levels of acute and chronic stress.

iii

5

The infants who participated in this study also tested differently than the normative

sample. The infants were generally happier than the normative sample. Cortisol levels in

maternal breast milk were measured but no comparison samples could be found due to

the lack of research in the area.

Spearman correlation analysis showed a significant relationship between State and

Trait Anxiety as well as a suggestion of a relationship between Trait Anxiety scores and

Negative Affectivity scores as well as between Positive Affectivity and Negative

Affectivity Rothbart scores.

Linear regression analysis results indicated that there was no significant relationship

between the dependent and independent variables, although the independent variables did

account for 16-26% of the variance in Rothbart scores.

iv

6

TABLE OF CONTENTS

Page

Chapter 1 INTRODUCTIONOverview . . . . . . . . 1Statement of the Problem . . . . . 2Rationale . . . . . . . . 2Purpose . . . . . . . . 14Research Question . . . . . . 15Overview of Dissertation . . . . . 15

Chapter 2 REVIEW OF LITERATUREIntroduction . . . . . . . 16Early Brain Development . . . . . 17The Role of Cortisol in Brain Development and Function . 33Breast Feeding: Biological and Social Factors . . 40The Study if Infant Behavior . . . . . 45Research Analysis . . . . . . 53Summary of Literature Review . . . . 56

Chapter 3 METHODOLOGYDescription of the Sample . . . . . 58Description of the Research Design . . . . 59Instrumentation . . . . . . 61Data Analysis . . . . . . . 66

Chapter 4 RESULTSIntroduction . . . . . . . 70Description of the Sample Characteristics . . . 70Data Analysis . . . . . . . 80Summary of Findings . . . . . . 87

Chapter 5 DISCUSSION AND IMPLICATIONSIntroduction . . . . . . . 88Summary and Highlights of the Findings . . . 88Correlation Results . . . . . . 90Discussion of the Findings . . . . . 91Possible Implications . . . . . . 97Limitations and Recommendations . . . . 98

Bibliography . . . . . . . . 101

v

7

APPENDICES

Appendix A Consent Form . . . . . .112

Appendix B Parent Information Sheet . . . .116

Appendix C State/Trait Anxiety Inventory . . . .119

Appendix D Rothbart Infant Behavior Questionnaire-Revised .122

Appendix E Socioeconomic Inventory . . . .132

vi

8

LIST OF TABLES

Chapter 3

3.1 Rothbart Behavior Descriptions

Chapter 4

4.1 Family Income4.2 Education Levels4.3a State Anxiety Scores Identified by Subject4.3b Trait Anxiety Scores Identified by Subject4.4 State Anxiety Score Distribution Compared to the Normative Sample4.5 Train Anxiety Score Distribution Compared to the Normative Sample4.6 Comparison of Means and Standard Deviations of Infants from the Current Study

with Infants from the Normative Sample4.7 Cortisol Identified per Week4.8 Cortisol Area Under the Curve Calculations4.9 Cortisol levels for Week One, Week Two, and Week Three4.10 Area Under the Curve Mean Cortisol Levels4.11 Cortisol Distribution Identified by Subject4.12 Spearman Coefficient Correlation Matrix4.13 Positive IBQ-R Regression Results4.14 Orientation/Regulation IBQ-R Regression Results4.15 Negative IBQ-R Regression Results

Chapter 5

5.1 Comparative Sample of Research Assumptions and Current Research Findings

vii

9

CHAPTER 1INTRODUCTION

Overview

The purpose of this study was to determine whether the concentration of cortisol

in maternal milk has a direct impact on infant behavior independent of the mother’s stress

levels. Maternal stress levels were evaluated to determine whether there was a

relationship between the stress levels of mothers and cortisol productions in breast milk

and as a control variable. Maternal stress was assessed using the State Trait Anxiety

Inventory (Consulting Psychologists Inc., 1977). The concentration levels of cortisol in

breast milk were also assessed. Individual cortisol levels fluctuate with the body’s

circadian rhythms, the cortisol concentration levels were taken three times on each of

three different days. In order to get one measurement per mother, the cortisol levels were

averaged. In order to determine the cortisol levels in the breast milk, basal cortisol levels

were measured using cortisol assay tests (Granger, 2004). Infant behavioral patterns were

assessed using the Rothbart Infant Behavioral Questionnaire-Revised (Rothbart, 2000).

The Rothbart was used to determine whether there is a relationship between the

concentration of cortisol in breast milk and infant behavior. The study findings have

provided a foundation for a better understanding of the relationship between the

concentration of cortisol in the mother’s breast milk and the impact it has on infant

behavior.

10

Statement of the problem

Cortisol has both positive and negative effects, and it is known that cortisol levels

can be affected by stress (LeDoux, 2002). It is known that infants with good attachment

or with good care experience less stress in terms of elevated cortisol (Gunnar & Barr,

1998). It is known that chronic stress can lead to elevated levels of cortisol which in turn

may lead to the disregulation of the LHPA system (LeDoux, 2002). It is known that over

time the disregulation of the limbic-hypothalamic-pituitary adrenocortical (LHPA) axis.

LHPA system disregulation can lead to neuronal damage (LeDoux, 2002). It is known

that cortisol levels in mothers affect maternal response (Flemming et al. 1997). It is

known that cortisol is found in breast milk ( Flemming et al.1997). It is known that the

mothers’ levels of stress affect the amount of cortisol in their milk (Flemming et al.1997).

It is not known whether infant behavior is affected by the level of cortisol in the mothers’

breast milk. Because there is a possibility that these factors may play a part in infant

behavior, it is important to find out if there is a relationship between cortisol in breast

milk and infant behavior.

Rationale

According to Sarit Avishai-Eliner and her colleagues, Kristen L. Brunson, Curt

Sandman and Tallie Baram (2002), the ability of an organism to adapt to its environment

is integral to its survival. Eliner-Avishai et al. (2002) point out that daily life involves

confrontations with changing situations that can be physiologically or psychologically

challenging. “In order for an organism to cope with these actual or perceived threats of

11

stress, the ability to alter function and expression of neuronal genes has been developed

in the form of molecular and behavioral stress responses” (p. 375).

What this means according to Eliner-Avishia et al. (2002) is that when an

organism senses stress, the body initiates a rapid secretion of affective molecules

(noradrenalin and adrenal glucocorticoids). One type of molecule secreted due to

perceived or actual stress is an adrenal glucocorticoid, cortisol in the case of humans.

This hormone is known to form the first line of defense for mammals (including humans)

under conditions that threaten homeostasis (Gunnar & Barr, 1998). The regulation of

cortisol in the body and brain works by using a feedback loop that is dependent on an

interrelationship of the amygdala, hypothalamus, pituitary glands, adrenal glands, and the

hippocampus in the brain referred to as the LHPA axis. The feedback loop is described

by LeDoux (2002) as a process that starts with the onset of perceived or real stress. At

this point the amygdala initiates the stress response by activating the paraventricular

nucleus of the hypothalamus (PVN). The PVN releases corticotropin releasing factor

(CRF) into the pituitary gland (PIT). ACTH is then released by the PIT into the

bloodstream which induces the release of cortisol from the adrenal cortex. Cortisol is

then carried back to the brain via the blood, where it stimulates receptors in the amygdala

and hippocampus. The hippocampus receptors mediate negative feedback control on the

release of cortisol whereas those in the amygdala mediate positive feedback, which in

turn inhibit or excite (respectively) in the PVN. The emotional and physiological

response to stress results from the balance between those two feedback circuits. Cortisol

has been identified as a significant contributor to the impairment of the hippocampi and

at the same time facilitates amygdala function (LeDoux, 2002). While the onset of stress

12

initiates this system, chronic stress contributes to a disorganization of the system due to

overuse. This disorganization leads to an LHPA system that can not regulate cortisol

effectively and in turn does not facilitate increases in cortisol when the system is under

stress or provides a lack of the appropriate levels of cortisol for stressful situations. The

LHPA system simply underfunctions. On the other hand, chronic stress can result in

prolonged stimulation of hippocampal cortisol receptors which then become subsensitive,

analogous to chronic heroin abusers’ need for higher doses of heroin to get high. Drug

tolerance is the result of chronic administration.

Consequently, at times of acute stress, the hippocampal negative feedback system

is under-responsive and a runaway, uncontrolled psychological and physiological stress

response can occur. Chronic stress resulting in the disregulation of the feedback

mechanism can sometimes present depressed resting levels of cortisol.

In an article entitled “Programming Effects of Neonatal Dexamethasone

Treatment on the Ventral Tegmental Area in Adulthood: The Potential Role of Annexin

1” by S. McArthur et al. (2003) the authors suggested that there may be a relationship

between inappropriate exposure to glucocorticoids at critical developmental stages that

could have long-term effects on the central nervous system function and may increase the

risk of developing central nervous system disorders later in life.

These projections were the outcome of their research conducted using rat dams and

their pups. The dams were given drinking water with 1 mg/ml of a synthetic

glucocordicoid called demamethasone (DEX ) for the first five days after giving birth.

The results of the research indicated that there was an effect of the CG manipulation on

the tyrosine hydroxylase immunopositive (TH+) cells. They found that there was a 50%

13

decrease in the TH+ cells in the lower midbrain also known as the ventral tegmental area

(VTA) of the brain in the adult rats who had nursed from the dams who had been given

the DEX treatment. “This decrease in TH+ cells indicated that the survival or expression

of VTA dopaminergic neurons is profoundly influenced by brief postnatal GC

exposure” (S. McAthur et al. 2003 Program Abstract, Society for Neuroscience

November 2003.) they also suggest that “their findings may have implications for the

development of major psychiatric disorders such as schizophrenia, the incidence of which

is increased by perinatal stressful events” (p.7). These findings suggest that dopamine is

involved in schizophrenia, reward, and the formulation of saliency.

It is clear through the research of Eliner-Avishia et al (2002), Gunnar and Barr

(1998), LeDoux (2002) and McArthur et al (2003) and that the cortisol that is

manufactured in the adrenal glands is helpful for the facilitation of overall bodily

function. Although cortisol is known as a stress related hormone that is responsible for

negative brain function like lack of memory, spatial reasoning and decision making, it

has also been identified as important to the overall function, among those functions are

the regulation of blood pressure and cardiovascular function and the regulation of the

body's use of proteins, carbohydrates, fats, as well as the facilitation of emotionally laden

processes like bonding and attachment. These relationships are found when cortisol is

within normal limits and when elevations in cortisol are non- persistent.

Another view of the role cortisol plays in human function was identified in a

study that looked at the role cortisol plays in the development of attachment relationships.

Fleming et al. (1999) conducted a study with mothers and their infants to determine the

significance of cortisol in maternal attentiveness. Hendonic testing assesses smell

14

sensitivity and salivary cortisol concentration. The authors of the research used

hendonics testing with samples from primiparous mothers, nonmothers, and multiparous

mothers. Their findings indicate that there is a significant correlation between higher

levels of cortisol in primiparous mothers and their ability to recognize the odor of their

infant. They also stated that when primiparous mothers were looked at alone there was a

relationship found between the amount of time spent affectionately touching the infant

and the level of cortisol found in the mother’s saliva. When they looked at multiparous

mothers there was again a correlation found between maternal behavior and the level of

cortisol found in the saliva. The increased behavior in multiparous mothers was found in

general caretaking activities and not with affection behaviors. In both cases there was a

correlation found between maternal cortisol levels and maternal behavior. In both cases

mothers with higher cortisol levels were found to be more attentive in general (Fleming et

al. 1999). In addition, Fleming et al. suggested that there may be a correlation between

breast feeding and maternal bonding. They hypothesized that milk facilitates the bonding

and attachment process between mothers and infants. To prove their hypothesis their

study focused on the chemical reactivity present during lactation which was established

using a chemical measurement instrument. One significant finding that emerged from

their assay results indicated that both the mothers’ and their offspring showed enhanced

neurochemicals (cortisol, oxytocin and endogenous opioids) during the lactation period.

These neurochemicals are thought to be instrumental in conforming the responsiveness of

the mother and infant to each other.

Knowing that maternal behavior and infant responsiveness are the precursor to

later relationship structures provides us with a biological view of bonding and

15

attachment. The findings also suggest that behaviors that facilitate the structure of the

mother-infant relationship are correlated to the chemical production in the maternal and

infantile biological systems (Fleming et al., 1999). They also suggest that these findings

indicate that “early experiences and associated brain changes clearly affect how infants

interact with the world in general, and with their own offspring as adults and the next

challenge is establishing exactly how the neurobiological changes produced by early

experiences might persistently alter the neurobiological mechanisms” (p. 673).

Other studies that investigated the role of breast feeding and brain development

were conducted by researchers such as Mezzacappa , E.S., Tu, A.Y. & Myers, M.M.

(2002); Rolf & Keil (2002); Nelson & Panksepp (1998); Fleming A.S., Steiner, M.,

Cortner, C. (1997), and Cook (1997). Their studies have extensivly examined the

sociological, nutritional, and biological impacts of breast feeding on human and animal

offspring. These studies have found that cortisol exists in breast milk and is presumed to

be passed to the infant during feedings.

A research team led by Mezzacappa (2002) uses both animals and humans to

better understand human brain development. The findings of researchers such as Nelson

& Pankseep (1998), Fleming et al. (1999) and Mezzacappa et al. (2002) have shed light

on the issue of breast feeding as an environmental construct in human infants. In their

research, which was conducted for the purpose of exploring the significant constructs of

breastfeeding, using both rats and humans. The rat studies were conducted using

immobilization and a maze as stress initiators. The human studies were conducted using

observation and interviews. The lactating mother rats were found to have less cortisol in

their systems and demonstrated less stress reaction than their non-lactating counterparts.

16

These data suggest that breast feeding reduces maternal subjective and physiological

responses to stress and negative moods in rats (Mezzacappa et al., 2002). In analyzing

findings from human subjects, data related to cortisol levels and heart rates from the

breast fed infants were compared to the bottle fed infants. The findings from this

comparison indicated that breast fed infants had lower baseline electrodermal

conductance and less heart rate reactivity to stimuli in a resting state (Mezzacappa et al.,

2002). The same study also compared breast and bottle feeding in relation to cortisol

levels and found that breast fed infants had lower cortisol levels when introduced to an

exercise experience.

Although it is clear that there is a relationship between cortisol levels and

behavior, the question of how cortisol levels in breast milk affects the behavior of breast

feeding infants is still unclear.

The increased cortisol in the infants with the disorganized attachment relationship

is thought to be the result of a disorganized LHPA system due to a history of

unpredictable maternal behaviors which in some cases can lead to chronic stress in the

infant. The issue of early experience was studied in 1996 by Gunnar M.R., Brodersen, L.,

Nachmias, M. Buss, K. and Rigatuso, J. Their study was conducted using seventy-three

18 month old infants who were inoculated during a routine well baby examination. The

infants were studied to determine whether there was a correlation between infant

attachment behaviors, maternal behavior, and salivary cortisol levels. Infant attachment

behavior was measured using the Ainsworth Strange Situation Assessment tool, the

Toddler Behavior Assessment Scale, and the Social Fear Scale. Maternal behavior was

assessed during the infant health exam using four indicators: Responsive, ignoring,

17

intrusive, and neutral. Responsive was identified as the demonstration of response to

infant signals e.g., soothing and adjusting the physical position. Ignoring identified as

behavior that failed to respond. Intrusive behavior was identified as continued stimulation

of an infant that was turning away. Neutral behavior was identified as no interaction but

aware of signal and needs of the infant. The quality of the mother infant relationship was

determined by observing the behavioral measures (Gunnar et al. 1996). Infant salivary

cortisol was measured using a variation of the Amersham Corporation Amerlex Cortisol

RIA kit and the Pantex Corporation Assay kit. The study findings suggested that

emotional security of the attachment relationship moderates relations between inhibited,

fearful behavior, and the neuroedocrine stress reaction (Gunnar et al., 1996) which

suggests that mothers who were more attached to their babies had babies with lower

cortisol levels. The findings also support the notion that secure attachment to mothers is

related to a history of responsive care and that maternal responsiveness may be related to

the degree of coherence in infant stress reaction early in life and that there is thought to

be a relationship between maternal responsiveness and infant stress organization and the

development of secure mother-infant relationships (Gunnar et al. 1996). These research

findings were the result of smaller increases in cortisol elevation in infants who were in

secure attached relationships with their mothers, and higher elevations of cortisol in

infants who were identified as being in a disorganized attachment relationship with their

mothers when the infants were in a stressful situation. The findings from this study

suggest that there is a relationship between maternal behavior and infant cortisol levels.

Early experiences, cortisol and behavior were also discussed in relation to

childhood stress, infant attachment, and quality care in a 2002 paper written Gunnar and

18

Donzella. Their paper states “that adverse experiences early in life predispose

individuals to affective pathology in part through affecting the LHPA system (Gunnar &

Donzella, p.200) This statement was made explicit through the discussion of research

that included studies with children who ranged in age from birth to approximately five

years of age. The studies included the use of salivary assay measures and a variety of

techniques that ranged from observing infants being cared for by their mothers during a

routine inoculation, eliciting a fear response in toddlers through separation from their

mothers, eliciting a fear response in preschool age children by using clowns, and

assessing the quality of care children experience. By using baseline cortisol levels to

compare to cortisol levels after treatment, the findings indicate that there is a relationship

between the quality of care an infant experiences, behaviors that are indicative of

negative emotional responses, and cortisol increases. In other words, infants who

demonstrated behaviors which reflected distress, such as crying, proximity seeking, and

withdrawal and are reported to have negative emotionality on temperament inventories

tend to have elevations in infant cortisol, and these were found to occur when the quality

of care was poor or when the attachment relationship between the parent and the child

was insecure (Gunnar & Donzella, 2002).

Although incomplete, their findings are consistent with the extensive work that

has been done with rodents and other non-human mammals conducted by researchers like

Gunnar & Donzella (2002). Gunnar and Donzella’s findings suggest that like rodents and

non-human mammals, developing humans have sensitivity to social regulation. Cortisol

levels become increasingly difficult to increase as the age of the infant increases under

conditions that are thought to be sensitive and responsive. It is suggested by Gunnar and

19

Donzella (2002) that these research results indicate that children learn to expect that their

attachment behaviors will be responded to by their caregivers. Children who are in

relationships that are not as responsive and nurturing show increases in cortisol levels

when in situations of maternal separation and when in situations with strangers. Increases

in cortisol were also found in children who were identified as temperamentally vulnerable

and those who were described as fearful and anxious (Gunnar & Donzella, 2002).

