Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
EMG BIOFEEDBACK TRAINING: EFFECT ON BEHAVIOR OF
CHILDREN WITH ACTIVITY-LEVEL PROBLEMS
THESIS
Presented to the Graduate Council of the
North Texas State University in Partial
Fulfillment of the Requirements
For the Degree of
MASTER OF SCIENCE
By
David L. Henry, B. A,
Denton, Texas
May, 1977
.0-%
:6-7 1
4 ON 44
Henry, David L., EMG Biofeedback Training: Effect on
Behavior of Children with Activity-Level Problems. Master
of Science (Clinical Psychology), May, 1977, 66 pp., 3 tables,
5 figures, references, 52 titles.
The relationships between muscle-tension level, motoric-
activity level, and academic performance in the laboratory
setting are investigated. Three participants were rein-
forced for reducing and increasing their tension levels,
alternately, while engaged in a simulated academic task, and
the effects of each on the rate of acitivty and academic
performance were measured. Measures were also obtained on
the rate of activity and occurrence of problem behavior in
the subject's homes. Significant treatment differences
were found which support a direct relationship between
tension and activity level so that a decrease in EMG level
was associated with a decrease in motoric activity, and an
increase in EMG level was associated with an increase in
motoric activity. The efficacy of using EMG biofeedback
to train relaxation in children with activity-level problems
to control their symptoms is supported, especially where
such a technique can be used in a specific task-oriented
situation.
TABLE OF CONTENTS
Page
LIST OF TABLES
Thesis
Introduction
Method .
SubjectApparatusProcedure
Results . . .
Subject 1Subject 2Subject 3
Discussion
Subject 1Subject 2Subject 3Summary
Appendix
References
iv
.1
. . . . . . . 9 . . . . . . . 9 . . . 12
. . . . . , . . 21
. ... 29
50
........ 60
Iii
. .0 . 0. 0. 0. 0. 0. . . . .0 .0 .0 .0 .0 .0 .a .0
LIST O TABLES
Table Page
1. Time-Series Analysis for Subject 1 . . . 41
2. Time-Series Analysis Subject 2 . . 44
3. Time-Series Analysis for Subject 3 . . . . . 47
iv
EMG BIOFEEDBACK TRAINING: EFFECT ON BEHAVIOR OF
CHILDREN WITH ACTIVITY-LEVEL PROBLEMS
Hyperactivity is a clinical condition characterized by
excessive movement, unpredictable behavior, unawareness of
consequences, inability to focus or concentrate on a
particular task, and poor academic performance (Stewart,
Pitts, Craig, & Dieruf, 1966). It has been estimated by
researchers and teachers that one in 25 children exhibit
symptoms of hyperactivity (Sroufe & Stewart, 1973).
Burks (1960) states that the hyperactive syndrome has
an organic basis and he lists seven primary symptoms seen
in most hyperactive children. These are a short attention
span, restlessness and overactivity, poor perceptual and
conceptual abilities, defective memory, and poor muscular
coordination. In addition, such children often show many
secondary or emotional symptoms, and when older may display
such antisocial tendencies as lying, stealing, vandalism,
fire setting, and cruelty to animals. Burks presents
evidence from several areas of investigation which support
an organic view of hyperactivity. He found that from 50% to
75% of children so diagnosed produce abnormal electro-
encephalograph (EEG) tracings. He postulates that these
children have a cortical impairment, and that those who
produce normal brain wave patterns may have a subcottical
1
2
defect which the EEG cannot reflect. His research showed
that hyperactive children fuse at lower frequencies than
do normals on the Flicker Fusion Test and have difficulty
seeing the aftereffect on the Archimedes Spiral, two tests
purported to measure cerebral efficiency and adaptability.
In checking the developmental histories of his sample,
Burks found that some type of birth trauma which could
cause brain impairment had occurred much more frequently
than in the normal population.
Stewart et al. (1966) and Denhoff, Davids, and Hawkins
(1971) note there is a high incidence of hyperactivity in
children with no history of disease or birth trauma.