Long term effects of attachment quality have also been identified as having long

term implications by researchers like Ainsworth and Bowlby (1991). They suggest that

(Gray, 1999, pg. 452) “infants who were judged as securely attached in the strange –

situation test (Ainsworth Strange Situation Test) (Ainsworth 1991) have been found to be

on average more confident, better problem solvers, emotionally healthier, and more

sociable in later childhood than those who were judged to be insecurely attached”

(Ainsworth, 1991, Frankel & Bates, 1990; Jacobson & Wille, 1986). R. J. Paterson and

G. Moran (1988) have taken a similar look at maternal social and cognitive

characteristics. Their research suggests that mothers who were identified as having

secure attachment relationships with their infants were themselves identified as

emotionally secure, likable, and felt competent as mothers. Paterson and Moran (1988)

cite the work of Egeland and Farber from a 1984 study that found that mothers of

securely attached infants were found to be less aggressive than mothers of insecure

infants (Paterson & Moran, 1988). Benn (1986) found that there was a significant

positive association between maternal emotional functioning and security of attachment.

The findings suggested that better functioning mothers have a higher proportion of

securely attached infants (Paterson & Moran, 1988). A more recent study by

20

Mangelsdorf , S.C., McHale, J.L., Diener, M., Goldstein, L.H., & Lehn (2000) confirms

the past theory by studying 102 mothers infant pairs at eight months and again at 12

months to determine whether maternal and infant characteristics were predictors of

attachment. Using the Rothbart Infant Behavior Inventory (Rothbart 2000) to study the

infants and the Tellegen’s Multidimensional Personality Questionnaire to study the

mothers, as well as video taped observations of the mother infant interactions the research

team found that at both eight and 12 months there was a relationship between infant

behavior and maternal behavior. An example of this was that they found that infants who

were classified as having avoidant attachments were lower on positive affect and higher

on fearfulness and had mothers with lower positive affectivity. Mangelsdorf et al. (2000)

suggest that these findings point to the importance of examining both parent and child

characteristics in the prediction of attachment (Mangelsdorf, et al., 2000). Essex, M.J.,

Klein, M.H., Cho, E., Kalin, N.H. (2002) discussed this issue further by providing

findings from their longitudinal research on maternal stress and infant reactivity. They

found that maternal stress interferes with mother-infant interactions and that mothers who

were stressed were often found to lack sensitivity and responsiveness. They found that

infants of stressed mothers showed higher levels of cortisol (Essex et al. 2002) Through

these findings the research team concluded that “ it is likely that our sample of altered

mother-infant interactions secondary to mothers’ high levels of stress mediated the

effects on children’s later cortisol and behavior” (p. 38). Because attachment is a

predictor of infant temperament and maternal behavior is related to stress, it is important

to look closely at maternal stress levels related to maternal characteristics and infant

21

temperament related to maternal behavior as identified by researchers like Paterson &

Moran (1988), Essex (2002) and Mangelsdorf (2000).

Another factor to consider when identifying issues related to infant cortisol levels

is the socio-economic status of the infant’s family. In a study to determine whether there

is a relationship between cortisol levels in children and the socio-economic status of the

family. Lupien S.J., King, S., Meaney, M.J., McEwen, B.S.(2002) studied 217 children

from a variety of income levels. The SES was indexed by family income and by the

father’s and mother’s level of education. Their research findings suggest that there is a

significant correlation between socio-economic status and infant cortisol levels. They

found that infants from low SES families have higher cortisol levels than children from

high SES families (Lupien et al., 2002). The Carnegie Foundation takes this issue a step

further in identifying the significance between SES and maternal stress. They suggest

that there is a direct correlation between maternal stress levels, SES and, maternal care

qualities (1994).

The next challenge in understanding the neurobiological mechanism related to

infant and maternal behavior will be met by determining the relationships among

maternal cortisol levels, maternal stress levels, concentration of cortisol produced in

breast milk, the infant’s cortisol level and the infant’s behavioral responsivity. While the

research of Fleming et al., (1997) has identified that there is a relationship between

maternal attentiveness and an increase in cortisol levels in mothers and Gunnar et al.,

(2001) found that there is a relationship between cortisol levels in mothers and the quality

of early care an infant experiences and Mezzacappa et al, (2002). identified a

relationship between breast fed infants and lower baseline cortisol levels; there were still

22

unanswered questions about how maternal stress affects infant behavior and the causal

relationships among cortisol, breast milk, maternal stress, and infant behavior.

In summary, chronic elevations of cortisol have been found to lead to emotional

disorganization. Therefore elevated levels of cortisol in breast milk could lead to

emotional disorganization in the developing nervous system of an infant which may have

long term effects for social and emotional competence in adulthood. Hence I measured

cortisol in breast milk to see if had an independent (direct) relationship with the infant’s

behavior.

Purpose

The purpose of this study was to determine whether the concentration of cortisol

in maternal milk has a direct impact on infant behavior independent of the mother’s stress

related behaviors. Maternal stress levels were evaluated to determine whether there is a

relationship between the stress levels of mothers and cortisol productions in breast milk.

Maternal stress was assessed using the State Trait Anxiety Inventory (Consulting

Psychologists, 1977). The concentration levels of cortisol in breast milk were assessed.

Due to the fact that individual cortisol levels fluctuate with the body’s circadian rhythms

the cortisol concentration levels were taken three times on each of three different days.

The cortisol levels were averaged. Infant behavioral patterns were assessed using the

Rothbart Infant Behavioral Questionnaire-R (Rothbart, 2000). The Rothbart Infant

Behavioral Questionnaire was used in order to determine whether there is a relationship

between the concentration of cortisol in breast milk and infant behavior. The basal

cortisol levels were measured using cortisol assay tests in the mother’s breast milk. The

23

study findings provided a foundation for a better understanding of the relationship

between the concentration of cortisol in the mother’s breast milk and the impact it has on

infant behavior.

Research question

Does the concentration of cortisol in mother’s breast milk have a relationship with the behavior of the infant?

Hypothesis: Under conditions of normal, not chronic stress, high maternal stress levels lead to high cortisol concentrations in breast milk. This leads to a disruption of the emotional response circuitry of the infant which results in excessive emotional responsiveness of the infant leading to negative infant temperament characteristics.

Overview of Dissertation

A review of the literature related to this study is provided in Chapter 2. Included

in the review is the research of experts in the fields of cognitive science, child

development, and neurobiology.

A description of the research methods used in this study is provided in Chapter 3.

Included in this chapter is a description of the measurement instruments, the methods of

research inquiry, and the research process.

The results of the study are presented in Chapter 4. Included are the results of the

mother and infant questionnaires, results of the cortisol assay tests, the statistical analyses

of the collected data, and the results of the overall study findings.

Finally, Chapter 5 provides a discussion of the results. Included in this chapter is

a discussion of the findings related to cortisol concentrations, infant behavior, and

maternal stress. The limitations of the study will be discussed as well as implications and

further research.

24

CHAPTER 2REVIEW OF THE LITERATURE

Introduction

This study is based upon a body of research on infant development and biological

reactivity of infants and mothers. This literature includes research on early brain

development, the role cortisol plays in brain development and function, the biological and

social factors of breast feeding, and the study of infant behavior. In order to provide a

context for the present study, studies from this body of research were reviewed.

Research in the area of infant development has endless facets of investigation. No

one facet of research is more important than another, and all of the findings work in a

synchronous fashion to facilitate our understanding of child development.

Research on child development has proliferated across the centuries. Infant

characteristics like attention, digestion, and behavior are among the areas studied. The

research findings in these and many other areas have deepened our understanding of how

infants’ respond to their environments, and to the people who care for them. The findings

from past research have also led to changes in strategies for interacting with infants in the

medical, educational and parenting realms.

One area of research interest in the past decade has focused on the issue of infant

brain development. In most cases the goal of the research has been to attempt to answer

long standing questions that past scientists and researchers were prevented from studying,

because of a lack of technological advancement and expertise. The questions focused on

the biological function of cognition, how chemical reactivity influences brain function,

the biological origin of temperament, and the biological impact of breast feeding on early

25

brain development. Advancements in science and technology along with the work of past

researchers have led to a time when many hard to answer questions about infant brain

development are ripe for investigation. Now is the time when researchers can investigate

the role breast feeding has on the biological function of the brain and the subsequent

temperament of the infant.

The purpose of this chapter is to create a clear picture of the work that has been

done by previous researchers in the areas of infant development and neurobiology by

reviewing literature in these areas. The review of literature is intended to provide an

overview of past and current research findings which will lead to an understanding of

current theories of infant development in the area of neurobiology as well as uncover the

relationship between brain development and breast feeding, and look at how temperament

plays a part in the overall development of the infant brain. This understanding is

intended to set forth a direction for further research.

In order to create a comprehensive picture of these issues this paper will review

literature in four separate parts and conclude with syntheses of the findings. Part one will

be a review of literature related to early brain development. Part two will review the

chemical substrates of the brain. Part three will review theories of temperament. The final

section will review the role breast feeding plays in a child’s development.

Early Brain Development

Allen Schore (2001) states that the greatest amount of neurobiological research is

on the adult rather than the developing brain, and most of it is not on normal but on

abnormal brain function. Even so, contemporary neuroscience is now beginning to

26

become interested in the early developmental failures of the brain. And so neurobiology

is currently exploring early beginnings and brain pathology (Schore, 2001)

Pediatric neurologist, Harry Chugani who studies early brain development notes

“there is no doubt that experience molds the young brain. The early years determine how

the brain turns out” (Schore, p.1, 2001). He goes on to make clear that the child’s

potential is determined in the early years—from the first moment of life to countless

hours spent in day care. He says “these are the years when we create the promise of a

child’s future. This is when we set the mold”. Many scientists have said that

kindergarten is not the starting point of the child’s brain development. They say that by

kindergarten the process is half over (Schore, 2001).

Although the work of Schore and Chugani and others make clear the obvious

movement today in the sciences to include early brain development, there is a well paved

road to the current practice of child care and neurobiology that is rich in history.

In an article titled “Neonatal Neurology, Past Present and Future: A window on

the brain”, the author, Lily Dubowitz and her colleagues describe the historical journey of

the study of infant brain development. Dubowitz et al. (1995) state that the first

examination of the brain had taken place in the 1940’s and 50’s. The study of the brain at

the time was aimed at evaluating the neurological state of the full term infant. The first

examination was developed by Andre'-Thomas and Saint-Anne Dargassies. The

examination mainly involved the evaluation of passive and active tones and defined in

detail various primitive reflexes. The findings of the examinations mapped out the

maturation of the neurological features in premature infants of various gestational ages

27

and were able to show that the development characterized related to the maturity and not

the size of the infant.

Dubowitz et al. (1995) write further about the history of infant brain research by

citing the work of Prechtl of Holland. Prechtl’s research followed that of Thomas and

Dargassies. In the 1970’s Prechtl developed an examination which concentrated on the

neurological responses of the full-term infant. The aim of the research was to take into

account the behavioral state of the newborn. The findings were able to demonstrate that

many of the neurological signs in a newborn were very much dependent on the state of

alertness at the time the newborn was evaluated.

Dubowitz et al. (1995) continue their review of historical research in the area of

infant brain development with the work of T. Berry Brazelton in the early 1970’s.

Brazelton developed a behavioral examination which was based to a large extent on an

earlier examination developed by K.F. Graham in the 1950’s. The items included tests

that measured motor activity and strength, responsiveness to auditory stimuli as well as

measured irritability and tension in infants. It also aimed to evaluate the interaction

between the newborn infant and its caretakers (Dubowitz et al., 1995).

Dubowitz et al. (1995) indicated that their research findings suggest that there

were problems with the neurological examinations of the past, because the examinations

were generally poor in predicting outcomes. The researchers in Dubowitz’s team also

note that “while it was possible to relatively easily identify infants with normal outcomes

as those who were persistently normal in the neonatal period, there were large number of

infants who showed abnormalities in the newborn period, yet had completely normal

outcomes later on”(pg 23).

28

The work of Thomas &Dargassies (1952), Prechtl (1977), Brazelton (1972), and

others were successful in creating an understanding of the human brain function, which

has in-turn, provided insight into early brain development. The synthesis of their work

has identified behavioral benchmarks in brain development that show stage and phase

related consistencies among developing infants. The measures, although empirical, left

questions that were technologically unanswerable.

The advancements that have occurred recently in science and technology have

created measures that have the capability of looking at the biological structure of the

brain such as magnetic imaging and magnification. These technological advances have

been instrumental in identifying the architectural structures of the developing brain. The

information gathered from these measures has added depth to our understanding of how

young brains organize, function, and development neurologically. Dubowitz et al. (1995)

describe that data that was instrumental in this theory development in their article entitled

“Neonatal Neurological, Past, Present and Future A window on the Brain”. The article

described the data collected by Dubowitz which focused on live infants. The use of

technology such as the MRI and the cranial ultrasonography was instrumental in their

findings. The Dubowitz et al. study of live infants examined the incidence, timing and

evolution of the lesions in the neonatal period. The researchers were able to follow the

evolution of the lesions into infancy. The findings from this work is significant in the

development of an understanding of development of brain lesions in live infants.

Because these technologies are available and in the case of the ultrasonography, portable,

this understanding is typically referred to as the maturational theory of brain

development. Discussions of this theory are significant to the field of early childhood

29

education, child development and other related fields because they suggest that early

neurological brain function is essentially the creation of cognitive constructs that become

the function of sight, language, emotional response, and creativity.

Within the confines of the maturational theory the architectural structure of the

developing brain has recently been identified. The theory has been developed with the

use of new technologies such as brain imaging machines. Three of the most frequently

used machines are the functional magnetic resonance imaging (fMRI), positron emission

tomography (PET), and magnetic resonance imaging (MRI) machines. Tests using the

machines have shown that the early brain architecture is made up of 100 billion neurons.

Each neuron has a nucleus, an axon and many branch-like dendrites. Each neuron has the

potential to make a synaptic connection with approximately 60,000 to 100,000 other

neurons (Pally, 1997). Researchers Gunnar and Barr (1998) describe the development of

the architectural process as the biological growth and differentiation of the developing

brain. They suggest that the development and migration of the neurons in the developing

brain consist of a series of biological events that proceed in a relatively well-ordered but

overlapping sequence. This sequence consists in the birth and migration of cells, the

growth of the axon, and the formation of the dendrites that allow connections among

neurons, the formation of the synapse that allow the information flow across neurons and

then the mylenation that speeds up the transmission of neural impulses. The repeated

process of this biological sequence is the precursor to the circuit development of the brain

that will be discussed later.

Gunnar and Barr state (1998) that axon development is largely complete by birth

in full-term infants and dendritic growth, synaptic function, and mylenation are largely

30

postnatal events. Most of the brain growth in these areas has been found to take place

during the first two years of the infant’s life.

In keeping with the maturational theory Diamond, (1988) suggests that the brain

has a use it or lose it structure. She states that we are born with an overabundance of

neurons and dendrites, and that during the prenatal period there is a pruning process that

begins. This pruning is the result of the experience dependent circuit development of the

neural pathways. The neurons that are used remain, while those that are not die off

(Pally, 1997). The neural circuit that Diamond describes is the foundation to the

constructs that substantiate learning throughout the infant’s life. Gunnar et al. (1998)

indicates that findings from animal research show that early experiences program the

brain’s neurological circuits in ways that affect later cognitive competence, emotional

responding, and activities of physiological systems that orchestrate reactions to stress and

challenge.

The architectures that Gunnar &Barr (1998), Diamond (1988), and others describe

are the microscopic underpinning of larger structures in the brain. These structures are

thought to have a great evolutionary history. According to the MacLean model or the

Triune model MacLean , developed by P.D. MacLean in the early 1950’s there are three

distinct yet interrelated parts of the brain that have developed systematically from our

ancestors. He suggests that there is a reptilian brain, a lower mammalian brain, and a

primate brain.(Pally,1997). This model of brain evolution developed in the 1950’s and

was popularized in the 1970’s and 1980’s. It is currently viewed by most researchers as

an outdated model due to the simplistic categorization of brain function. An

understanding of the actual dynamic function of the brain has changed as a result of the

31

research from the recent decade. However, the categories MacLean identified are still

thought to be accurate. Because of this, the model is still used as a basic structural view

of the brain. The part of the brain that McLean identified as the reptilian brain is

currently known as the brain stem. It is responsible for the vital functions of

physiological survival such as sleep/wake cycles, heart rate, respiration, and body

temperature. MacLean’s lower mammalian brain is currently called the limbic or

subcortical region of the brain and is thought to have primary control of the emotional

function of the brain. This region of the brain is responsible for such things as parental

care; including nursing, emotion, memory, play and the infant distress cry (Pally, 1997).

What MacLean called the primate brain is currently referred to as the cortex. This area of

the brain is thought to control the higher order thinking function. This region of the brain

is significant to infant brain development in that it modulates the emotion, behavior, and

body physiology processed by the lower brain (Pally, 1997). According to MacLean,

through natural selection “newer” brain structures, which could perform more adaptive

functions, were added on to and integrated with more primitive structures (Pally, 1997).

As mentioned, researchers suggest that the model presents an over simplified

view of the brain. The brain is in fact a very dynamic structure that is whole in its

function. Despite the current view, I use MacLean’s Triune or modified Triune model in

this work for two reasons. First, the model provides a simple structural view of the

human brain. Secondly, the structures that MacLean identified are not inaccurate. They

do exist in the manner in which he describes them. However they are found to be more

complex in their function than MacLean originally thought. I believe that for the purpose

of structural identification this model is useful.

32

Using the research findings of MacLean and others, Pally (1997) has conducted

extensive research in the area of early brain development that has led to the development

of a theory that suggests that the brain's design strikes a balance between circuit’s

permanence and circuit’s plasticity. Circuit permanence is the result of repeated

exposure to a given stimuli. Many researchers refer to this phenomenon as “hardwiring”.

As a child has repeated exposure to sight, sound, touch and other stimuli the neurons

responsible for securing an understanding of the situation create a permanent pathway

related to that given information. An example of this for older children would be riding

a bicycle or playing a Mozart concerto (Pally, 1997). As the behavior is repeated the

neural circuit’s hardwire. The hardwired motor pattern is stored in the basal ganglia in

the brain stem and is activated as automatic motor routines (Pally, 1997). Infants have

the same developmental process with experiences like vision and parental response. As

an infant has repeated visual information the neural pathways in the occipital region of

the brain (primarily responsible for vision) hardwire the pathways that in-turn provide

visual input. This hardwired neural pathway responds automatically creating visual

capabilities for the infant. Repeated parental response will also result in a hardwiring

process in the limbic region of the brain (primarily responsible for emotional reactivity).

An infant will respond to parental behavior in a predictable manner based on repeated

past responses by the parent. In other words the infant hardwires his/her understanding

of the care expected through repeated behavior on the part of the parent. If an infant is

picked up, soothed and responded to appropriately, the brain will make neural

connections that become “hardwired” and in-turn the brain will expect the same response

in the future. The phenomenon of neural plasticity is quite the opposite of “hardwiring”.

33

Neural plasticity describes the flexibility of the neural circuitry to adapt and adjust to

stimuli. According Diamond (Diamond & Hopson, 1988) neural plasticity is essentially

neural flexibility. Diamond (1988) states that neural plasticity has the same functional

resilience, and pliancy of living silly putty. Diamond goes on to say that “more than any

other organ, the brain can be shaped by stimulation and use, by disease, and trauma, by

dull routine and disuse into a center of thought, sensation, and regulation most

appropriate for a given individual’s life” (p. 58). Healthy brain development has the

presences of both neural hardwiring and neural plasticity. Pally states (1997) that more

plasticity exists in cortical circuits where new dendrites can grow and synaptic

connections can continue to be made throughout life. An example would be that cortical

functions such as vocabulary and math have a lot of plasticity. People can continue to

learn a number of words for the same concept and a number of solutions to a problem.

On the other hand, the subcortical limbic "emotional” circuits that develop in infancy

have less plasticity and therefore may have a long-lasting effect on subsequent

psychological development. It seems that these are the circuits that, through repetitive

performance create permanence that allows children to form and maintain attachments to

their parents over the long period of their development and to seek familiar reliable

sources of safety and comfort. As mentioned earlier, through maturation the cortex

develops the capacity to modulate emotional responses of the subcortex a process that is

easily observable as children and adolescents grow (Pally, 1997).

Pally (1997) also points out that the human brain is born prematurely, and that

because of this much of the brain’s development occurs postnatally and for many years

forward. This development includes both neural hardwiring and plasticity of the neural

34

circuits. The significance of the neural process of hardwiring and plasticity are

compounded when considering neural pruning. Diamond (1988) describes the

phenomenon of neural pruning in Pally’s work related to early brain development.

Diamond describes the architecture of the developing brain as having a use it or loses it

process. Pally quotes Diamond (1988) as saying that brain development is a matter of

“use it or lose it”. Diamond continues on to say that we are born with an overabundance

of neurons and dendrites and during the neonatal period a pruning process begins. As a

result of experience-dependent circuit development, the neural paths that are used remain,

while those that are not used die off (Pally, 1997). Gunnar (1998) adds to this theory by

suggesting that many aspects of brain development are activity dependent (use

dependent), and thus incorporate experience as part of the basic program for growing a

brain. Diamond says that experiences early in life do appear to influence how the child

reacts to later experiences and how the brain processes and profits (or fails to profit) from

experiences throughout life.

Johnson (1995) brings yet another perspective. He also supports the maturational

theory, but suggests that brain research should focus on the notion that brain development

has a progression nature of maturation from posterior to anterior regions. He cites

research findings that indicate that the primary visual cortex is functional shortly after

birth and the frontal lobes and prefrontal cortex being the last part to become functional

as supportive evidence for his theory. Johnson refers to his theory as the cognitive

developmental theory and suggests that in the first few weeks after birth many brain

pathways are partially activated by task situations in infants, but most of the cortical

pathways are under specialized. He further suggests that the actions of young infants are

35

often based on the output from the first available pathway which may often result in an

impoverished or incorrect response when the task situation is more complex or

demanding (Johnson, 2000). The theory proposed by Johnson has implications for views

on cognitive and behavioral development in infancy. He believes that development is not

stage-like in nature and suggests that brain development is dependant on multiple

competing representations that are engaged by particular stimuli or tasks demands, and

which compete to influence behavioral output. Johnson’s cognitive developmental

theory is new and based on the data generated by the author. Johnson (2000) suggests

that during the next decade great advances will be made as he comes closer to

incorporating evidence from neuroscience into our understanding of mental development

in infants and children.

Neurobiologist Michael Posner (2001) looks at infant brain development from a

different perspective. He combines the maturational and cognitive developmental

theories as he views brain development with a primary interest in relating the human

brain to the cognitive and emotional development of infancy and childhood. Posner has

focused his research on the behavioral output of the brain and on the attentional networks

in the infant brain. In doing so Posner feels that he is building upon new knowledge

about the brain that has developed from the work of maturational researchers as well as

recent research in the areas of cognitive development and genetic influences. Posner

traced the orientating skills of infancy during the first year of life and found that at four

months, infants can learn where to look. The maturational theory suggests that the

infants have hardwired their ability to perform visual tracking through repeated visual

input. The cognitive theorist would suggest that the infant brain was physically

36

developed enough at four months to perform the task with or without previous

experience. Posner (2001) also found that the four month old infants did not have the

ability to handle ambiguous context and using explicit knowledge to control their

behavior development. Gunnar (1996), Diamond (1988), and Pally (1997) suggest that

through repeated experience and appropriate social interactions the behavioral

development would emerge. Johnson (2000) suggests that in time the behavioral

development would emerge in isolation of interaction and social experience.

In 2001 Allen Schore introduced another view of infant brain development in an

article he wrote for the Infant Mental Health Journal entitled; “Contributions From The

Decade Of The Brain To Infant Mental Health: An Overview”. In his article Schore

outlined a comprehensive theory of infant brain development which he called the

psychoneurobiological theory. This theory was developed by integrating areas of study

like neurobiology, infant psychiatry, child development, and early childhood education.

He states that the psychoneurobiological theory incorporates the findings of many

sciences to create a theory about infant brain development and the development of

attachment. In discussing his theory, Schore makes clear that findings from this work

have contributed to our understanding of child development in general. He points out

that his work in the areas of brain development and attachment theory, using a biological

perspective, are significant because infant development has long been addressed

primarily through the lenses of the psychological sciences. While respecting and retaining

the significance of the theories from past work in the areas of infant psychiatry,

behavioral pediatrics, and child psychology; Shore, like Pally (1997), Diamond (1988),

and Dubowltz (1995), incorporates the work of neuroscience into his theory of infant

37

mental health development. According to Schore (2001) the emergence of the

psychoneurobiological model of infant brain development indicates that the work of

contemporary neuroscience is now going to focus on the early developmental growth and

failures of the brain. He supports this comment by saying that “neurobiology is currently

exploring early beginnings and brain pathology” (p. 2).

In creating a case for his theory, Schore identifies two practitioners who have

contributed to the field of child development, early childhood studies, and now

psychoneurobiology: Daniel Siegel, a pediatric psychologist and pediatrician and John

Bowlby, a well know researcher in the area of attachment theory and infant/caregiver

behavior.

Daniel Siegel (Siegel, 1999) was identified by Schore as being instrumental in

integrating current cognitive neuroscience into this new developmental model. Schore

(2001, p. 2) quotes Siegel as saying, “we need to expand the area of ‘affective

neuroscience’ if we are to meaningfully explore the neurobilogical basis of interpersonal

relationships, subject experience, and the developing mind." Siegel’s work has been

instrumental in creating an understanding of the biological constructs of the developing

infant brain that manifests in behavioral outcomes like attachment and temperament.

Siegel points out that “the brain has been implicitly seen as a ‘knower’ of the world, as a

socially isolated organ whose purpose is to grasp the inanimate world outside it” (Schore,

p. 2, 2001). Shore points out that the validity to Siegel’s statements are being verified by

current research findings in developmental neuroscience that suggest that the

development of the infant’s emotional brain is directly and actively influenced by his

38

transactions with the animate social and not the inanimate physical environment

(Panksepp,1998; Schore, 2001; Siegel, 1999)

The second researcher Schore cites as he describes the development of his

psychoneurobiological theory is John Bowlby (1969). Bowlby is the founder of a long

held interdisciplinary perspective which outlines the processes of early developmental

and the implications this early brain development has on later mental health. His work is

rich in findings related to attachment and infant/care giver behavior. Although the

empirical work of Bowbly was done in the 1960’s, the findings and implications of his

work are relevant today. Schore used the theories derived from Bowlby’s (1969)

research to create a holistic view of infant mental health development. Of particular

interest to Shores was Bowlby’s hypothesis that suggests that the developmental

processes of an infant are the product of the interaction of a unique genetic endowment

with a particular environment of adaptiveness, and especially of his/her interaction with

the principal figure in that environment, namely his/her mother and that the infant’s

capacity to cope with stress is correlated with certain maternal behaviors. Bowlby (1969)

also suggested that there is a correlation between attachment theory that could frame a

heuristic hypotheses about the etiology and neurophysiology of psychiatric disorders

(Schore, 2001). Schore points out that it is clear that as Bowlby was attempting to shed

light on the theory surrounding infant attachment he was also suggesting a neurological

foundation; a foundation that was undetectable with the technology of his time.

With the advancement of technology and further research, Schore (2001) has been

successful in identifying significant components of a new developmental theory that

incorporates behavior, biology, and environment. In recent years Bowlby (1969)

39

responded to Schore’s work by suggesting that the current developmental conceptions

that integrate the psychological and biological realms are bringing us closer to a complex

biopsychosocial model that can serve as a source of not only the next level of questions

for science, but pragmatic applications of this knowledge. He added that these deeper

investigations of infant mental health, of the primary forces that impact the development

of human nature, can do more than alter the intergenerational transmission of

psychpathology, they can significantly increase the numbers of individuals who posses an

intuitive sense of emotional security, and thereby the quality of life of the infant, child,

and adult members of our society.

In his article, Schore (2001) pulls together the work of Siegel (1999) and Bowlby

(1969) to offer an overview of his interdisciplinary perspective of infant brain

development. In his description, Schore outlines a connection between infant brain

development, attachment theory, stress regulation, and infant mental health. He points

out that research findings indicate that the relationship between infant brain development

and social responses are found to develop architecturally in the right hemisphere of the

developing brain.

Researchers have been helpful in identifying the structures and functions of the

right brain. They have found that the right brain develops the most during the first three

years of life, and that it is centrally involved in processing social-emotional information,

facilitating attachment function, and regulating bodily and affective states (Schore, 2001).

Schore further points out that the right brain is also critical in the control of vital

functions supporting survival and enabling the organism to cope actively and passively

with stress. Schore (2001) has found, through his research, that these regulatory systems

40

are experience dependent and that the experience is embedded in attachment relationships

between the infant and the primary caregiver. Schore’s psychoneurobiological model

suggests that there is a link between secure attachment, development of efficient right

brain regulatory function, and adaptive infant mental health.

Schore is not alone in his assumptions about the developing brain. Michael

Johnson, (Johnson, 2000) a researcher for the Neurological and Pediatric Department at

Johns Hopkins University suggests that the human infant brain undergoes remarkable

organizational changes during intrauterine and postnatal life. According to Johnson, the

success of these organizational changes relies on the intricate dance of the neurological

system. The neurological system is made up of a variety of cell types. The most

significant of the cells being the nerve cells called neurons that were described in detail

earlier by Pally (1997). According to neurobiologists like Johnson (2000) and Diamond

(1988) each of those 100 billion neurons has the capability to generate a chemical

response that is significant to the overall function of the brain. As mentioned before there

are terminals at the base of each of the axons. These terminals house a myriad of

chemicals called neurotransmitters. Neurotransmitters work with the neurological

system to send messages that work to regulate the neural system. Some familiar

neurotransmitters are; dopamine, serotonin, adrenaline, and cortisol.

Neurotransmitters are released from a terminal and enter an available receptor on

a branch of a dendrite. This transmission is based on information input through the

senses. There is a reticular activator at the base of the brain stem that acts as a switching

station between information input and chemical release. As the body experiences sensory

information the brain identifies the activation needed on a chemical level. A specific

41

chemical (neurotransmitter) will be released and will in-turn activate a physiological

response in the body. An example of this is if a person smells smoke while sitting in

his/her living room his/her brain assesses the appropriateness of the smoke in the living

room. The information about the smoke enters the brain system via the olfactory senses.

The brain then determines the chemical reaction that best facilitate the survival of the

person. The neurotransmitter, adrenaline will be released from the terminals of the axons

and enter receptor sites on the branches of dendrites. There will then be a physiological

response. The presence of adrenalin will initiate an increased heart rate, more blood will

be sent to the muscles, and the person will have a heightened sense of alertness. Due to

the chemical response the person will, physiologically, have the ability to get help and

possible put out a fire.

According to Johnston (1995), neurotransmitters are present in the developing

brain while still in utero. His research indicates that the developing brain is very

responsive to these chemicals and the presents of them in the developing brain have been

tied to maternal behavior.

The next section of this paper will discuss one of the neurotransmitters that is

commonly implicated as critical for the appropriate function of the developing brain. The

neurotransmitter is cortisol, known as glucocorticoids. It has many functions in both the

developing and mature brain.

42

The role of cortisol in brain development and function

In describing the significance of cortisol I refer to the work of Eliner-Avishai et

al. (2002) who provide insight into neurochemicals and the adaptation of organisms. The

findings from their work with primates is widely accepted and translated to human

development. According to Eliner-Avishai et al. (2002), the ability of an organism to

adapt to its environment is integral to its survival. Eliner-Avishai et al. point out that

daily life involves confrontations with changing situations that can be physiologically or

psychologically challenging. To cope with these actual or perceived threats of stress, the

ability to alter function and expression of neuronal genes has been developed in the form

of molecular and behavioral stress responses. Eliner-Avishai et al. also suggest that this is

advantageous because it allows rapid behavioral, autonomic and cognitive responses to

stressful circumstances, followed by the prompt re-establishment of the functional steady

state. This means that upon sensing stress, our brain not only initiates rapid secretion of

effected molecules (noradrenalin and adrenal glucocorticoids) but also responds to the

inciting signal with patterned and coordinated changes in programmed gene expression.

Adrenal glucocorticoids, (cortisol in humans), are stress hormones also known as

catecholamine and is known to form the first line of defense for mammals (including

humans) under conditions that threaten homeostasis (ie,. conditions of stress) (Gunnar &

Barr, 1998). Gunnar and Barr describe glucocordicoids as the end products of the

hypothalamic-pituitary-adrenal (HPA) system (Gunnar & Barr, 1998). It has been

determined that glucocorticoids help to mobilize and distribute energy stores, influence

the activity of the immune system and coordinate adaptive behaviors in the brain Gunnar

& Barr, 1998, deWeerth & van Geert, 2002). Given these findings Gunnar and Barr

43

describe (1998) cortisol, and the process that regulates it is part of the story of how early

experiences shape brain development.

Many researchers have contributed to our understanding of the function of

cortisol by citing findings from both human and animal studies. One of the most prolific

of those researchers is M.J. Meaney. Like Eliner-Avishia et al, Meaney et al. (1996)

added clarity to our understanding of the role of cortisol with respect to early brain

development by discussing findings from both animal and human research. The findings

from the animal research suggest that early experiences ‘program’ the brain’s stress

circuits in ways that affect later cognitive competence, emotional responding, and activity

of physiologic systems that orchestrate reactions to stress and challenge. Meaney

continues his discussion about early brain development and chemical reactivity by citing

research that was performed with preterm infants. The research findings identified

prenatal treatments with synthetic glucocordicoids offered great benefit to preterm infant.

However, animal studies of prenatal synthetic glucocorticoids exposure indicate that

there may be some long-term physiological costs of early exposure to excess

glucocordicoids. Further, the effects may not become apparent until later in life

(Meaney, 1996). Caspi et al. (2002) found that some male children who are maltreated

grow up to develop antisocial behaviors while others do not. Their findings suggest that

a functional polymorphism in the gene encoding the neurotransmitter-metabolizing

enzyme monoamine oxidase was found to modulate the effects of maltreatment. In

addition to these research findings, studies in rodents and primates suggest that

responsively and regulation of the Hypothalamic-pituitary-adrenal (HPA) system later in

life may be shaped by social experience during early development. Although it has been

44

determined that cortisol is a major hormonal product of the HPA system in humans more

information about its influence is needed. Current research has provided a partial

understanding of the activity of this neuroendocrine axis. It is speculated that the

regulation of this system may bear importantly on human growth and development. This

development may be influenced by biological as well and environmental factors

(Mattews, 2002, Gunnar, 1998, de Weeth, 2002).

The environment must be taken into account when considering the postnatal

development of the infants brain. The ability of the early environment to program the

HPA axis has been documented in several species. There is considerable evidence that a

similar process can also occur in humans. Studies of animals indicate that the phenotype

of HPA function follows early manipulation which depends on the timing and intensity of

the manipulation, in addition to the gender of the fetus or neonate (Mattews, 2002).

Mattews has found that there is considerable interplay between the HPA and the

hypothalmo-pituitary-gonadal axes, and emerging evidence indicates that this interaction

is modified by early environmental manipulation. Studies are rapidly unraveling the

mechanisms that underlie developmental programming of the HPA axis (Mattews, 2002).

Mattews (2002) states that understanding these mechanisms could hold the key to the

development of therapeutic interventions aimed at reversing the impact of an adverse

intrauterine of neonatal environment.

According to deWeerth, (2002) research in recent years has seen an upsurge in

studies of cortisol response. DeWeerth et al. suggests that this is occurring for two

reasons. The first reason is that cortisol has been found to be a hormone with many

different and important functions such as: the regulation of the energy system, the

45

activation of the immune system, and support of the cognitive system. In addition,

cortisol has been found to play an important role in stress response. It is known to be

secreted when an organism faces a difficult or problematic situation that is perceived as

stressful (deWeerth, et al, 2002)

The second reason for the interest in this area is related to advancements in

laboratory instrumentation and techniques that have occurred in the last decade.

DeWeerth points out that previously cortisol levels could only be measured in serum or

urine. Currently cortisol can be easily measured and reliably assessed in saliva using

fairly cheap, ready made kits (Schwartz, 1998). There is consensus that this simple

measure provides a window on the psychobiology of the stress response through which

those who study children may observe the interacting effects of biological, contextual,

historical, temperamental, and behavioral factors (Schwartz et al. 1998).