Morrison and Stewart (1973) suggest that minimal brain
dysfunction is the cause of hyperactivity, and because
the syndrome is six times more prevalent among males, propose
that such a dysfunction is transmitted genetically.
The most commonly recommended treatment for hyper-
activity is the use of stimulant drugs and it is estimated
that about 200,000 children in the United States are currently
receiving amphetamines to control their overactivity
(Krippner, Silverman, Cavallo, & Healy, 1973).. Bradley (1937)
was among the first to report the use of drugs in treating
children with an "organic behavior syndrome." He found
by administering benzedrine that about 50% of the children
showed dramatic improvement in both school performance and
emotional response. The remainder of the subjects showed
little or no improvement and a small percentage showed
3
increased hyperactivity and aggressiveness. Conners and
Eisenberg (1963) and Denhoff et al. (1971) reported similar
results using methylphenidate (Ritalin) and dextroampheta-
mine, respectively. Using global ratings by parents,
teachers, and clinicians, Conners (1971) has shown that
stimulant drugs are effective with some children in
reducing overactivity and increasing attention span both
at home and at school, with a resultant improvement in
school achievement and social behavior. Comly (1971)
reports that teachers' ratings of children receiving stimu-
lant drugs indicate that improvements occurred in the areas
of listening ability, excitability, forgetfulness, and peer
relationships. Using actometer devices which are a direct
measure of motoric activity, several authors have shown
that stimulant drugs control excess activity in the
laboratory and applied settings (Hollis & St. Omer, 1972;
Sprague, Barnes, & Werry, 1970; Sykes, Douglas, Weiss, &
Minde, 1971L).
Stimulant drugs appear to be less effective in the
decrease of irritability, explosiveness, and other aspects
of emotionality (Weiss, Werry, Minde , Douglas, & Sykes, 1968;
Denhoff et al., 1971). Werry, Weiss, Douglas, and Martin
(1966) found that chlorpromazine decreased overactivity but
did not improve distractability,aggressiveness, or excita-
bility. Zrull (1963) found that children treated with
stimulants tended to become more fearful as their symptoms
improved, and postulated that the symptoms may be releases
4
for anxiety, whether the basis for the anxiety is psycho-
logical, physiological, or both. He noted that central
nervous system stimulants may be effective in treating a
child's symptoms at one time but not at other times, and
often have very different effects oxr children with identical
diagnoses. There is a growing ethical concern for the .
morality and wisdom implied in administering medication to
children (Fish, 1971; Hentoff, 1L970; Koegh, 1971; Ladd, 1970).
Because the often-implied objective behind the use of
drugs for the hyperactive child is that of enabling him to
profit academically, it is surprising that few data directly
support this belief. Most studies have measured the effect
on medication on component skills or learning, for example,
attention, concentration, and discrimination. Drug effects
on general intelligence test performance have been tested
(Conners & Rothschild, 1968; Epstein, Lasagna, Conners, &
Rodriquez, 1968; Knights & Hinton, 1968). Sprague et al.
(1970) measured childrens' responses of "same" or "different"
to pairs of visual stimuli presented on a screen. Conners,
Eisenberg, and Sharpe (1964) studied the effects of Ritalin
on paired-associate learning and Porteus Maze performance
in children with hyperactive symptoms. Others concentrated
their efforts on the effects of drugs on the attention of
hyperactive children to various tasks (Conners, Eisenberg, &
Barcai, 1967; Sprague & Toppe, 1966). These laboratory
studies investigated the effects of drugs on component
5
skills related to learning, but they did not measure academic
performance in the classroom.
Sulzbacher (1972) analyzed the effects of drugs on
behavior of hyperactive children in the classroom. Measures
of correct solutions and error rates were taken in arith-
metic, writing, and reading on three subjects. In addition,
measures were taken on the childrens' rates of talk outs and
out-of-seat behavior. His conclusion was that stimulants
"can effectively modify disruptive behavior without adversely
effecting academic performance in the classroom." Drug
effects on actual academic tasks were highly variable,
however.