An example of work that has evolved from this advancement comes in the form of

a phenomenon described by Gunnar and Vazquez (2001) termed hypocortisolism. The

two researchers start their description of hypocortisolism by saying that for many years

we have known that stress has been associated with increased activity of the limbic-

hypothalamic-pituitary-adrenocortical (LHPA) axis. They go on to say that recent studies

in adults have shown that this neuroendocrine axis may be hyporesponsive in a number of

stress related states (Gunner, Vazquez, 2001). This phenomenon is described as a

paradoxical suppression of the LHPA axis under conditions of trauma and prolonged

stress. They apply this hypothesis to early development by saying that they believe that

the experience of adverse conditions early in life are thought to contribute to the

development of hypocortisolism in adulthood. In describing the ramifications of this

46

phenomenon the researchers point out that low cortisol levels may not be an indication of

low stress but may need to be investigated further to identify the possibility of

hypocortisolism due to early experiences of undue stress.

DeWeerth (2002) conducted research in this area using cortisol measurement

instruments in an attempt to shed some light on some profound facts about infants and

their mothers. By measuring cortisol levels in infants using salivary samples deWeerth

was able to find that before birth human infants have an established basal cortisol level.

According to de Weerth (2002), early in life cortisol is secreted in a pulsatile fashion by

the adrenal cortex and its levels show a strong circadian rhythm, being lowest around

midnight and highest in the early morning hours. She has found that after birth, the

infant's basil cortisol level decreases linearly with age and is negatively related to sleep

(deWeerth, 2002). DeWeerth et al. looked further into cortisol levels in infants and found

that while infants showed relative stability across individuals they displayed great intra-

individual variability across assessments. The mother’s cortisol levels were also assessed.

They found that unlike the infants the mothers display inter-individual variability,

together with a relative stability across assessments. It was also found that the infant’s

inta-individual variability was not affected by gender, or time of visit, nor was it related

to the mother's basal cortisol. Daily measures of basal cortisol taken in a subgroup of

infants indicated the day-to-day intra-individual variability to be of the same magnitude

as the week to week variability.

The works of Gunnar (2001) and deWeerth (2002) have been made possible due

to the advancement of technology and the increased the availability of a variety of

research measures. Determining the appropriate measure has become an issue when

47

considering research design. Schwartz (1998) suggests that the appropriateness of the

cortisol measure may be at least in part based on the research question.

In 1999, a group of 14 researchers contributed to the further understanding of

these measures by gathering to share their expertise in the areas of chemical reactivity

and brain function. Their discussions attempted to identify appropriate instruments,

strategies, and protocol for measuring salivary cortisol. The researchers were unable to

identify one measurement instrument as superior but were in agreement when discussing

the need to determine the instrument based on the research direction. Despite the

nebulous nature of instrument selection there was a clear protocol established for testing

identified by the research group. That protocol included the description of the

appropriate process of sampling by suggesting that a minimum of four or five samples of

saliva should be taken in a day. It was also stated that the measures should be taken at

the one, four, nine, eleven hours after waking. The group of researchers suggests that the

morning and evening cortisol samples are the most important and that the other three or

four samples be random. One researcher, Schwartz (1998) suggests that the data should

be collected over three-four days to get a reliable assessment of a "trait" daily

concentration and for six or more days. The 14 researchers that make up the MacCarthur

Research Network (1999) have settled on a one-day, six-sample protocol. The timing of

the samples is: 1) awakening 2) 45 minutes after awakening 3) 2.5 hours after awakening

4) 8 hours after awakening 5)12 hours after awakening, and 6) bedtime. When working

with an infant population the collection times would need to be adjusted to accommodate

a shorter awake time. The MacCarthur Network also identified some control factors to

consider when using these measures. They are: the stable characteristics of the

48

individuals like age and gender, state characteristics which include menstrual cycles,

contraceptives, and other medications, chronic disease, dynamic characteristics like food

intake, sleep status, exercise, wake time, smoking, and finally physiological

characteristics (Seeman, 2000). Participant protocol is outlined in three steps by the

MacArthur Network as follows: 1.Particpants fill out a log during the day that the

samples are being collected using the MacArthur Salivary Cortisol protocol.

2. Participants fill out a log book about how many controls they experienced (smoking,

exercise, drinking, stress of the day) 3. Participants are interviewed about the stress of the

day in comparison to other days they have (Stewart, 1999).

Cortisol is only one piece of the biological puzzle. It is however, a piece that can

provide insight into the inter-functioning of the developing human brain. The current

measurability of neurochemicals like cortisol along with the growing understanding of

the profound implications of these chemicals will lead researchers to further their

investigations in this area. The findings from future research will have a profound impact

on our understanding and subsequent treatment of infants. It will answer questions for

researchers like Gunnar and Barr (1998) who profoundly identified the research

phenomenon by stating that cortisol and the process that regulates it is part of the story of

how early experiences shape brain development.

Breast feeding: Biological and Social Factors

For many infants one of those early experiences is breast feeding. Cortisol has

been found in breast milk and is presumed to be passed to the infant during feedings.

Because cortisol is also known to influence brain activity we can assume that there is a

49

correlation between breast feeding and brain development with cortisol acting as a

structural agent. As mentioned in part two, when looking at infant brain development we

need to consider the influences of the environment. Breast feeding is without question a

significant environmental factor. Significant findings in the area of breast feeding and

brain development have been reported in recent studies by (Mezzacappa et al. (2002);

Rolf & Keil (2002); Nelson & Panksepp (1998); Fleming et al. (1999). The studies have

examined the sociological, nutritional, and biological impacts of breast feeding on human

and animal offspring.

An example of animal research that has provided significant data that was

transferred to human constructs was conducted Cook (1997). Cook specifically

investigated the neuropeptide cortisol using sheep as a subject group. In his work with

sheep, Cook found that the cortisol response to stress appears to differ between lactating

and non-lactating animals. He found that lactating animals had a slightly higher basal

level of cortisol and a lower cortisol response to stress than their non-lactating

counterparts.

Another research team that uses both animals and humans to better understand

human brain development is led by Elizabeth Mezzacappa (Messacappa, E.S., Tu, A.Y.,

Myers, M.M. (2002). Using the findings of researchers like Cook, Mezzacappa’s

research team has shed light on the issue of breast feeding as an environmental construct

in human infants. In their research, which was conducted using human infants, the

research team generated data that suggests that breast feeding reduces maternal subjective

and physiological responses to stress and negative moods (Mezzacappa et al., 2002). The

findings also indicate that weaning appears to increase maternal subjective and

50

physiological responses to stress. To compliment these findings Mezzacappa et al. also

cites research that has identified differences in mood and perceived stress using breast

feeding subjects. These studies indicate that breast-feeding buffers the physiological

responses in the laboratory. In analyzing these findings, data from the breast fed subjects

were compared to the bottle fed subjects. The findings from this comparison indicated

that breast fed infants had lower baseline electrodermal conductance and less heart rate

reactivity to stimuli. The same study also compared breast and bottle feeding in relation

to cortisol levels and found that breast fed infants had lower cortisol levels when

introduced to an exercise experience. This finding suggests that the breast fed infants had

a blunted hypothalamic-pituitary-adrenal (HPA) axis response to stress.

Fleming et al. (1999) provided more meaning to this phenomenon by reporting

findings from research that was conducted in an attempt to identify the sensory,

endocrine, and neural mechanisms that underlie early mother-infant relationship

by encompassing findings related to the cues and characteristics that are present during

responsive mother-infant interactions. One of the areas of study focused on the chemical

reactivity present during lactation. The findings from this particular interest area were

established using the chemical assay measurement instruments that Schwartz et al. (1998)

referenced in their work. One significant finding that emerged from their assay results

suggested that both the mothers’ and their offspring showed enhanced neuropeptides

(cortisol, oxytocin and endogenous opioids) during the lactation period (Fleming et al.,

1999). These neuropeptides are thought to be instrumental in the responsiveness of the

mother and infant to each other. It is also thought that the presence of these

neuropeptides are directly related to the presence of affiliate behaviors in the mother-

51

infant pair such as enhanced olfactory function, regulated suckling, and heightened

maternal behaviors (Fleming, 1999). Knowing that maternal behavior and infant

responsiveness are the precursor to later relationship structures these findings provide us

with a biological view of bonding and attachment. The findings also make clear the

notion that behaviors that facilitate the structure of the mother-infant relationship are

directly correlated to the chemical production in the maternal and infantile biological

systems (Fleming et al., 1999). Fleming et al. (1999) suggest that these findings indicate

that early experiences and associated brain changes clearly affect how infants interact

with the world in general, and with their own offspring as adults. Fleming et al. (1999)

say that establishing exactly how the neurobiological changes produced by early

experiences might persistently alter the neurobiological mechanisms providing for later

maternal learning and plasticity constitutes the next challenge.

Nelson and Panksepp (1998), have also been instrumental in identifying the

components of research that are significant to our understanding of the biological impacts

of breast feeding. Nelson and Panksepp’s research used rats to study the biological

substrates of attachment and provided another segway from animal research to human

implications. The researchers investigated maternal behavior, sexual behavior, social

memory, and the role of neurochemicals. In determining the significance of each of these

research areas Nelson and Panksepp investigated the role feeding had on infant

development and subsequent attachment. They suggested that milk plays at least some

role in the formation of attachment in mammalian infants. They go on to say that milk

transfer is one of the primary functions of the mammalian infant-mother relationship and

that milk infusion has been shown to induce a dramatic behavioral activation in rat pups

52

(1998). They go on to mention that milk transfer is an important part of regulating sleep-

wake, and arousal cycles, and has been shown to induce odor preferences in rat pups.

They suggest further that milk transfer is likely to contribute to the funneling of behavior

toward the mother, and as a stimulus around which infant’s behavior is organized (1998).

When considering the significance of the biological substrates of mammalian

milk, Nelson and Panksepp (1998) indicate that there is considerable evidence that

suggests that the neurochemical systems are physiological components of these sensory

domains in infant rats and are a core part in the physiological regulation of mother

directed behaviors. Based on these findings Nelson and Panksepp believe that

attachment and affiliative behaviors displayed by rats and other mammalian infants may

represent the emergence of adult affiliative brain systems and the ontogeny of a core

attachment/affiliative motivational system in the brain.

Berthold Koletzko, Maria Rodriguez-Palmero, Hans Demmelmair, Natasa Fidler,

Robert Jensen, and Thorsten Sauerwald (2001) took the work of Nelson and Panksepp

and applied their findings from rodent studies to human milk. The research of Koletzko

et al. (2001) focused primarily on the nutritional substrates of human milk; identifying

that human milk from healthy well nourished mothers is the preferred form of feeding for

all healthy new born infants and adds that the nutrients found in human milk support

normal growth and development in infants. Because the focus of their research was

primarily nutritional substrates of human breast milk they describe in detail the

composition of human milk related to vitamins, fatty acids and other nutritional

components. Koletzko et al. (p. 3, 2001), describe the existence of these nutritional

substrates as precursors to what they call the “indispensable structural components of

53

cellular membranes” and suggest that they are deposited to a great extent in the growing

brain and retina during perinatal development.

Through the research endeavors of people like Fleming (1999), Mezzacappa

(2002), Nelson & Panksepp (1998), and Kolezko (2001), we now have a better

understanding of the role breast feeding plays in social, cognitive, and biological

development. These findings have contributed to the field of early childhood studies by

generating empirical data that has led to broader thinking in the area as well as the

generation of more questions to be answered in future research. One area of research that

has generated more questions for further research is the existence of neurochemicals like

cortisol in mammalian milk. Using what is currently known about the significance of

these neurochemicals in mammalian milk researchers can further investigate the

relationships between cortisol, breast feeding, and infant behavior.

The Study of Infant Behavior

Infant behavior is a broad field of study that has a rich research history. This

history has led to many categorical views of infant behavior. The most common

categories are infant temperament, attachment, and regulation. Each of these categories

has significant and unique views of infant development. Each of these categories also

offers significant insight into the developmental process by which infants assimilate to

the social world they enter at birth.

Nathan Fox and Heather Henderson (1999) have studied infant behavior and

suggested that infant behavior is synonymous with temperament. They define infant

temperament as behavioral styles that appear early in life and are a direct result of

54

neurobiological factors (Fox & Henderson, 1999). They state (1999) that temperament is

thought to describe those aspects of behavioral responding that are not due to interactions

with caregivers or other environmental influences, but rather those aspects that are

present from birth and are biologically based. They believe that individual differences in

young infants’ reactivity are a result of the excitability of the central nervous system that

can be measured using autonomic, endocrine, and central nervous system changes in

response to stimulation.

Jerome Kagan (1996) has a similar model in which he suggests that

temperamental inhibitions in infants vary in the degrees of their reactivity to novelty and

mild stress. He feels that this reactivity is a function of the excitability of certain nuclei

within the amygdala and surrounding regions such as the central gray region of the brain.

Davis, Hitchcock & Rosen’s (1987) work with rats brought clarity to our

understanding of the neurological function of the developing brain that describe the

neural circuits and structures of the brain stem and mid-brain and are associated with the

animals’ heightened fear responses as a result of conditioning, and are universally

understood to be translated to human behavior. Davis et. al (1987) and LeDoux (1989)

report that the neural systems involved in the heightened fear responses are measured

during a classical conditioning paradigm with rats involving a tone or light and shock.

According to their findings this system appears to underlie both the behavioral (e.g.

startle, freezing, and negative vocalization) and physiological (e.g. increased heart rate,

and stress hormones) signs of fear demonstrated in rats (Fox & Henderson, 1999).

LeDoux argues that human fear reactions originate from the same neural system (Kagan,

1996; Fox & Henderson, 1999).

55

The study of infant temperament has long been viewed in tandem with infant

attachment behavior. Mary Ainsworth began work in this area in the 1950’s and has

generated volumes of research findings that indicate that there is a correction between

infant temperament and infant attachment behaviors. Colleagues such as Bowlby

(1969), and Lewis (2000) who have continued Ainsworth’s work have agreed that there is

a relationship between infant temperament and social attachment. Lewis et al. (2000),

suggest that the concept of attachment has captured our attention because it seeks to

explain, in part, the origins of social and emotional behaviors. Lewis et al. say (p. 707,

2000) that attachment theory has generated an enormous amount of theoretical and

empirical work. Work that suggests attachment is essentially an infant’s early overt

behavior that later becomes a representations of the interpersonal experience he/she had.

It is thought that a child’s model of attachment relationship is organized around the

history of the caregiver’s response to the infant’s actions. It is also thought that the

infant’s behaviors are the overt representations of the cognitive process that he/she do not

yet have language to support.

This historical work has led to the identification of attachment categories. These

categories were determined by using a behavior measurement instrument called the

Ainsworth Strange Situation Test. The Ainsworth Strange Situation Test is thought to

capture overt behaviors that reflect an infant’s internal working model of attachment

relationships. Louise Hertsgaard et al. (1995), describe the process by which the

Ainsworth Strange Situation is administered. They describe the measure as a cumulative

stress paradigm test consisting of seven-three minute episodes, including two separations

and two reunions. The results of the measurements are determined by observing sets of

56

behaviors that move the individual toward and maintain proximity with an attachment

figure (Lewis et al. 2000). Four principle attachment patterns have emerged from this

empirical research. It is thought that these patterns reflect and predict important aspects

of social development (Cassidy & Berlin, 1994).

The attachment patterns are: Secure, Avoidant, Resistant/ambivalent, and

Disorganized/disoriented attachment. Secure attachment is characterized by the infant’s

willingness to explore rooms and toys with interest in preseparation episodes. The infant

shows signs of missing the parent during separation, often crying by the second

separation. The infant shows an obvious preference for the parent over the stranger and

greets the parent actively, usually initiating physical contact. There is usually some

contact maintaining by second reunion, but then settles and returns to play (Siegel, 1999)

Avoidant attachment is characterized by the infants failure to cry on separation from the

parent. The infant actively avoids and ignores the parent on reunion (i.e., by moving

away, turning away, or leaning out of arms when picked up). There is little or no

proximity or contact seeking, no distress, and no anger. The response to the parent

appears unemotional. The infant focuses on toys or the environment throughout

procedure. Resistant or ambivalent attachment is characterized by an appearance of

distress even prior to the separation, with little exploration. There is a preoccupation with

the parent throughout procedure. The infant may seem angry or passive. The infant fails

to settle and takes comfort in parent on reunion, and usually continues to focus on the

parent and cry. The infant also often fails to return to exploration after reunion (Siegel, p.

74, 1999). The last attachment behavior is called disorganized. This attachment pattern

is characterized by the infant displaying disorganized or disoriented behaviors in the

57

parent’s presence, suggesting a temporary collapse of behavioral strategies. For example,

the infant may freeze with a trance-like expression. The infant may rise at the parent’s

entrance, then fall prone and huddle on the floor; or may cling while crying hard and

leaning away gaze averted (Siegel, p. 74, 1999).

In recent discussions about this long held view of attachment, researchers have

questioned the foundational principles on which this theory of attachment was

established. Mangelsdorf et al. (2000) suggests that we need to look at Ainsworth’s

attachment theory through new lenses. They indicate that the Strange Situation

procedures have more to do with endogenous characteristics of the infants than to the

sensitivity of maternal care. Mangelsdorff et al. ( 2000) suggest that there are three

explanations of the association between attachment and temperament. The first

explanation is that temperament and attachment are orthogonal constructs

(temperamental variations may influence aspects of behavior, but not the overall

organization of security at attachment). The second explanation is that temperament may

determine what type of insecurity (avoidant or resistant) the child manifests, and the third

explanation is that relationship history totally transforms temperamental variations that

consequently make little or no contribution to the relationship. Although the Ainsworth

model of attachment is still held in high regard, there are other models emerging based on

the Strange Situation model but with new views, like those of Mangelsdorff et al.(2000)

and others who bring modern measure for future analysis.

Rothbart & Hwang (2002) describe temperament as the individual differences in

emotional and motor reactivity and in the attentional capacities that support regulation.

Rothbart and Hwang are leading experts in the area of infant behavior assessment and

58

have suggested that the ability to measure infant behavior has increased during the past

century. Measures have been created that have allowed early temperament to be

empirically linked to later personality, the development of conscience and empathy, as

well as the development of behavior problems (Rothbart, Hwang, 2002). Rothbart has

been a pioneer in the development of these behavior measures with the development of

the Infant Behavior Questionnaire (IBQ; Rothbart, 1981) which is designed to measure

infant temperament using parent report. Following the original development and

implementation of the IBQ, Rothbart redesigned the instrument to included nine new

scales and minor modifications. The revised IBQ (IBQ-R) (Rothbart & Gartstien, 2002

measures three broad areas of personality development. They are surgency/extraversion,

negative affectivity, and orienting/regulation. The Rothbart measurement instrument is

widely used and highly regarded as a valid instrument for the measurement of infant

behavior.

The theory of infant temperament that expands on the work of Fox &Henderson

(1999), and Kagan (1996) is described in an article by Schore (2001) entitled, “Effects of

Secure Attachment Relationships on Right Brain Development, Affect Regulation, and

Infant Mental Health”. In his article Schore identifies infant temperament by using a

psychoneurobioloical theory of infant brain development. Recognizing the work of

neurobiologists Schore introduces the concept of infant regulation. Schore describes

infant regulation as integration of both the biological and the psychological realms.

Schore states (2001) that because the concept of regulation is a theoretical concept that is

shared by many clinical and basic sciences it can be used to further develop models of

normal and abnormal structure. He states that this interdisciplinary research and the

59

clinical data affirms the concept that in infancy and beyond, the regulation of affect is a

central organizing principle of human development. Although the concept of regulation

is not a new one it takes on new elements when viewed through Schore’s model. The

model introduces the role biology plays in infant regulation. To support this Schore cites

Damasio (1994)as saying “ Emotions are the highest order direct expression of

bioregulation in complex organisms” (Schore,2001, p.10). The connections between

infant brain development and the significance of infant regulation of the development of

infant mental health structures is pointed out by Damasio (1994) who notes that the

overall function of the brain is to be well informed about what is going on in the rest of

the body proper; about what goes on in itself; and about what the environment

surrounding the organism, so that suitable survivable accommodations can be achieved

between the organism and the environment. The psychobiological model incorporates

components from the maturational theory and components from the cognitive

developmental theory by looking at how the biological architecture of the developing

brain work together with environmental conditions to establish patterns of infant

regulation.

Schore’s (2001) psychoneurobiological theory is supported further by the research

of Gunnar and Donzella (2002) that specifically identified the role that cortisol plays in

early brain development. This theory of infant regulation is supported by citing animal

studies suggesting that responsivity and regulation of the biological system later in life

may be shaped by social experience during early development. They indicate that

cortisol is found to play a part in the regulation of the biological system in humans.

Although it provides only a partial understanding of the activity of this neuroendocrine

60

axis, its regulation may bear importantly on human growth and development. Gunnar and

Donzella (2002) identify developmental studies of cortisol and behavior in children, birth

through age five and describe the development of the social buffering of cortisol

responses, and established a functional analogue of the rodent stress hyporeponsive

period by the time children are about 12 months of age. They further describe the

sensitivity of cortisol activity to variations in care quality among infants and toddlers,

along with evidence that children with negative emotional temperaments may be most

likely to exhibit elevations in cortisol under conditions of less than optimal care.

Regulation is identified by Gunnar and Donzella (2002) in reference to the

activity in the limbic hypothalamic-pituitary-adrenocortical system. Gunnar and

Donzella (2002) cite the past research findings of S. Levine from 1994 and Suomi from

1991. The work of Levine and Suomi was conducted using rodents and primates. Like

other studies the research findings are universally translated to be used as human

behavior predictors. Gunner states (2002) that Levine (1994) and Suomi (1991) have

identified that caregivers play an important role in regulating activity in the biological

system during development. In rodents, licking and grooming by the dam and the

delivery of milk into the gut maintain the adrenal hyporesponsive period, a period

between postnatal day four through the fourteenth day. During this time it is difficult, yet

not impossible to elevate glucocordicoids levels ( i.e., cortisol). Gunnar and Donzella

(2002) discuss the findings of Graham and Heim (1999) who translated animal findings

to human behavioral predictors in their paper “Effects of Neonatal Stress on Brain

Development: Implications for Psychopathology.” In their paper they cited a theoretical

review that suggested that adverse experiences early in life predispose individuals to

61

affective pathology in part through affecting activity of the biological system. Although

this research has identified a strong correlation between treatment early in life and the

biological development of the brain in animals, there is little information about these

issues in human infants.

The study of infant behavior is a complicated area of research. Defining infant

behavior is difficult; equally difficult is identifying the appropriate measurement of the

phenomenon. Whether the researcher is viewing infant behavior as temperament,

behavioral, or regulatory there are common characteristics found in the investigations.

Biology, social interactions, and environmental influences are woven throughout all of

these works. There are many opportunities for further research in this area. Future

findings will shed light on the complexities of the biological, social, and psychological

development of infant behavior. Table 2.1 provides a summary of the research findings.

Table 2.1 Summary of Research

Research analysis

Research topics Authors Methods Finding

Infant stress and cortisol reactivity

Gunnar, Vazques, Donzella,

Infant inoculation, strange situation, cortisol measures and philological testing

Infants did not show elevated cortisol levels during inoculation, or strange situation when they were in a stable maternal relationship. Infants in unstable relationship had elevated cortisol levels.

Maternal responsiveness/cortisol levels

Fleming et al.

Mezzacappa et al.

Maternal responsiveness is observed. Ability to smell infant and demonstrate maternal

Breast feeding mothers were better able to identify their infants by smell. Breast feeding mothers demonstrated higher maternal behaviors. Lactating mothers

62

behaviorshad higher cortisol levels

Lactation and chemical concentrations

Fleming et al. Breast milk is analyzed for chemical content

Breast milk is found to have cortisol, oxytosin, and endogines opieds.

LHPA cortisol feedback loop

Ladoux, Gunnar Chemical reaction is tracked through the LHPA system

Cortisol is released through the adrenal glands and moves through the brain by way of the blood stream. Stress initiates an increase in cortisol release. Too much cortisol impairs hippocampal function which impairs memory, problem solving and spatial skills

Negative impacts of cortisol

LaDoux, GunnarStudy of adults who have been identified as having loss as children. The study of abused and neglected children

Both adults and children who were identified as traumatized showed decreased cortisol reactivity which is thought to be from the body and brain dealing with stress so much that the system that regulates stress response becomes impaired due to too much cortisol in the system.

Positive impacts of cortisol

Pally, Gunnar, LaDoux

The study of developing brains

Cortisol has been found to contribute positively to normal brain development and has been seen as a necessary chemical in both brain development and brain function.

63

Infant temperament measures

Putnam,Rothbart,

GartstienStudies investigated the correlation between infant temperament and parental reporting. Longitudinal studies were used both with parents and with secondary caregivers to see if there was accuracy in the reporting.

The studies found that parents were accurate reporters of their infants’ behavior and that there were situations observed by parents that were thought to be missed in a laboratory setting.

Factors that impact parenting competence

Carnegie Foundation

Gunnar & Donzella

Lupien et al.

Studies of census information was used to identify significant patterns in the treatment of young children in the US

Children from lower SES families were found to be placed in child care settings of lower quality.

There was also a correlation between SES and incidence of maltreatment on the part of the parent.

Attachment and infant temperament

Mangseldorff Video taped interactions of mothers and infants when the infants were 8 months. Maternal personality tests and an assessments of infant temperament and was filled out by mothers. 12 month assessment

78% of the infants were correctly classified as secure, resistant, or avoidant. The study results suggested that there is a need to consider both the mother’s and the child’s characteristics when predicting attachment

64

of same children using the Strange Situation measure

Summary of Literature Review

In order to identify the impact breast feeding has on the developing brain as well

as infant behavior I investigated four separate yet interconnected disciplines: brain

development, how cortisol interacts during brain development, the role breast feeding

plays in brain development, and the subsequent development of infant behavior. Brain

development is thought to be a delicate biological interplay of genetics and environment.

Research suggests that the brain undergoes a surge of development during the first five

years of a child’s life. Optimal development was been found to require a balance of

nutritional, social and environmental supports.

The second area I investigated was the stress hormone cortisol. Cortisol was

found to be present in developing brain and was also found to play an important role in

infant brain development and behavioral regulation. Cortisol was also determined to be

present in maternal breast milk. The relationship between the presence of cortisol in

breast milk and the impact it has on the feeding infant is yet to be studied but has

interesting implications for the identification of maternal stress response and infant

reactivity.

The third area of study was the impact breast feeding has on infant brain

development. Human milk has been found to have a positive impact on the development

65

of the brain as well as qualities that strengthen the developing immune system. The

emotional benefits were also identified. It is clear from a number of research sources that

the social interaction required for breast feeding positively contributes to infant-mother

bonding process.

The final area of study was infant behavior. Behavior was found to be an offshoot

of all of the developmental constructs that take place during infancy. Researchers agree

that infant behavior, although somewhat nebulous in its nature, is biologically, socially,

and environmentally driven.

What has been identified through this investigation of literature is that there are

clear relationships between brain development, chemical reactivity in the brain, breast

feeding, and infant behavior. The scope of these relationships is still unclear but within

the realm of further research.

66

CHAPTER 3METHODOLOGY

The purpose of this study was to determine whether the concentration of cortisol

in maternal breast milk has a direct impact on infant behavior independent of the

mothers’ stress levels. In order to answer the question a study was conducted in parts of

two Midwestern states. Breast feeding mothers were recruited from a variety of parent

support organizations. Data gathered included maternal stress, infant behavior, breast

milk cortisol levels, and family income and education levels. In order to investigate

relationships a correlational analysis was conducted as well as a regression analysis.

Description of the methodology of this study is organized into several different

sections. Included in this chapter are descriptions of the sample, research design,

research variables, instruments, and the methods of data analysis.

Description of the Sample

The study took place with the consent of participants from three demographic areas in

two Midwestern states. Two of the areas were urban and one was rural. All three of the

demographic areas had universities within 30 miles, large cities within 20 miles and large

67

hospitals within 20 miles. All three areas also had large schools within them and active

parent support porgrams.

Participants of this study were recruted through public health programs, public school

programs, and breast feeding support programs. The researcher contacted the program

managers, described the study and asked them if they would announce the study to the

mothers in their programs. Program managers described the research and distributed

research packets to interested mothers. Research packets were distributed to the program

managers based on the estimated number of breast feeding participants enrolled in their

programs. In a few cases, the mothers asked that the researcher call them, which she did.

The phone conversation provided an opportunity for the mothers to discuss the details of

the study with the researcher before deciding whether or not to participate in the study.

Through these recruiting processes, sixty-seven mother/ infant pairs self selected to be

involved in the study by taking research packets. Twenty-one of the mother/infant pairs

completed the study. Two of those mother/infant pairs had missing data and had to be

eliminated from the study.

The final sample used for the study was 19 breast feeding mothers and their

infants. The mothers ranged in age from 25-39 years. The infants ranged in age from 2

months to 14 months of age. There were 10 female infants and 9 male infants. The

mothers were primarily from upper-middle class house holds, were married and had

college educations. More details are provided in Chapter 4.

Description of research design

68

The research project took place over a 4 month period of time from November 2003 to

February 2004. This included the initial contact with the program managers until the

final data were collected. The time since February has been spent analyzing the data and

writing up the results.

The sixty-seven mothers who agreed to participate in the study were provided with a

ziplock baggie that contained all of the material necessary to participate in the study.

The baggie contents consised of; an SES inventroy (Appendix E), the State/Trait Anxiety

Inventory (Appendix C), The Rothbart Infant Behavior Inventroy (Appendix D) , a

consent form (Appendix A), a parent information sheet (Appendix B), directions, and

three small baggies labeled week one, week two and week three (the instruments will be

described later in the chapter). Each of the small baggies contained; three baggies that

were labled breast milk containers (a.m., noon, p.m.).

The mothers were given three weeks to complete the study. They were asked to fill

out the State/Trait Anxiety Inventroy, the Rothbart Infant Behavior Inventory and the

SES inventroy and place the documents into the large baggie provided. The mothers

were also asked to choose three different days on three different weeks to provide fluid

samples. On each of the selected days the mothers provided three samples of breast milk;

one in the morning, one at noon and one in the evening. The breast milk samples were

stored in the mother’s freezer until all of the samples had been gathered, the mothers

contacted the researcher when the samples were ready for collection. The reseacher

picked the samples up at the mothers homes.

The maternal State/Trait anxiety inventory, the infant behavior inventory and the SES

inventroy were collected by the researcher and scored. The breast milk samples were

69

collected at the same time and sent to the Salimetrics Lab at Penn State University for

analysis in a frozen state.

Breast milk assay results were provided by Salimetrics Lab. The State/Trait

Anxiety scores, SES inventory results, and the Rothbart Infant Questionarre scores were

tabulated by the researcher. All of the data were entered into Statistical Package for the

Social Sciences 12.0 (SPSS, 2004) for further analysis.

Instrumentation

Several instruments were used to measure the study variables. Maternal stress

was measured using the State/Trait anxiety inventory (Consulting Psychological Inc,

1977). Cortisol concentration levels in breast milk were found using a highly-sensitive

immunoassay (Granger, 2004). Infant behavior was measured using the Rothbart Infant

Behavior Questionnaire (Rothbart 2000), the socioeconomic status of the family was

assessed using a tool generated by the researcher to determine economic and education

levels of the mother (Appendix E). The following will describe each of the instruments

used in this study and the scoring procedures used.

State Trait Anxiety Inventory:

The State Trait Anxiety Inventory is a standardized, norm referenced instrument

that was designed by Charles D. Spielberg and published by Mind Garden, Inc. to

measure anxiety in adults (Consulting Psychological Inc, 1977).

The inventory was designed to make distinctions between temporary or acute

conditions of anxiety referred to as state anxiety and general long term or chronic anxiety

referred to as trait anxiety. In order to measure these two kinds of stress the inventory

70

was divided into two sections. Section one: State Inventory which assessed acute stress.

Section two: Trait Inventory which assessed chronic stress. Each of the two sections had

20 multiple choice questions written at the sixth grade reading level. The instrument was

scored using a Likert scale ranging from 1-4. The State Anxiety Inventory had a

response for number 1 of not at all, 2 somewhat, 3 moderately so, and 4 very much so.

The Trait inventory was scored with a response for number 1 of almost never, 2,

sometimes, 3, often, and 4, almost always. This scoring system provided for a possible

minimum score of 20 and a possible maximum score of 80 on each of the inventories.

On this particular measurement tool, lower numbers represent lower stress, while higher

numbers represent high stress. A sample State question is “I feel pleasant”. A sample

Trait question is “I feel satisfied”. Based self report, it took the mothers about ten

minutes to complete both the State and Trait Anxiety Inventory.

The data presented by Spielberg for the normative sample derived from a sample

of working adults, college students and military recruits were divided into three

subgroups by ages (19-39, 40-49, 50-69) and subdivided by gender. The mean scores of

the normative sample for State-Anxiety scores for women between 19-39 is 36.17 with a

standard deviation of 10.96. The mean scores of the normative sample for women

between 19-39 for the Trait anxiety test is 36.15 with a standard deviation of 9.53.

Enzyme-Linked Immunosorbent Assay (ELISA):

A highly sensitive enzyme immunoassay (Granger, 2004) was used. Due to the

changes in cortisol levels caused by the circadian rhythm the ELISA was used to measure

breast milk samples from three different times of a single day on three different days.

71

This enzyme immunoassay was developed to measure microamounts of chemicals such

as cortisol. According to Douglas Granger of Salimetrics Research Laboratory, the test

on the mothers breast milk was conducted two times per sample (original and

duplication) using 25ul of breast milk (for singlet determinations), the test was found to

have a lower limit of sensitivity of .003 ug/dl, range of sensitivity form .003 to 1.2 ug/dl,

and average intra-and inter-assay coefficient of variation 4.13% and 8.89% respectively.

Method accuracy, determine by spike recovery, and linearity, determined by serial

dilution are 105% and 95%. Values from matched serum and saliva samples show the

expected strong linear relationship, r(17)=>.94,p<.0001.(Granger, 2004)

The assay measures antigens and uses a solid phase assay method that requires the

separation of reagents. The measurement begins with an antibody bound to a polystyrene

well plus the antigen to be measured (cortisol). An enzyme conjugate is then added to the

well with bound antigen-antibody or immune complex. Next a substrate is added to the

enzyme conjugate which is bound to the immune complex. There are changes due to the

presence of the enzyme conjugate bound to the immune complex. If the test is positive a

change in color will occur. The amount of color is an index of the amount of cortisol,

and is measured by spectroscopy (Granger, 2004).

Infant Behavior Inventory-Revised (IBQ-R):

The IBQ-R has been designed to use parent reports to measure behavior in infants

between the ages of three and twelve months. According to Mary Rothbart (2002), the

researcher who designed the IBQ-R, the inventory assesses the following dimensions of

behavior (Rothbart, 2002):

Table 3.1 Rothbart Behavior Descriptions

72

Activity level: Movement of arms and legs, squirming and locomotor activity.

Distress to limitations Baby's fussing, crying or showing distress while a) in a confining place or position; b) involved in caretaking activities; c) unable to perform a desired action.

Approach:  Rapid approach, excitement, and positive anticipation of pleasurable activity

Duration of orienting: The baby's attention to and/or interaction with a single object for extended periods of time.

Smiling and laughter:  Smiling or laughter from the child in general caretaking and play situations.

Vocal reactivity:  Amount of vocalization exhibited by the baby in daily activities. Sadness: General low mood; lowered mood and activity specifically related to personal suffering, physical state, object loss, or inability to perform a desired action.

Perceptual sensitivity: Amount of detection of slight, low intensity stimuli from the external environment.

High intensity pleasure: Amount of pleasure or enjoyment related to high stimulus intensity, rate, complexity, novelty, and incongruity.

Low intensity pleasure:  Amount of pleasure or enjoyment related to situations involving low stimulus intensity, rate, complexity, novelty, and incongruity.

Cuddliness: The baby's expression of enjoyment and molding of the body to being held by a caregiver.

Soothability: Baby's reduction of fussing, crying, or distress when the caretaker uses soothing techniques.

Falling reactivity/rate/of recovery from distress:

Rate of recovery from peak distress, excitement, or general arousal; ease of falling asleep

Fear: The baby's startle or distress to sudden

73

changes in stimulation, novel physical objects or social stimuli; inhibited approach to novelty.

Sadness: General low mood; lowered mood and activity specifically related to personal suffering, physical state, object loss, or inability to perform a desired action.

The instrument uses a seven point Likert Scale of measurement. Parents are

asked to assess the frequency of infant behavior based on the experiences of the past

seven days. The measurement points on the Likert Scale are: (1) never, (2) very rarely,

(3) less than half of the time, (4) about half of the time, (5) more that half of the time, (6)

almost always, (7) always, (x) does not apply (Rothbart, 2002).