Teachniques employing principles of behavior modification
have also been successfully used to control symptoms of
hyperactivity. Patterson (1965) used candy and pennies
to reinforce the on-task behavior of a hyperactive child in
a classroom setting. The child was reinforced at 10-second
intervals if attending to his assigned task. Prior to
conditioning, the boy spent 25% of his time making inatten-
tive or disruptive responses. After 10 days of training,
inappropriate behavior occurred only 5% of the time.
Several authors (Kinsely, 1970; Patterson, Jones,
Whittier, & Wright, 1965; Pihl, 1967; Twardosz & Sajivaj, 1972)
have reinforced behavior incompatible with hyperactive
responses, such as sitting still, and found that hyperactivity
and distractibility were decreased.
6
Ayllon, Layman, and Kandel (1975) reported a study which
compares the effects of stimulant drugs to behavior modifi-
cation techniques on classroom behavior and academic
performance. While on medication the subjects were observed
in a classroom for 18 days to establish a baseline for
correct math and reading responses, which occurred at a
negligible level for all subjects, and for occurrence of
hyperactive responses, which occurred nearly 20% of the time.
Medication was then withdrawn over a weekend and the rate of
hyperactive responses increased to 80%. A slight improve-
ment was found in math and reading responses. The class
time was divided so that half was spent on mathematics and
half on reading. The children were then reinforced for
correct mathematics responses while reading went unreinforced.
Hyperactive responses were reduced to 20% during mathematics
and mathematics performance improved from 12% to 85%
correct responses. Hyperactive responses remained at 80%
during reading and reading performance showed no improve-
ment. Correct reading responses were then reinforced and
similar improvement was obtained. Behavioral methods were
as effective as drugs in controlling hyperactivity in this
population with the additional effect of improved academic
performance. The problem of obtaining a transfer of effects
from the training situation to other areas was very apparent.
Investigators have sought to overcome the problem of
generalization by teaching parents to be behavior modifiers.
Ferman dnd Feighner (1973) videotaped parental interactions
7
so that parents could be aware of chronic maladaptive
behaviors which they reinforced in their children. They
were then taught to modify their own and their child's
behavior with encouraging results. Wiltz and Gordon (1974)
placed the entire family in an experimental apartment
equipped with one-way mirrors, video tape equipment, and
hidden microphones. The family lived in the setting for
five days, were observed carefully, and suggestions were
made for correcting maladaptive behavior. Hyperactive,
destructive, noncompliant, and other inapprorpiate responses
made by the child were quickly extinguished with this method.
Once back in their home, the parents continued the procedures
and remained in contact with the authors by telephone.
Appropriate behavior generalized to the home setting and
the parents continued to make progress in changing other
behavior in their children.
Effective behavior modification techniques are costly
in terms of equipment and manpower, On the other hand,
with techniques such as Jacobson's (1938) progressive
relaxation exercises or conditioning of relaxation with
biofeedback equipment, it would be possible to teach the
child to control his own impulses and behavior. If as
Zrull (1968) suggested, anxiety underlies hyperactive
syiptomatology, or if hyperactive children are overly tense,
as Lupin, Braud, Braud, and Duer (1974) suggested, such
techniques for desensitizing the tension or anxiety would
be beneficial. Wolpe (1958) postulated that anxiety is
8
incompatible with relaxation. Therefore, if a child can
be taught to relax in anxiety-producing situations such as
the classroom, the tension or anxiety would subside and the
symptoms would be alleviated. Lupin et al. (1974) suggested
that medication can reduce overactivity without reducing
the tension which would explain why hyperactive children
on medication remain excitable, irritable, explosive,
aggressive, and have a low tolerance of frustration.
Lupin et al. (1974) reported a study using the relaxation
exercises developed by Jacobson (1938) in the treatment of
13 hyperactive boys and their parents. Their treatment
program was carried out in the subjects' homes with the aid
of tape-recorded instructions. There were six tapes for
the children and six tapes for the parents. The parents'
tapes consisted of instructions in behavior modification,
instructions on teaching the child to do the relaxation
exercises, adult exercises, and three tapes using visual
imagery to enhance the feeling of relaxation. The visual
imagery tapes used desensitization techniques to enable the
parents to respond to their children with less anxiety. The
childrens' tapes consisted of one tape describing the relaxa-
tion exercises, four tapes which incorporated visual imagery
and sound effects, and one tape on the value of positive
attitudes. Both parents and children were instructed to
use the tapes daily for a 3-month period. At the end of
the treatment, a significant improvement was found between
pre and postbehavioral ratings done by parents. Teachers'
9
observations in a classroom setting revealed that inappro-
priate behaviors occurred much less frequently after
treatment, and scores on three Wechsler Intelligence Scale
for Children (WISC) subtests showed significant improvement.
However, children who used the tapes most frequently showed
the most improvement on parent ratings but not on other
measures, which suggests a placebo effect.
Biofeedback is a recently developed technique which
has been used successfully in treating tension-related
problems of adults, such as tension and migraine headaches
(Blanchard & Young, 1974; Budzynski, 1973; Sargent, Green, &
Walter, 1973). Through the use of electronic devices, the
subjects' physiological processes can be made known to
them via some external stimulus. This allows them to attend
to their internal states and ultimately bring them under
voluntary control. The biofeedback process incorporates
two major principles of operant conditioning--the immediate
knowledge of results and the gradual shaping of responses
(Budzynski & Stoyva, 1969). Essentially, it is a technique
of behavior modification going on under the skin. Biofeed-
back training has been useful in training adults to gain
control over a variety of physiological activities such as
heartrate, blood pressure, muscle tension, and brain wave
patterns.
Biofeedback is currently being explored as a treatment
for hyperactivity. Nall (1973) used alpha-brain-wave-feed-
back training with hyperactive children with limited success.
10
However, total training time was short (6 hours) and no
attempt was made to facilitate generalization from the
training situation to the child's normal environment.
Braud, Lupin, and Braud (1974) reported a pilot study
done with a hyperactive boy using electromyograph (EMG)
biofeedback to reduce his tension, Using the frontalis
muscle as an index of muscular tension, the EMG monitoring
system would sound a tone when tension levels occurred above
a predetermined threshold. The subject was simply told to
sit still and keep the tone off. He was given 24 minutes
of training twice weekly for 3 weeks, and then once weekly
for 5 weeks. Muscle tension declined greatly both within
and between sessions, and many psychosomatic symptoms were
eliminated. The subject was not encouraged to practice the
technique at home or elsewhere, so the improvements in his
behavior did not last. He was,,however, able to reduce his
tension level effectively and immediately after a 7-month
period.
Connoly,, Besserman, and Kirschvink (1974) used EMG
biofeedback to reduce the overactivity of six children in
an uncontrolled pilot study. All but one of the subjects were
taking Ritalin to control activity levels. Each subject
received two biofeedback-training sessions a week for 4 con-
secutive weeks. During these training sessions, the children
were reinforced for reductions in frontalis muscle tension
below an established level. Significant improvement was seen
on two subtests of the WISC and on parent and teacher ratings.
11
Braud (1975) conducted a controlled study comparing
normal children with hyperactive children in three treat-
ment groups. One group received home relaxation training
via taped exercises, one group received home training plus
EMG biofeedback, and one group received no treatment. Both
treatment groups were found to do better than the nontreat-
ment group on several measures, but no difference was
reported between the relaxation-only and relaxation-plus-
biofeedback groups. It was noted that EMG biofeedback
produced quicker and greater drops in tension levels, at
least in the forehead region.