The revised IBQ (IBQ-R) instrument was used for this study. The IBQ-R

modifications added nine new scales and minor modifications to the existing seven scales

of the IBQ creating total scales. The IBQ-R also added three scales which was derived

from research analysis done by Rothbart and Gartstein (2003). The scales of the

instrument are negative affectivity, positive affectivity, and orientation/regulation. For

these analyses the children were divided into three age groups to control for the normal

developmental differences found throughout infancy. The age groups were: age 1: 3-6

months, age 2: 6-9 months, and age 3: 9-12 months. The mean scores 360 children was

as follows: Positive Reactivity (Group 1: Mean=3.97, SD=.81; Group:2 Mean-4.53,

SD=.57; Group 3 Mean=4.78, SD .55). Negative Reactivity (Group 1: Mean=2.55,

SD=.67; Group 2: Mean=2.77, SD .63, Group 3: Mean=3.04; SD .63).

Orientation/Regulation (Group 1and 2: Mean=4.39, SD .59, Group 3: Mean=4.61,

SD .63).

74

According to Gartstein and Rothbart “A number of studies have provided

validational support for the original IBQ” (Gartstein & Rothbart, 2003 p.165) They cite

the work of Bridges, Plamer, Marales, Hurtado, & Tsai, 1993; Crockenberg & Acredolo,

1993; Goldsmith & Rothbart 1996, Gibbons, Johnson & McDonough, 1989; Rothbart

1986 to support their claim. They continue on to state that the reliability, convergent

validity, and relative stability have also been demonstrated for the IBQ with infants as

young as two weeks of age through the work of Worobey,1986; Woroby & Blajda, 1989.

In addressing the issue of the potential inaccuracies of the parent reporting

structure of this instrument Samuel Putnam, Lesa Ellis and Mary K. Rothbart describe the

positive functions of this reporting method in a paper written in 2001 called The Structure

of Temperament from Infancy through Adolescence. They state that “Although Kagan

(1994) had argued that parent report is not appropriate for the study of temperament,

there is evidence that parents can be reliable observers of their own children and that

parent reports contain substantial objective reliability” (Putnam et al., 2001). Putnam et

al. (2001) continue on to say that both parent report and self report have several

advantages that observational methods do not have. They suggest that “parents view

their children in a wide range of situations that are ethically and logically impossible to

recreate in the laboratory, and that self report allows for the assessment of internalized

reactions and tendencies that cannot be measured via laboratory observation”. These

findings suggest that the potential for the legitimate assessment of children by parents is

likely. These findings support the use of parent report as a method of data collection for

this research.

75

Data Analysis

The purpose of this study was to test the research hypothesis that there is a

relationship between cortisol levels found in breast milk and infant behavior. Data were

analyzed using Spearman Correlation Coefficients and linear regression analysis. The

Spearman Correlation Coefficient was used to determine the direction and degree of

linear relationship between variables. The linear regression was used to determine the

relationship between cortisol and temperament while controlling for the mothers’ level of

stress, which in turn, may affect her interactions with her child. The following will be a

description of the analysis processes used.

Spearman correlation coefficient

Spearman correlation coefficient was used to examine the relationships among the

variables. The analysis ranked the scores first and then calculated the relationship

differences among the ranks (Howell, 2002). The Spearman correlation was used to

determine the results of this study because of the nature of the data. Some of the data

naturally occurred in rank form (Rothbart scores) and other data had rank as a

substitution for raw data (mothers income level). Along with the fact that the Spearman

is appropriate to use with ordinal data, the Spearman also makes no assumptions about

the population and does not assume a bell curve of distribution of the wider population.

According to Howell (2002) the fact that the wider population is irrelevant means that the

test is designed for a specific population. Howell (2002) also suggests that the Spearman

is good for studies with small Ns and home devised tests (SES) and can be tailored to

particular tests.

76

The variables studied were: Positive Affectivity Rothbart scores,

Orientation/Regulation Rothbart score, Negative Affectivity Rothbart scores, mother’s

Trait Anxiety score, mother’s State Anxiety score, the cortisol concentrations in breast

milk, and maternal family income level (SES).

The interpretation of the correlation results was based on the statistical standard

that a p-value of less than .05 is considered statistically significant. The strength and

significance of the correlation was determined by using SPSS 12.0 to identify which

independent variables had relationships with the dependant variable or with each other

(Howell, 2002), in this case infant behavior. The correlation coefficient is a number that

summarizes the direction and degree of linear relation between two variables. An

example of this analysis would be: as cortisol levels in breast milk go up there is a similar

increase in negative infant behavior.

Multiple linear regression-

Multiple linear regression analysis was used. The test was used to determine the

significance of the variables in relation to the dependent variable. A linear regression is a

classic statistical technique that is used to determine the relationship between the

dependent variable and a collection of predictor variables. For example infant behavior

(X) and cortisol levels in breast milk (Y). A linear regression attempts to explain this

relationship with a line of best fit to the data. There is also an underlying assumption that

there is a relationship between the variables. The linear regression model postulates that

Y= a+(bX)+e (Howell, 2002). For this study the use of linear regression was based on

the theory that infant behavior and cortisol levels in maternal breast milk have a

77

significant relationship. Using a forced entry method, the test was created by

programming SPSS 12.0 to use the specified order of independent variables. Infant

behavior, identified by IBQ-R scores, was used as the dependent variables. Independent

variables were added in the following order: maternal stress, SES, and the cortisol

concentrations in breast milk. Mother’s breast milk cortisol levels were entered last so

that the regression analysis would look at all of the other variables first and removes

shared variance before adding cortisol levels to the equation.

78

CHAPTER 4RESULTS

Introduction

The purpose of this study was to determine whether the concentration of cortisol

in maternal milk has a direct impact on infant behavior independent of the mother’s stress

levels. Specifically, this study was designed to answer the following research question:

Does the concentration of cortisol in mother’s breast milk have a relationship with the behavior of the infant?

Hypothesis: Under conditions of normal, not chronic stress, high maternal stress levels led to high cortisol concentrations in breast milk which leads to a disruption of the emotional response circuitry of the infant which results in excessive emotional responsiveness of the infant leading to negative infant temperament characteristics.

In order to answer this question several different data analyses were conducted.

First, the background variables of the participants were analyzed in order to determine the

demographic characteristics of the sample. Second, correlation analyses were conducted

to determine the strength of the relationships between the behavior of infants in the study

and the levels of cortisol found in maternal breast milk, as well as socioeconomic levels

and maternal stress levels. Finally, the data for the various measures were analyzed using

regression to determine the nature of the relationship between stress levels of the

mothers, the behavioral patterns of the infants and the cortisol levels in the breast milk.

79

Description of Sample Characteristics

The study was conducted with the participation of 19 breast feeding mothers and

their infants. The mothers ranged in age from 25-39 years. The infants ranged in age

from 2 months to 14 months of age. There were 10 female infants and 9 male infants.

The women who participated in the study had a range of annual family income

from$11,000.00 to over $100,000.00. The education levels of the mothers ranged from

completion of a high school education to completion of a graduate degree.

Socioeconomic Status

Chart 4.1 shows the distribution of the sample of the women who participated in

this study by family income bracket. The income brackets were divided into five

categories. The categories were:

One: $10,000.00 or lessTwo: $11,000.00-30,000.00Three: $31,000.00-$50,000.00Four: $51,000-$100,000.00Five: $100,000.00 or over

Data show that most of the mothers had a family income between $31,000.00 and

$50,000.00. There were no women with an annual family income of $10,000.00 or less,

5% of the women had an annual family income of $11,000.00 to $30,000.00, 52% of the

women had an annual family income of $31,000.00 to $50,000.00, 26% had an annual

family income of $51,000.00 to $100,000.00 and 15% of the women had an annual

family income of more than $100,000.00. The socioeconomic inventory can be found in

Appendix E.

80

Chart 4.1 Family income

2 3 4 5

SES1

0

2

4

6

8

10

Freq

uenc

y

Mean = 3.53Std. Dev. = 0.841N = 19

$11,000-$30,000 $31,000-$50,000 $51,000-$100,000 over $100,000

Y axis-The number of subjects per income level X axis-Family income level

Chart 4.2 presents the distribution of the education levels of the mothers

who participated in this study. The education levels were divided into

three groups. High school education, college education, and graduate

school education. Fifteen percent of the women had a high school

education, 73% of the women had a college degree, and 10% of the

women had a graduate degree.

81

Chart 4.2 Education levels

0.5 1 1.5 2 2.5 3 3.5

SES2

0

2

4

6

8

10

12

14

Freq

uenc

y

Mean = 1.95Std. Dev. = 0.524N = 19

High school College Graduate School

Y axis-number of subjects per education levelX axis-level of education acquired by subject

Maternal Anxiety Level

The mothers’ anxiety levels were measured using the State/Trait Anxiety

Inventory (Consulting Psychological Inc., 1999). This standardized, norm referenced

instrument is designed to measure anxiety in adults. The inventory makes distinctions

between temporary conditions of anxiety referred to as state anxiety and general long

term anxiety referred to as trait anxiety. The State/Trait Anxiety Inventory can be found

in Appendix C. The instrument is designed by Charles D. Spielberg and published by

Mind Garden, Inc.

As stated in Chapter 3, the State and Trait Anxiety scores from a normative

sample of working adults were divided into three subgroups by age. The groups were 19-

39, 40-49, and 50-69. Scores were also subdivided by gender. The range of scores from

82

the normative sample was 20-80. The mean scores from the normative sample for the

State-Anxiety scores for women between 19-39 was 36.17 with a standard deviation of

10.96. The range of scores from the normative sample for Trait-Anxiety was 21-80. The

mean scores from the normative sample for women between 19-39 for the Trait-Anxiety

test was 36.15 with a standard deviation of 9.53.

All of the women who participated in this study fit into the 19-39 age group. The

range of State Anxiety scores from the current study population was 20-47. Chart 4.3(a)

illustrates the distribution scores for State Anxiety of the women from the current study

population. They had an overall mean of 31 and a standard deviation of 1.25. Using the

normed scale to measure percentile rank of the study population the mean for women in

this study is at the 44th percentile for State Anxiety. Chart 4.3 (b) shows the women in

the current study had a range of Trait Anxiety score from 23-57. The mean Trait Anxiety

scores for the mothers from the current study population was 32 with a standard deviation

of 1.27. Using the normed scale to measure percentile rank of the study population the

mean for women in this study is at the 42nd percentile for Trait Anxiety.

Chart 4.3 State Anxiety and Trait Anxiety scores identified by subject

(a) State (b) Trait

>80 >80

83

Normative Sample has a scale from 20-80 Normative Sample has a scale from 20-80Study Sample has a scale from 20-47 Study Sample has a scale from 23-57Y axis-Subject identification number Y axis-Subject identification numberX axis-State Anxiety Inventory scores X axis-Trait Anxiety Inventory scores

Table 4.1 identifies the study distribution in comparison to the normative sample.

This table illustrates that there is a larger percentage of low raw scores for the current

study participants compared to the normative sample. Table 4.2 illustrates a similar

distribution for the Trait Anxiety scores. Table 4.2 also illustrates that there is a larger

percentage of low raw scores for the current study participants compared to the normative

sample. Tables 4.1, 4.2 and 4.3 illustrate that the participants of the current study have

lower scores for both State and Trait Anxiety than the scores of the normative sample.

This finding suggests that the women who participated in this study are less stressed both

acutely and chronically than those in the normative group.

Table 4.1 State Anxiety score distribution of study participants compared to normative sample

Normative Sample Raw Score

Percentile RankOf Normative Sample

Percent of Subjects from Current Sample

20-25 2-13 2126-31 17-41 5332-37 44-62 7938-43 68-77 9544-49 79-87 10050-55 89-93 No data56-61 94-95 No data62-67 100 No data68-73 100 No data74-79 100 No data80 100 No data

Table 4.2 Trait Anxiety Scores distribution of study participants compared to Normative Sample

Normative Sample Raw Score

Percentile Rank Percent of Subjects from Current Sample

20-25 0-12 1626-31 16-35 48

84

32-37 42-61 8038-43 65-80 9044-49 83-92 9550-55 92-95 10056-61 95-99 No data 62-67 100 No data68-73 100 No data74-80 100 No data

Infant Behavior Inventory

Mothers were asked to assess their infant’s behavior using the Infant Behavior

Questionnaire (IBQ-R). The IBQ-R can be found in Appendix D. As discussed in

Chapter 3 in more detail, the IBQ-R has been designed to use parent reports to measure

behavior in infants between the ages of 3 and 12 months. The infants were divided into

three age groups: age 1: 3-6 months, age 2: 6-9 months, and age 3: 9-12 months to

control for normal developmental differences among ages throughout infancy. The

inventory assessed the following dimensions of behavior: Approach, activity level,

distress to limitations, fear, duration of orientation, smiling and laughter, high pleasure,

low pleasure, soothability, falling reactivity/rate of recovery from distress, cuddliness,

perceptual sensitivity, sadness, and vocal reactivity using a seven point Likert Scale.

The results, presented in Table 4.3 indicate that the mean scores of the infants

who participated in the current study are slightly higher for Positive Affectivity (by .46

points) and Orientation/Regulation (by .43 points) when compared to the normative

sample. With the same comparison, the scores from the current study participants are

lower for the Negative Affectivity (1.56 points). These differences were found for each

of the three domains by adding the mean scores for the three different age groups (2-6. 6-

9, 9-14 months) together and then dividing them by three to get one number. The number

from the current sample was subtracted from the number from the normed sample

85

resulting in a difference. When comparing the distributions it is clear that, although there

are differences in mean scores between the current study group and the normative

sample, there is a similar range of high and low scores for each of the factors. Using the

means from the Rothbart study as a comparison, the data from the current study suggest

that the infants who participated in the current study were generally happy (Positive

Affectivity), and showed few negative behaviors (Negative Affectivity).

Table 4.3 Comparison of means and standard deviations of the infants in current study to the infants in the Rothbart study

AGE GROUP

AGE 1: 2-6 Months

AGE 1: 2-6 Months

AGE 2: 6-9 Months

AGE 2: 6-9 months

AGE 3: 9-14 months

AGE 3: 9-14 months

Study Group Rothbart Current Study

Rothbart Current Study

Rothbart Current Study

Positive Affectivity

3.97(.81) 4.12(.23) 4.53(.57) 5.07(.36) 4.78(.55) 5.47(.61)

Negative Affectivity

2.55(.67) 1.22(.44) 2.77(.63) 1.16(.93) 3.04(.63) 1.29(.59)

Orientation/Regulation

4.16(.63) 4.92(.26) 4.16(.63) 5.07(.24) 4.39(.59) 4.93(.34)

Cortisol Levels

Cortisol levels in the mother’s breast milk were determined by using a highly-

sensitive enzyme immunoassay (Granger, 2004). The test used 25 ul of saliva (for singlet

determination), and had a lower limit of sensitivity of .003 ug/dl, range of sensitivity

from .003 to 1.2 ug/dl, and average intra-and inter-assay coefficients of variation 4.13%

and 8.89% respectively. Method accuracy, was determine by spike recovery, and

linearity, determined by serial dilution are 105% and 95%. Values from matched serum

and saliva samples show the expected strong linear relationship, r(17)=.94, p<.0001.

86

Individual cortisol levels were generated from each of the samples provided. In

order to adjust for the circadian rhythm cortisol levels were calculated to find an Area

Under the Curve (AUC). The Area Under the Curve is a number that takes into account

the fact that the body produces more cortisol in the morning and has a drop of cortisol

throughout the day. The AUC was the chosen method of cortisol measurement for this

study because an average of the cortisol levels would not have been able to account for

the fact that the cortisol levels change in a stereotypical manner over the course of a day.

A mean AUC (AUCM) was calculated to generate one averaged measure for analysis for

each subject. Table 4.4 illustrated how the samples were identified. Table 4.5 illustrates

the calculation formula for the area under the curve.

Table 4.4 Cortisol identification per week

Week one Week one/ time oneCortisol levels from a.m. sample

Week one/ time twoCortisol levels from noon sample

Week one/ time threeCortisol levels from p.m. sample

Week two Week two/ time oneCortisol levels from a.m. sample

Week two/ time twoCortisol levels from noon sample

Week two/ time threeCortisol levels from p.m. sample

Week three Week three/ time oneCortisol levels from a.m. sample

Week three/ time twoCortisol levels from noon sampe

Week three/ time three Cortisol levels from p.m. sample

Table 4.5 Area Under the Curve calculation

AUC1-Week one AUC Week1time1+.5*(week1time1-week1time2)

AUC2-Week two AUC Week2time1+.5*(week2time1-week2time2)

AUC3-Week three AUC Week3time1+.5*(week3time1-week3time2)

AUCM-Mean of AUC AUC1+AUC2+AUC3/3

Table 4.6 further describes the distribution of the AUC measurements. Using the

area under the curve calculation to adjust for the changes in cortisol levles found in

87

maternal breast milk over the three week period of time, the results are as follows: In

week one there was a range of .30 to 1.09 with a mean of .65 and a SD of .20. Week two

there was a minimum cortisol level of .24 with a maximum level of 1.25 with a mean

of .59 and a SD of .27. In Week three there was a minimum cortisol level of .84 and a

maximum level of 1.08 with a Mean of .66 and a SD of .26

Table 4.6 Cortisol levels for weeks one, two and three

N Range Minimum Maximum Mean Std. DeviationAUC1 19 .79 .30 1.09 .6568 .20939AUC2 19 1.00 .24 1.25 .5981 .27808AUC3 19 .84 .24 1.08 .6658 .26161AUCM 19 .43 1.00 .6402 .17677

Table 4.6 also illustrates the results of the AUCM calculation. These calculations

are the results of an averaging of the cortisol enzyme immunoassay results. An average

of these AUC levels most accurately measures the amount of cortisol found in the

mothers’ breast milk taking into account the circadian rhythm. The mean found for the

area under the curve for the women in this study is .64 with a SD of .17.

88

Table 4.4 is an illustration of the distribution of mean cortisol levels identified by

subject. The distribution shows that the majority of the mothers had mean cortisol levels

below .65.

Table 4.4 Cortisol distribution identified by subject

0 10 20 30 40 50 60

subject

0.40

0.50

0.60

0.70

0.80

0.90

1.00

1.10

AU

CM

Y axis-Mean Scores for Area Under the Curve X axis-Identified by subject number

Data Analysis

Data collected in this study were analyzed by SPSS 12.0. Frequency counts,

means, and standard deviations were used to present the characteristics of the sample. A

Spearman coefficient correlation matrix and a linear regression were used to answer the

research question.