There is a need for a treatment that will control
activity-level problems without impairing academic perfor-
mance. Allyon et al. (1975) suggest that the continued use
of drugs to control hyperactivity may result in students who
are compliant, yet academically incompetant. Behavioral
procedures can be used to reduce overactivity and improve
academic performance, but they are costly in terms of man-
power and equipment, and it is difficult to generalize the
effects from the specific settings in which they are trained
to other environmental situations. Relaxation training
and EMG-biofeedback training offer the possibility of a
treatment that will not only control hyperactivity without
impairing academic performance, but will also be more
easily generalizable to the total environment. Once control
over tension level is mastered, the child can be taught that
he can control his tension in any situation by applying the
12
skills he has learned. As in the drug studies, the effects
of reduced tension levels on global behavior ratings and
the component skills of learning have been measured, but
no studies were found which measured the effects on academic
performance.
This study was designed to determine if it would be
feasible to use EMG biofeedback in conjunction with reinforce-
ment techniques to train muscle relaxation and tension,
alternatively, in children With activity-level problems
while they engaged in an academic task. Time spent on task,
motoric activity, and mathematics effort were monitored
to assess the effects of reduced and increased tension
levels on the children' academic performance. Measures
were also taken in the childrens' homes to determine to
what extent responses trained in the simulated academic
situation would generalize to and produce changes in their
activity levels and behavior.
Method
Subjects
Advertisements were placed in area newspapers asking
for the research participation of children between the ages
of 5 and 12 who exhibited symptoms of hyperactivity. Appli-
cants were screened with an EMG to determine whether they
maintained high tension levels while at rest for 4 minutes,
and while responding to problems on a mathematics placement
examination. Three children were accepted as suitable sub-
jects for the study.
13
Subject 1 was a 10-year-old male in the fourth grade.
He had been diagnosed as hyperkinetic by several physicians,
and began receiving Ritalin when he entered school. Prior
to the start of the study, he received 20 mg of Ritalin per
day. Medication was withdrawn with the physician's approval
several days before the screening session. The subject was
extremely small in size, emotionally immature, and a problem
both in school and at home, Because of his overactivity and
poor academic performancehe was placed in a special educa-
tion class for half of each school day. He exhibited such
hyperactive symptoms as low tolerance of frustration,
irritability, impulsive actions, frequent fighting, and
temper tantrums. His parents felt that medication was
inadequate in controlling his behavior problems. During
screening, he maintained an average frontalis muscular
activity level of 8,85 micro-volts root mean squared {uV RMS)
while at rest, and 8.03 uV RMS while responding to the
placement test. The nonhyperactive children screened
were found to maintain tension levels between 2.3 and 3.9
uV RMS.
Subject 2 was a 6-year-old female in the first grade.
She presented mild behavioral problems at school, but was a
superior student academically. At home, she tended to be
overactive and agressive towards her playmates. She was
somewhat disobedient, showed a low tolerance of frustra-
tion, fought, and displayed temper tantrums frequently.
During screening, she was found to produce an average
14
frontalis muscular activity level of 14.30 uV RMS while
at rest, and 8.65 uV RMS while responding to the placement
exam.
Subject 3 was a 9-year-old in the third grade. She
was extremely fidgety and highly distractible during the
screening session. Her parents and teachers felt that
she was overactive to the point where she presented problems
for them at home and in school. Her activity level and an
apparent lack of motivation academically resulted in her
placement in a special class for half of each school day.
Besides being overactive and distractible, she appeared
somewhat overemotional, exhibited great fluctuations in mood
from time to time, and was often uncooperative. During
screening, she was found to produce an average frontalis
muscular activity level of 10,40 uV RMS while at rest and
10.02 uV RMS while responding to the placement test.
Apparatus
The Bio-Feedback Technology (BFT) EMG 401 system was
used to promote effective training of control over tension
levels. The BFT EMG 401 system provided instrumentation
for muscle tension detection, information feedback, and
data collection. The system consisted of two instruments--
the feedback myograph (BFT EMG 401) and the time period
integrator (BFT TPI 215). The feedback myograph sensed
muscle-produced electrical activity through three silver-
chloride electrodes (6.5 mm in diameter) which were placed
25 mm above the eyebrows, and spaced 102 mm apart on the
15
subject's forehead. The EMG 401 amplified the detected
signal and provided both audio-information feedback to the
subject, and an electrical signal which was recorded by the
TPI 215. Auditory feedback was provided to the subject
via headphones or a speaker so that a tone varied in pitch
in direct proportion to the degree of forehead muscle
tension present--the higher the pitch of the tone, the higher
the EMG level of muscle tension present. The feedback
myograph was powered by low voltage, rechargeable batteries
and was electrically shielded for effective research use
in a normal office setting.