Spearman coefficient correlation matrix

89

Spearman correlation coefficients were used to examine the relationships among

the variables. The Spearman correlation was used to determine the results of this study

because of the nature of the data. Some of the data naturally occurred in rank form

(Rothbart scores) and other data had rank as a substitution for raw data (mothers income

level) (Howell, 2002). The analysis ranked the scores first and then calculated the

relationship differences among the ranks. The variables studied were: Positive

Affectivity Rothbart scores, Orientation/Regulation Rothbart score, Negative Affectivity

Rothbart scores, mother’s Trait Anxiety score, mother’s State Anxiety score, the cortisol

concentrations in breast milk, and maternal family income level (SES).

As seen in table 4.5, using the statistical standard that a p-value of less than .05

there was one significant relationship found. It was between the State and Trait Anxiety

scores with a correlation coefficient of .56 and a p-value of .012. There was the approach

of significance found with a correlation coefficient of .41 and a p-value of .075 between

Positive Affectivity and Negative Affectivity. There was a similarly strong relationship

found between Negative Affectivity Rothbart scores and Trait Anxiety scores with a

correlation coefficient of .40 with a p-value of .084.

90

Table 4.5 Spearman Coefficient Correlation Matrix

SES1 State trait AUCM poszscoreRegzscor

e negzscoreSES1 Correlation

Coefficient 1.000 -.114 -.226 -.010 -.123 .161 -.071

Sig. (2-tailed) . .642 .352 .969 .616 .509 .772N 19 19 19 19 19 19 19

state Correlation Coefficient -.114 1.000 .562(*) -.225 -.073 -.204 .120

Sig. (2-tailed) .642 . .012 .355 .767 .403 .624N 19 19 19 19 19 19 19

trait Correlation Coefficient -.226 .562(*) 1.000 -.138 -.366 -.396 .407

Sig. (2-tailed) .352 .012 . .573 .123 .093 .084N 19 19 19 19 19 19 19

AUCM Correlation Coefficient -.010 -.225 -.138 1.000 .098 .193 .253

Sig. (2-tailed) .969 .355 .573 . .689 .429 .297N 19 19 19 19 19 19 19

poszscore

Correlation Coefficient -.123 -.073 -.366 .098 1.000 .093 .418

Sig. (2-tailed) .616 .767 .123 .689 . .705 .075N 19 19 19 19 19 19 19

regzscore

Correlation Coefficient .161 -.204 -.396 .193 .093 1.000 -.219

Sig. (2-tailed) .509 .403 .093 .429 .705 . .367N 19 19 19 19 19 19 19

negzscore

Correlation Coefficient -.071 .120 .407 .253 .418 -.219 1.000

Sig. (2-tailed) .772 .624 .084 .297 .075 .367 .N 19 19 19 19 19 19 19

SES1 mother’s family income Poszscore positive Rothbart scoresState mother’s acute stress level Regzscore regulation/orientation Rothbart scoresTrait mother’s chronic stress level Negzscore negative Rothbart scoresAUCM Mean cortisol levels of area under the curve

Multiple linear regression

A multiple linear regression analysis was also used. The test was created by

programming SPSS 12.0 (2003) to determine the relationship of the predictor in relation

to the dependent variable. Infant behavior, identified by IBQ-R score, was used as the

dependent variable in this study. The variables that were added to the equation were the

mother’s trait anxiety score, maternal SES, and the cortisol concentrations in breast milk

91

in that order. Tables 4.6, 4.7, and 4.8 will illustrate the findings from the linear

regression analysis. The IBQ-R scores were determined by dividing the infants into three

age groups to control for the normal developmental differences found throughout infancy.

The age groups were divided as follows: age 1: 3-6 months, age 2: 6-9 months, and age 3:

9-12 months (Gartstien & Rothbart, 2003) and normalized, using z-transformations, from

the means and standard deviations of each age group as identified by Rothbart. The

mothers of the infants rated their children’s behavior based on specific questions asked

on the IBQ-R. The score sheets were returned to the researcher for tabulation. The

individual inventories were tabulated by adding the numbers on the IBQ-R scores sheets.

The score were then subdivided into 14 different domains. The domains were: activity

level, distress to limitation, fear, duration to orientation, smiling and laughter, high

pleasure intensity, low pleasure intensity, soothability, falling reactivity, cuddliness,

perceptual sensitivity, sadness, approach, vocal reactivity. Based on the recommendation

of Dr. Rothbart (per phone conversation, 2004) the tabulation results of the 14 domains

were divided again into three behavioral factor groups. The factor groups were identified

as Positive Affectivity, Negative Affectivity, and Orientation/Regulation. Three separate

linear regression analyze were conducted to allow for each of the factors to be used as a

dependent variable.

Table 4.6 describes the findings from the linear regression analysis which uses

Positive Affectivity as the dependent variable and SES, maternal trait anxiety, and

AUCM cortisol levels as independent variables.

92

Table 4.6 Positive IBQ-R ScoresIndependent variables: Area Under the Curve Mean Cortisol levels, Family income, Maternal Trait Anxiety levelDependent variable: Positive Affectivity Rothbart Scores

Model R R SquareAdjusted R

SquareStd. Error of the Estimate

1 .401(a) .161 -.007 1.18465a Predictors: (Constant), AUCM, SES1, trait

Model Sum of

Squares df Mean Square F Sig.1 Regressio

n 4.029 3 1.343 .957 .438(a)

Residual 21.051 15 1.403Total 25.080 18

a Predictors: (Constant), AUCM, SES1, traitb Dependent Variable: poszscore

Model

Unstandardized Coefficients

Standardized Coefficients

t Sig.B Std. Error Beta1 (Constant

) 1.226 2.227 .550 .590

trait -.055 .037 -.358 -1.467 .163SES1 -.237 .336 -.169 -.706 .491AUCM .668 1.611 .100 .415 .684

a Dependent Variable: poszscore

An RSquare of .16 indicates that 16% of the variability found in Positive

Affectivity Rothbart scores can be accounted for by AUCM, SES and Trait scores.

The significance levels of this linear regression equation indicates that there is no

relationship between Positive Affectivity Rothbart scores and the mothers Trait Anxiety

scores, mother’s SES and cortisol levels in breast milk collectivly. The F test for the

regression was not significant. This is made clear by an F-value significance level

of .43.

93

Table 4.7 describes the findings from the linear regression analysis which uses

Orientation/Regulation as the dependent variable and SES, maternal trait anxiety, and

AUCM cortisol levels as independent variables.

Table 4.7 Orientation/Regulation IBQ-R ScoresIndependent variables: Family income, Area Under the Curve Cortisol levels, Trait Anxiety levels.Dependent variable: Orientation/Regulation Rothbart Scores

Model R R SquareAdjusted R

SquareStd. Error of the Estimate

1 .444(a) .197 .036 .77233

a Predictors: (Constant), AUCM, SES1, trait

Model Sum of

Squares df Mean Square F Sig.1 Regressio

n 2.192 3 .731 1.225 .335(a)

Residual 8.947 15 .596Total 11.139 18

a Predictors: (Constant), AUCM, SES1, traitb Dependent Variable: regzscore

Model

Unstandardized Coefficients

Standardized Coefficients

t Sig.B Std. Error Beta1 (Constant

) .935 1.452 .644 .529

trait -.040 .024 -.395 -1.652 .119SES1 .034 .219 .037 .156 .878AUCM .553 1.050 .124 .527 .606

a Dependent Variable: regzscore

An RSquare of .20 indicates that 20% of the variability found in

Orientation/Regulation Rothbart scores can be accounted for by AUCM, SES and Trait

scores.

The significance levels of this linear regression equation indicates that there is no

relationship between Orientation/Regulation Rothbart scores and the mothers Trait

Anxiety scores, mother’s SES and cortisol levels in breast milk collectivly. The F test for

the regression was not significant. This is made clear by an F-value of .33

94

Table 4.8 describes the findings from the linear regression analysis which uses

Negative Affectivity as the dependent variable and SES, maternal trait anxiety, and

AUCM cortisol levels as independent variables.

Table 4.8 Negative IBQ-R ScoresIndependent variables: Family income, Area Under the Curve Cortisol levels, Trait Anxiety levels.Dependent variable: Negative Affectivity Rothbart Scores

Model R R SquareAdjusted R

SquareStd. Error of the Estimate

1 .506(a) .256 .108 1.09295

a Predictors: (Constant), AUCM, SES1, trait

Model Sum of

Squares df Mean Square F Sig.1 Regressio

n 6.174 3 2.058 1.723 .205(a)

Residual 17.918 15 1.195Total 24.092 18

a Predictors: (Constant), AUCM, SES1, traitb Dependent Variable: negzscore

Model

Unstandardized Coefficients

Standardized Coefficients

t Sig.B Std. Error Beta1 (Constant

) -3.704 2.055 -1.803 .092

trait .058 .034 .390 1.694 .111SES1 -.020 .310 -.014 -.064 .950AUCM 2.656 1.486 .406 1.787 .094

An RSquare of .26 indicates that 26% of the variability found in Negative

affectivity Rothbart scores can be accounted for by AUCM, SES and Trait scores.

The significance level of this linear regression equation indicates that there is no

relationship between Orientation/Regulation Rothbart scores and the mother’s Trait

Anxiety scores, mother’s SES and cortisol levels in breast milk collectivly. The F test for

the regression was not significant. This is made clear by an F-value of .20

95

Although the overall regression was not significance the significance tests for the

individual coefficients show a p-value of .11 for Trait Anxiety and a p-value of .09 for

AUCM.

Summary of Findings

The results of the analysis of variables determined that the women who participated

in this study had middle to high family income levels and most had a college education.

The mothers also scored low on the State and Trait Anxiety test compared to the

normative sample which suggests that the women who participated in this study exhibited

lower than average levels of acute and chronic stress. The infants who participated in this

study also tested differently than the normative sample. The infants were generally

happier than the normative sample. Cortisol levels in maternal breast milk were measured

but no comparison samples could be found due to the lack of research in the area.

The results of the correlation analysis which tested the strength of the relationships

between the study variables showed a significant relationship between State and Trait

Anxiety as well as a suggestion of a relationship between Trait Anxiety scores and

Negative Affectivity scores as well as between Positive Affectivity and Negative

Affectivity Rothbart scores.

A linear regression analysis was used to look at the significance of the relationships

between the independent variables in relation to the dependent variables. The analysis

results indicated that there was no significant relationship between any of the three

dependent and the same set of independent variables, although the independent variables

did account for 16-26% of the variance in Rothbart scores.

96

Chapter 5Discussion and Implications

Introduction

This chapter discusses the results of the study. In the first part of the chapter, a

summary of the findings is given. The second part of this chapter discusses the

correlation results. The third part is a discussion of the findings in relationship to

existing research. The fourth part addresses possible implications. Finally, the limitations

of this study and recommendations for further research are discussed.

Summary and Highlights of Findings

The purpose of this study was to determine whether the concentration of cortisol

in maternal milk has a direct impact on infant behavior independent of the mothers’ stress

levels. Specifically, this study was designed to answer the following research question:

Does the concentration of cortisol in mother’s breast milk have a relationship with the

behavior of the infant?

Hypothesis: Under conditions of normal, not chronic stress, high maternal stress levels led to high cortisol concentrations in breast milk which leads to a disruption of the emotional response circuitry of the infant which results in excessive emotional responsiveness of the infant leading to negative infant temperament characteristics.

97

In order to answer these questions several different data analyses were

conducted. First, the background variables of the participants were analyzed in order to

determine the demographic characteristics of the sample. Second, the data for the various

measures were analyzed in order to determine the stress levels of the mothers, the

behavioral patterns of the infants and the cortisol levels in the breast milk. Finally,

correlation and regression analyses were conducted to determine the strength and

significance of the relationships found between the behavior of infants in the study and

the levels of cortisol found in maternal breast milk as well as SES and maternal stress

levels.

Socioeconomic Status

The data from this study showed that the 19 women who participated in this study

had middle to high family income levels. This was seen by the fact that 78% of the

women had a family income between $11, 000.00 and $50,000.00. The women also had

a high level of education with 73% of the women having a college degree.

Maternal Stress

When the stress of the mothers was tested, the group of women who participated

in this study demonstrated low levels of stress on a tool that measured both acute (State)

and chronic stress (Trait) using the State Trait Anxiety Inventory (Consulting

Psychological Inc. 1977). The range of the tool is from 20 to 80. Lower numbers

represent low stress while high numbers represent high stress. The study participants had

a range of State anxiety scores from 20-47 while the normative sample had a range of

98

scores from 20-80. The range of Trait Anxiety scores for the mothers who participated in

the current study was from 23-57. The normative sample had a range of 21-80.

Infant Behavior

Infant behavior was assessed using the Infant Behavior Questionnaire-Revised

(IBQ-R) (Gartstien & Rothbart, 2003) and the findings of the study indicate that the

infants who participated in this study had higher levels of positive affectivity and lower

levels of negative affectivity compared to the normative sample used for this instrument.

For example, infants who participated in the current study had a higher average mean

score for Positive Affectivity by .46 points. The infants who participated in the current

study also had a higher average mean score for Orientation/Regulation by .43 points.

Compared to the infants who participated in the Rothbart study, the infants who

participated in this study had a lower average mean score for Negative Affectivity by

1.56 points.

Cortisol levels in Breast milk

The range of cortisol found in maternal breast milk over the three week period of

time was as follows. Week one .30 to 1.09 with a mean of .65 and a SD of .20. Week

two .24 to 1.25 with a mean of .59 and a SD of .27. Week three .84 to 1.08 with a mean

of .66 and a SD of .26. Due to a lack of research in this areas there are no norms to

compare these findings with.

99

Correlation Results

There was one significant relationship found using the Spearman correlation

coefficient among the pair wise correlations. A significant relationship was found

between the State and Trait Anxiety scores. There was also a suggestion of a relationship

found between Negative Affectivity Rothbart scores and Trait Anxiety scores. There was

a similar relationship found between Positive Affectivity and Negative Affectivity.

Linear regression results found no relationships. Three regression analyses were

conducted using each of three Rothbart factors as the dependant variables (Positive

Affectivity, Negative Affectivity, and Orientation/Regulation). Each regression used the

same independent variables, they were, maternal SES, State Anxiety scores, Trait

Anxiety scores, and breast milk cortisol levels. None of the regressions showed

significant results.

Discussion of Findings

General Findings

State and Trait Anxiety scores were the only two variables found to have a

significant relationship with the other. This is due to the nature of the stress test. Based

on the normative findings which illustrate a similar relationship between State and Trait

Anxiety scores it would be predicted that the State and Trait Anxiety scores would have a

relationship with each other (Consulting Psychological Inc. 1977). The lack of

relationship found between the other pairs of variables is possibly due to the low N also,

the sample does not lend itself to the establishment of statistical significance due to

100

homogeneity of the sample. Because the sample is small any relationship between

variables has to be very large to show significance. Also a sample of 19 may not

accurately reflect relationships which exist in the total population, irrespective of the

significance.

Socioeconomic Status

Studying the socioeconomic status (SES) of the women who participated in this

study was important based on the findings of research done by Lupien, King, Meaney,

and McEwen (2002). Their study found that there was a significant correlation between

socio-economic status and infant cortisol levels using a sample of 217 children. The

Carnegie Foundation added to the importance of this issue by studying SES and maternal

stress and found a direct correlation between maternal stress and SES as well as maternal

care qualities (Carnegie, 1994)

The lack of correlation between SES, maternal stress, and cortisol levels in this

study was perhaps due to the fact that the sample for this study was homogeneous and

was not necessarily representative of the population. The sample was largely made up of

women from middle to high income families. They also had high levels of education.

These two characteristics may have precluded a relationship between stress and SES. The

small sample size also potentially impacted the results of the correlation.

Maternal Stress

Essex, M.J., Klein, M.H., Cho, E., Kalin, N.H. (2002) identify the importance of

studying maternal stress in relation to infant cortisol levels by providing findings from

their longitudinal research on maternal stress and infant reactivity. They found that

101

maternal stress interferes with mother-infant interactions and that mothers who were

stressed were often found to lack sensitivity and responsiveness. They also found that

infants of stressed mothers showed higher levels of cortisol. Through these findings the

research team concluded that “It is likely that our sample of altered mother-infant

interactions secondary to mothers’ high levels of stress mediated the effects on children’s

later cortisol and behavior” (p. 38). Because attachment is a predictor of infant

temperament and maternal behavior is related to stress, it is important to look closely at

maternal stress levels related to maternal characteristics and infant temperament related

to maternal behavior as identified by researchers like Essex (2002) and Mangelsdorf

(2000).

The findings from this body of research suggest a relationship between State and

Trait Anxiety. This relationship indicates that mothers who are under acute stress are

also likely to be under chronic stress. There were no other significant relationships.

However, there may have been something going on between high Trait Anxiety scores

and negative infant behavior with a correlational coefficient of r = .40 and a p= .08. In

interpreting this data it could suggested that with a larger N there may be some

significance found between chronic stress in mothers and negative infant behavior. These

findings would support the work of Essex (2002) and Mangelsdorf (2000). On the other

hand the findings of no relationship may suggest that there truly is no relationship

between maternal stress and negative infant behavior and with a study using a larger N

the findings would not change indicating that infants are resilient to maternal stress.

Infant Behavior

102

The study of infant behavior has long been viewed in tandem with infant

attachment. Ainsworth began work in this area in the 1950’s and has generated volumes

of research findings that indicate that there is a correlation between infant behavior and

infant attachment. In recent years the study of infant behavior has changed to include the

influences of biology and environment. The concept of behavior has changed to include

regulation as a measure of infant behavior and temperament. Schore (2001) describes

infant regulation as the integration of both the biological and the psychological realms.

The connections between infant brain development and the significance of infant

regulation is also pointed out by Damasio (1994) who notes that the overall function of

the brain is to be well informed about what is going on in the rest of the body proper;

about what goes on in itself; and about what the environment is surrounding the

organism, so that suitable survivable accommodations can be achieved between the

organism and the environment. Through the work of researchers like Ainsworth (1996),

Shores (2001), and Damasio (1994) it is clear that infant behavior is an area of study that

requires a complex body of data that spans from behavioral tests to biological

assessments.

Although this study found no significant relationships between infant behavior

and maternal stress, the findings of this study support the thinking of people like

Ainsworth (1996), Shores (2001) and Damasio (1994) due to the suggestion of a

relationship between Trait (Chronic) stress in mothers and infant and Negative

Affectivity with a correlation coefficient of r = .40 and a p =.08. There are three possible

reasons for the lack of statistical significance found in this research. It may have been the

fact that the mothers who participated in this study had low levels of stress. It may have

103

been due to the low N of this study. It may also indicate that there is no relationship

between maternal stress and infant behavior.

Cortisol Levels

Mothers with higher cortisol levels were found to be more attentive in general

(Fleming et al. 1999). Fleming et al. looked further into mother infant behavior by

suggesting that there may be a correlation between breast feeding and maternal bonding.

Other studies that investigated the role of breast feeding and brain development were

conducted by researchers such as Mezzacappa , E.S., Tu, A.Y. & Myers, M.M.(2002);

Rolf & Keil (2002); Nelson & Panksepp (1998); Fleming A.S., Steiner, M., Cortner, C.

(1999), and Cook (1997). Their studies have thoroughly examined the sociological,

nutritional, and biological impacts of breast feeding on human and animal offspring.

These studies have found that cortisol exists in breast milk and is presumed to be passed

to the infant during feedings. However there are no studies that have examined cortisol

levels in breast milk and maternal stress.

The results of this research study did not find a significant relationship between

cortisol levels in maternal breast milk and infant behavior. However, a p-value of .09 for

AUCM and Negative Affectivity Rothbart scores were found within the regression

analysis. The significance levels indicate that although there is no relationship overall

there may be a relationship between these variables. This lack of significance may have

been due to the small sample size.

These findings may also suggest that there is no relationship between maternal

stress and cortisol levels in breast milk, or that the mothers who participated in this study

had low levels of stress which led no significance relationship.

104

Overall findings

As discussed in Chapter 4 the results of the analysis of variables determined that

the women who participated in this study had middle to high family income levels and

most had a college education. The mothers scored low on the State and Trait Anxiety test

compared to the normative sample which suggests that the women who participated in

this study exhibited lower than average levels of acute and chronic stress. The infants

who participated in this study also appeared happier than the normative sample. Many of

these findings differ from the pre-established assumptions about this research area. Table

5.1 describes the assumptions and the research findings.

Table 5.1 Comparative sample of current findings and research assumptions

Research topics

Authors Assumptions Current findings

Maternal Stress

Essex, M.J., Klein, M.H., Cho, E., Kalin, N.H. (2002)

Maternal stress interferes with mother-infant interactions and that mothers who were stressed were often found to lack sensitivity and responsiveness. They also found that infants of stressed mothers showed higher levels of cortisol.

There may be a threshold of maternal stress below which no relationship with infant temperament exists which would indicate a quadratic relationship

The mothers scored low on the State and Trait Anxiety test compared to the normative sample which suggests that the women who participated in this study exhibited lower than average levels of acute and chronic stress.

Cortisol Fleming et al. Breast milk is found

to have cortisol, Cortisol was found in breast milk. Values

105

levels in Breast milk

Grangeroxytosin, and endogenous opiods.

There may be a threshold of cortisol in breast milk below which no relationship with infant temperament exists which would indicate a quadratic relationship.

from matched serum and saliva samples show the expected strong linear relationship.

Infant behavior

Putnam,Rothbart,

Ainsworth

Gartstien

Infant behavior can be assessed based on three factors. Negative Affectivity, Positive Affectivity, and Orientation/Regulation. Infants in the normative sample had higher negative scores and lower positive scores. These scores were derived from a large sample which is representative of the population.

The infants who participated in this study tested differently than the normative sample. The infants were generally happier than the normative sample.

SES and the impact on parent competence

Carnegie Foundation

Gunnar & Donzella

Lupien et al.

There is a correlation between SES and incidence of maltreatment on the part of the parent.

Mothers who participated in this study had high SES and were well educated.

Possible Implications

In determining whether the concentration of cortisol in maternal milk has a direct

impact on infant behavior independent of the mother’s stress levels, (and thus behavioral

106

interactions with the infant) many factors had to be taken into consideration. They were

SES, maternal stress, infant behavior, and cortisol levels in breast milk. Fleming et al.,

(1997) has identified that there is a relationship between maternal attentiveness and an

increase in cortisol levels in mothers, Gunnar et al.(2001) found that there is a

relationship between cortisol levels in mothers and the quality of early care an infant

experiences. Mezzacappa et al. (2002) identified a relationship between breast fed infants

and lower baseline cortisol levels. However there are still unanswered questions about

how maternal stress affects infant behavior and whether there is a relationship between

cortisol, breast milk, maternal stress, and infant behavior.

The lack of relationships found in the data from the current study indicates a

contrary finding to those of Flemming, Gunnar, Mezzacappa. The data from this study

also suggests with a more varied sample there may be a relationship between mother’s

Trait Anxiety and Negative infant behavior. This relationship would suggest that

mothers’ who are feeling stress over a long period of time and have generalized stress

may impact their infants’ behavior in a negative manner. With an increase N there may

also be a relationship between cortisol levels in breast milk and negative infant behavior.

Limitations and Recommendations

Limitations The volunteer participation of the mother/infant pairs created a non random

sample which limited the ability to generalize the results of the study.

107

The small N (N=19) was a significant limitation of this study. The small number

of subjects impacted the analysis by limiting the ability to appropriately run a regression

of the data as well as limiting the ability to identify the strength and direction of the data

when analyzed for correlational significance. A small demographic area was used to

draw the sample which led to a homogeneous group with similar SES, stress and infant

behavior data. Using a larger sample from a diverse demographic area would have

provided more variability in the sample and would have allowed more generalization of

the study results and to an increase in statistical power, given a greater range of scores.

There may be a threshold of maternal stress below which no relationship with

infant temperament exists which would indicate a quadratic relationship

There is limited information about cortisol concentration levels in breast milk.

Because of this it was impossible to compare the cortisol concentration levels of the

women in this study with other groups of mothers. This again led to an inability to

generalize the results of this study. However, this Study is the beginning of important

work in the areas of cortisol, breast milk, maternal stress, and SES.

Recommendations

This study was designed to explore the relationship between cortisol

concentration levels in maternal breast milk and infant behavior. It was an attempt to

focus on the importance of the biological influences of maternal stress on infant behavior

using cortisol as a measure. Hopefully, more research studies will focus on the topic in

the future. By making continuous efforts to examine the influence of mother’s stress and

the impact of biological factors on infant behavior, we can increase our knowledge base

108

related to maternal and infant mental health. This will provide for a clearer picture of

what is needed in order to support breast feeding mothers and their infants.

This study examined the influence of a limited set of selected variable related to

mothers and infants. Further research could extend the investigation by increasing the

variables used to study this topic (e.g., infant salivary cortisol levels, maternal salivary

cortisol levels, measure of maternal attitude toward her infant, and prenatal and postnatal

history). The use of these variables in further research will increase our understanding of

the roles of maternal stress, chemical reactivity and infant response.

Since this study has identified that there are some possible significant

relationships which were related to maternal stress, infant behavior, and cortisol levels in

breast milk (e.g., State and Trait Anxiety, Negative Affectivity, and AUC cortisol),

further research studies may gather information about what affects these variables

(environment, biology). This could be done by conducting appropriate naturalistic

studies in which these variables are studied independently using a larger sample.

Finally, large-scale studies and longitudinal studies on maternal stress, infant

behavior and the role of cortisol are recommended. Since the role of cortisol on the

biological function of the body is a fairly new domain of research, and the role of

maternal mental health and infant behavior are emerging fields of research it would be

appropriate to continue studying these issues in tandem using large sample sizes. The

large sample would provide for more generalizable results which in turn would be useful

to people who are breast feeding.

Because one of the study factors for this research project, (cortisol) is most often

studied in the biological fields, maternal mental health (maternal stress) is most often

109

studied in the social science fields it is important to continue to do research that combines

the two issues to find results that support mothers and children. Studying these issues

over a longer period of time would provide data about cortisol concentration in breast

milk over time, maternal stress related to adjusting to parenting, infant behavior over time

and the role of bonding on cortisol levels and infant behavior, as well as the course of the

development of temperament into adulthood.

REFERENCE

Ainsworth, M.D. & Bowlby, J. (1991) An ethological approach to personality

development. American Psychology, 46, 333-341.

Bowlby, J. (1969) Attachment and loss vol. 1 attachment. New York: Basic.

Brazelton, T.B (1973) Neonatal behavior assessment scale. Clinical Development Med,

50.

Carnegie Foundation (1995). Starting Points, Carnigie Foundation, N.Y.

Caspi, A. (2002). Role of Genotype in the Cycle of Violence in Maltreated Children.

Science, 297 (2), 851-854.

Cassidy, J. B. L. J. (1994). The insecure/ambivalent pattern of attachment: theory and

110

research. Child Development, 65 (4), 971-991.

Consulting Psychologist Press, Inc (1977) Self-evaluation questionnaire. Palo Alto:CA.

Cook, C. J. (1997). Oxytocin and prolactin suppress cortisol responses to acute stress in

both lactating and non-lactating sheep. Journal of Dairy Research, 64 327-339.

Damasio, A.R. (1994). Descarte’s Error. New York:Grosset:Putnam.

Davis, M., Hitchcock, J.M. & Rosen, J.B. (1987) Anxiety and the amygdala:

pharmacological and anatomical analysis of the fear potentiated startle

paradigm. The Psychology of Learning and Motivation, 21 263-305.

deWeerth, C. & van Geert. (2002) A longitudinal study of basil cortisol in infants: intra-

individual variability, circadian rhythm and developmental trends. Infant

Behavior and Development, 152, 1-24.

Diamond, M., Hopson, J. (1988) Magic trees of the mind. New York: Penguin.

Dubowitz, L. M., Cowan, F., Rutherford, M. Mercuri, E., Pennock, J. (1995). Neonatal

neurology, past present and future: A window on the brain. Brain and

Development, 17, 22-30.

111

Eliner-Avishai, S., Brunson, K. L., Sandman, C. A. & Baram, T.Z. (2002). Stressed-

out,or In (Utero)? Trends in Neurosciences, 25(10), 518-524.

Essex, M.J., Klein, M. H., Cho, E., Kalin, N.H. (2002) Maternal stress beginning in

infancy may sensitize children to later stress exposure: effects on cortisol and

behavior. Biological Psychiatry 52, 776-784.

Fleming, A. S., O'Day, D.H., Kraemer, G.W. (1999). Neurobiology of mother-infant

interactions: experience and central nervous system plasticity across

developmental and generations. Neuroscience and Behavioral Reviews, 23, 673-

685.

Fleming, A. S., Steiner, M., Cortner, C. (1997). Cortisol, hedonics, and maternal

responsiveness in human mothers. Hormones and Behavior, 32, 85-98.

Fox, N., & Henderson, H.A. (1999). Does Infancy Matter? Predicting social behavior

from infant temperament. Infant Behavior and Development, 22 (4), 445-455.

Gartsein, M. A., Rothbart, M.K. (2003). Studying infant temperament via the revised

infant behavior questionnaire. Infant Behavior and Development, 166, 1-23.

Granger, D. (2004) Salimetrics Laboratory. State College, PA

112

Graham, F.K. (1956). Behavioral differences between normal and traumatized

newborns. Psychol Monographs: General and Applied, 70.

Granham, Y.P. & Heim, C. (1999). The effects of neonatal stress on brain

development: implications for psychopahthology. Developmental

Psychopathology, 11, 545-565.

Gray, P. (1999). Psychology.Worth Publishers, NY.

Gunnar, M. R., Donzella, B. (2002). Social regulation of the cortisol levels in early

human development. Psychoneuroendocrinology, 27, 199-220.

Gunnar, M. R. B., R.G. (1998). Stress, early brain development, and behavior. Infants

and Young Children, 11(1), 1-14.

Gunnar, M.R., Vazquez, D.M. (2001). Low cortisol and a flattening of expected

day time rhythm: potential indices of risk in human development. Development

and Psychology, 13, 515-538.

Gunner, M.R., Brodersen, L., Nachmias, M., Buss, K., Rigatuso, J., (1996). Stress

reactivity and attachment security. Developmental Psychobiology, 29 (3), 191-

204.

113

Hertsgaard, L. G., M.R., Erickson, M.F., Nachmias, M. (1995). Adrenocortocal

responses to the strange situation in infants with disorganized/disoriented

attachment relationships. Child Development, 6 (4), 1100-1106.

Howell, D. (2002). Statistical methods for psychology. Pacific Grove CA: Duxbury

Press.

Johnson, M. (2000). Functional brain development in infants: elements of an interactive

specialization framework. Child Development, 71(1), 75-81.

Johnston, M. V. (1995). Neurotransmitters and vulnerability of the developing brain.

Brain and Development, 1995(17), 301-306.

Kagan, J. (1996). Three pleasing ideas. American Psychologist, 51 901-908.

Kagan, J. (1994). Galen’s prophecy: temperament in human nature. New York:

Basic Books.

Koletzko, B., Rodriguez-Palmero, M., Demmelmair, H., Fidler, N., Jensen, R., &

Sauerwald, T. (2001). Physiological aspects of human milk lipids. Early Human

Development, 66(Supplement), 3-18.

114

LeDoux J. (1989). Cognitive-emotional interactions in the brain. Cognition and

Emotion,3, 267-289.

LeDoux, J. (2002). The synaptic self, Viking Press, N.Y.

Lewis M. & Feiring, C., & Rosenthal, S. (2000). Attachment over time. Child

Development, 71(3), 707-720.

Levine, S. (1994). The ontogeny of the hypothalamic-pituitary-adrenal-axis: the influence

of maternal factors. New York Academy of Science, 4 275-288.

Lupien, S.J., King, S.,Meaney, M.J., McEwen, B.S. (2002). Child’s stress hormone levels

correlated mother’s socioeconomic status and depressive state.

BiologicalPsychiatry. 48, 976-980.

Mangelsdorf, S. C., McHale, J.L., Diener, M.,Goldstein, L.H., & Lehn, L. (2000). Infant

attachment: contributions of infant temperament and maternal characteristics.

Infant Behavior and Development, 23, 175-196.

Matthews, S.G. (2001). Antenatal glucocorticoids and the developing brain:

Mechanisms of Action. Seminars in Neuroatology, 64 (4), 309-317.

Matthews, S.G. (2002). Early programming of the hypothalamo-pituitary-adrenal axis.

115

Trends in Endocrinology & Metabolism.

McArthur, S., Shah, A.N., Morris, J.F., Flower, R.J., Buckingham, J.C., Gillies,

G.E.(2003). Programming effects of neonatal dexamethasone treatment on the

ventral tegmental area in adulthood: the potential role of annexin 1. program no.

149.4. 2003 Abstract Viewer/Itinerary Planner. Washington, DC: Society for

Neuroscience.

Meaney, M. J., Diorio, J.,Francis, D., et al. (1996). Early environmental regulation of

forebrain glucocrodicoid receptor gene expression: implications for adrenocortical

responses to stress. Developmental Neuroscience, 18, 49-72.

Mezzacappa, E.S., Tu, A.Y, Myers, M.M. (2002). Lactation and weaning effects on

physiological and behavioral response to stress. Psychology and Behavior, In

press.

Neil Gordon (1997). Nutrition and cognitive function. Brain and Development19, 165-

170.

Nelson, E. E., & Panksepp, J. (1998). Brain substrates of infants-mother attachment:

contributions of opioids, and norepinephrine. Neuroscience and Behavioral

Reviews, 2(3), 437-452.

116

Monroe, D. (1984). Adapted from enzyme immunoassay. Analytical Chemistry, 56,

(8),165-170.

Moore, D. (2000). The basic practice of statistics. W. H. Freedman Company: New York.

Nelson, E. E., & Panksepp, J. (1998). Brain substrates of infants-mother attachment:

contributions of opioids, and norepinephrine. Neuroscience and Behavioral

Reviews, 2(3), 437-452.

Pally, R. (1997). How brain development is shaped by genetics and environmental

factors. International Journal of Psychoanalysis, 78, 587-593.

Panksepp, J. (1998). Affective neuroscience: the foundation of human and animal

emotion. New York: Oxford University Press.

Paterson, R. & Moran, G. (1988). Attachment theory personality development and

psychotherapy. Clinical Psychology Review, 8 611-638.

Posner, M. I. (2001). Developing brains: the work of the sackler institute. Clinical

Neuroscience Research, 258-266.

Putnam, S.P., Ellis, L.K. & Rothbart, M.K. (2001). The structure of temperament from

infancy through adolescence. In A. Eliasz & A. Angleitner (Eds.) Advances in

Research on Temperament 165-182. Germany: Pabst Science.

117

Prechtl, H.F. (1977). The neurological examination of the full-term newborn infant: a

manual for clinical use. Clinical Development Med, 63.

Rolf, B. & Keil, W. (2002). Investigation of oral swabs taken from newborn infants after

breastfeeding. Legal Medicine, 4 52-54.

Rothbart, M.K. (2000). Infant behavior questionnaire-revised.

Rothbart, Calkin, Gunnar, Kalin, Panksepp, & Reiman. (2001). Emotion and

temperament. Developmental Science, 4, 313-329.

Rothbart, M. K., & Hwang, J. (2002). Measuring infant temperament. Infant Behavior

and Development, 25, 113-116.

Schore, A. N. (2001). Contributions from the decade of the brain to infant mental health:

an overview. Infant Mental Health Journal, 22(1-2).

Schore, A.N. (2001). Effects of a secure attachment relationship on brain

development, affect regulation and infant mental health. Infant Mental

Health Journal, 22, 345-375.

Schwartz, E. B. G., Douglas A., Susman, Elizabeth J., Gunnar, Meggan R. ,

Laird, Brandi.(1998). Assessing salivary cortisol in studies of child development.

118

Child Development, 69(6), 1503-1513.

Seeman,T.(2000).SalivaryCortisol.Http://www.macses.ucsf.edu/Research/Allostatic/note

book/salivarycort.html.

Siegel, D. (1999). The developing mind: toward a neurobiology of interpersonal

experience. New York: The Guilford Press.

SPSS 12.0 (2003). Minnesota: Rochester

Stewart, J. S., T. (1999). Salivary cortisol measures (Chapter 2): Rockefeller

University.

Suomi, S.J. (1991). Early stress and adult emotional reactivity in rhesus monkeys. paper

presented at the childhood environment and adult disease: Ciba

FoundationSymposium.

Thomas, A & Dargassies, S.A. (1952). Etudes neurologique du tonus masculaire.

Paris:Flammarion.

Worobey, J. (1989). Temperament rating at 2 weeks, 2 months, and 1 year differential

stability and activity and emotionality. Developmental Psychology (8). 257-263.

119

120

APPENDIX A

Parent Consent Form

121

122

123

124

APPENDIX B

Parent Information

125

126

127

APPENDIX C

State/Trait Anxiety Inventory

128

129

130

APPENDIX D

Rothbart Infant Temperament Inventory

131

132

133

134

135

136

137

138

139

140

141

APPENDIX E

Socioeconomic Inventory

142

143

144