The TPI received the amplified EMG signal, averaged it
over a selected time period of 10 to 120 seconds, and
provided a digital readout in uV RMS. It had a noise
threshold feature which allowed both the detection of inter-
nal equipment noise and the cancellation of the effect of
that noise on the EMG uV RMS readings. It was 125 V AC
powered and for the safety of the participant was electri-
cally isolated by optical circuitry.
An actometer was used to provide a direct measure of
activity. The actometer was a modified automatic winding
calendar wristwatch which provided a direct drive from
the pendulum of the self-winding mechanism to the hands of
the watch. It was worn like a wristwatch fastened by a
watchband on the dominant wrist or ankle, Any movement of
the arm or leg produced a movement of the hands on the watch
and the degree of the movement was proportional to the extent
16
of the motion of the limb. Results were in actometer units
which is a rate measure defined as the number of activity
minutes per hour of elapsed time divided by 100. The
actometer has been shown to be a highly reliable instrument
and validity has been high when activity measures were
correlated with oxygen consumption (Schulman & Reismen,
1959).
A Biofeedback Training Release form concerning informed
consent was signed by the parent or guardian of all partici-
pants in the study (see Appendix A). A written rationale
of the training procedure and description of the equipment
was also given to each parent (see Appendix B). Lupin's
Child Behavior Rating Scale was used to provide a measure
of each subject's home behavior (see Appendix C). Ratings
from "0", for no occurrence, to "4", for frequent occurrence,
were obtained daily on 11 problem behaviors,
A programmed mathematics course (Sullivan, 1970)
consisting of 37 levels of mathematics workbooks, a place-
ment test, and a teacher's manual was used to provide a
measure of the subject's academic performance. A stop-
watch was used to record the cumulative time spent on
task during each mathematics session,
Procedure
At the start of every session, the EMG batteries were
checked for proper voltage level and the noise calibration
was made. The subject's forehead was prepared for electrode
contact by scrubbing the skin with an abrasive cleanser
17
(Brasivol) and swabbing it with alcohol. The function and
safety of the EMG system was explained .to the partici-
pant. Three electrodes held in place by a rubber headband
and treated with electrode cream were placed on the child's
forehead. He was then seated at a desk which was situated
away from electrical and ground sources as a general safety
precaution, and was asked to touch a ground screw on the
EMG unit to remove static electricity. Electrode resistance
was checked to insure that both active electrodes were below
20,000 ohms and approximately similar in resistance. The
child's tension level was monitored and recorded for two
120-second intervals while he remained inactive.
The initial baseline period (Phase 1) lasted five
sessions. While seated at a school desk in a fully
illuminated room, attached to the EMG system, and equipped
with an actometer on the dominant ankle, the child was
encouraged to work mathematics problems from an appropriate
level of the programmed mathematics course for 20 minutes.
The experimenter recorded the subject's tension level at
120-second intervals. On-task behavior was defined as
looking towards the workbook, writing answers in the work-
book, or asking problem-related questions of the experimenter.
The experimenter explained generally how problems were to be
worked, and encouraged the child to remain on task. The
child was not given answers to the problems, nor forced to
work. At the end of 20 minutes, the workbook was collected,
the actometer was removed and read, and the problems were
18
graded. The number of problems correctly answered was
used as the measure of mathematics performance. In addition,
the subject's activity was measured in his home for 1 hour
each evening by placing the actometer on the dominant
wrist. After being equipped with the actometer, the child
was allowed to resume his normal activities.
The biofeedback-assisted relaxation-training (Phase 2)
began on the sixth session. With the child reclined on a
cot and the lights turned off, the audio-feedback system
was explained. He was requested to relax, close his eyes,
and keep the tone's pitch as low as possible. The child
was told that he would receive tokens for reducing his
tension level. At the end of the session, he could trade
these for candy or toys. The headphones were placed on the
child and a 20-minute-biofeedback session was begun. To
facilitate the transfer of the training to a normal working
situation, the response of relaxation was shaped through
several positions. For the first 4 minutes, the subject
lay on the cot with his eyes closed. For the second 4
minutes, he lay with his eyes open. During the third
4 minutes, he was seated in a reclining chair with his
eyes closed, and for the fourth 4 minutes, the child sat in
the recliner with his eyes open. In the last 4 minutes,
the child sat at a school desk with his eyes closed. Ten
120-second interval EMG readings were taken, and at each
interval the child would receive verbal praise and one token
for meeting the following criteria; a reduction of 1.0 uV RMS
19
when the previous reading was 10.0 uV RMS or above, a
reduction of 0.5 uV RMS when the previous reading was
between 6.0 and 9.9 uV RMS, a reduction of 0.2 uV RMS when
the previous reading was between 3.1 and 5.9 uV RMS, or any
reading of 3.0 uV RMS or lower.
When the above sequence was completed, the subject was
seated at the desk with his eyes open, ready to begin the 20-
minute-mathematics session. The procedure was the same as
that followed in Phase 1 except that biofeedback was provided
through a speaker, and reinforcement of gains in relaxation
was continued in accordance with the above criteria. This
sequence was adhered to until tension levels stabilized at
or below 4.0 uV RMS during the first 20 mintues of training
and remained below 4.0 uV RMS during the mathematics
session. At this time, the first 20 minutes of training
was discontinued. When EMG levels remained at or below
4.0 uV RMS during the mathematics session, the biofeedback
was faded out over a 4-day period. On the 1st day,
biofeedback was discontinued during the last 6 minutes of the
mathematics session. On the 2nd day, it was discontinued
for the last 12 minutes of the session. On the 3rd day,
it was discontinued for the last 18 minutes of the session,
and on the 4th day, no biofeedback was provided. Reinforce-
ment according to relaxation criteria continued, When EMG
levels remained at or below 4.0 uV RMS for 3 consecutive
days without biofeedback, this phase ended.
20
The second baseline (Phase 3)consisted of the last
three sessions of the Phase 2 procedure, plus two sessions.
The child continued to receive reinforcement for meeting
the relaxation criteria, but no biofeedback was given. The
procedures used in Phase 1 were otherwise followed. Home
activity measures were obtained each day from this point
until the end of the program.
The biofeedback-assisted tension-training (Phase 4)
began after Phase 3. The subject sat at a desk with his
eyes open. He was asked to tense his muscles and to keep
the feedback tone as high as possible by whatever tactics
he could find, as long as these did not include gross
facial distortions which produced high readings that were
less reflective of entire body tension. The session lasted
for 20 minutes during which 10 EMG readings were taken at
120-second intervals. The subject received verbal praise
and one token for an increase of 1.0 uV RMS above the pre-
vious reading, or for any reading above 30.0 uV RMS, At
the end of this period, a 20-minute-mathematics session began.
The procedure followed while working mathematics remained
the same as during Phase 2 except that the child was rein-
forced for increases in tension according to the criteria
used during the tension-training period. When EMG readings
stabilized at or above 30.0 uV RMS during the 20 minutes
of tension training, and remained above that level while
working mathematics, the tension-training period was dis-
continued. Biofeedback and reinforcement were continued
21
during the mathematics session.. If EMG readings remained
above 30.0 uV RMS, the biofeedback was faded out over a
4-day period as in Phase 2. Reinforcement according to
tension criteria continued. When tension levels remained
above 30.0 uV RMS for 3 consecutive days without biofeedback,
this phase ended.
The third baseline (Phase 5) consisted of the last
three sessions of the Phase 4 procedure, plus two sessions.
The child continued to receive reinforcement for meeting
the tension criteria, but no feedback was given. The
procedures used in Phase 1 were otherwise followed.
A second biofeedback-assisted-relaxation phase (Phase 6)
began on the first session following Phase 5. Procedures
used during Phase 2 were followed exactly. The program
ended after the third session in which EMG tension levels
remained at or below 4.0 uV RMS without biofeedback. Home
activity measures and behavior ratings by parents were
continued during this phase.
Results
Subjects 1, 2, and 3 received a total of 48, 43, and
45 sessions, respectively. Sessions occurred daily,
Monday through Friday, over a 2k4-month period. The number
of sessions during Phases 2 and 4 were individualized,
whereas Phases 1, 3, and 5 consisted of five sessions each.
During Phase 2, Subjects 1, 2, and 3 reached the procedural
criteria in 18, 13, and 16 sessions, respectively. During
Phase 4, all subjects reached the procedural criteria in nine
sessions.
22
The data for each of the three subjects on seven
variables across three baseline and intervention periods
were subjected to a time-series analysis (Bower & Glass,
1974; Glass, Willson, & Gottman, 1975). One-tailed t tests
were used to decide whether there was any significant change
in levels between a given phase and the phase that followed.
A full summary of the data for each subject including the
model identification, level of the data at each phase,
amount of change in that level between phases, t value
for the change, error variance, and degrees of freedom for
the test are presented in Tables 1, 2, and 3 for Subjects 1,
2 and 3 respectively.
Subject 1
The first subject's presession frontalis EMG data (see
AppendiX D) were identified as following a first-order
moving-averages model. This model is identified by the
series of numbers 0, 0, 1. These refer to the order of
autoregression, the order of differencing, and the order
of moving averages, respectively. The level of the series
at Phase 1 was estimated to be 16.37 uV RMS. The inter-,
vention effect between Phase 1 and 2 was -10.33 uV RMS
which was significant, t (21) :--5.95, p< .0005. The
effect between Phase 2 and 3 was -0.65 uV RMS which was not
significant. The intervention effect between Phase 3 and4
was 8.32 uV RMS which was significant, t (12) = 2.46,
p < .025. Between Phase 4 and 5, the change in level
was -5.24 uV PMS which approached significance, t (12) = -1.40,
23
p< '.10. A significant change in level of -3.78 uV RMS
was found between Phase 5 and 6, t (9) = -3.28, p< .005.
The frontalis EMG readings during the mathematics
session (see Appendix E) followed a first-order autoregres-
sive model (1, 0, 0) for the analysis. The initial level
of the data was estimated to be 25.77 uV RMS. Between
Phase 1 and 2,a significant change of -22.75 uV RMS occurred,
t (21):= -6.63, p< .0005. The change in level between Phase
2 and 3 was 0.57 uV RMS which was not significant. Between
Phase 3 and 4, a significant increase in level of 40.17 uV
RMS was found, T (12) = 9.09, p < .0005. Between Phase 4
and 5, a decrease in level of 13.87 uV RMS was significant,
t (12) = -2.86, p,< .01. From Phase 5 to 6, a significant
change in level of -26.21 uV RMS occurred, t (9) = -8.92,
p< .005.
The percentage of time spent on task during the mathe-
matics session (see Appendix F) assumed the model 1, 0, 1.
At Phase 1, the subject spent an estimated 74% of his time
on task. An increase in level of 21% occurred between
Phase 1 and 2 which was significant, t (21) = 5.83, p < .0005.
A nonsignificant drop of 1% was found from Phase 2 to 3.
Between Phase 3 and 4, a drop of 8% was found to be signif-
icant, t (21) = -2.19, p < .025. From Phase 4 to 5, a
nonsignificant change of 1 % was found. From Phase 5 to 6,
a significant increase in level of 4% was obtained, t (9) =
5.35, p < .0005.
24
The mathematics session actometer data (see Appendix G)
assumed a model 0, 0, 1 for the analysis. An activity rate
of 11.94 was estimated as the level at Phase 1. From
Phase I to 2, a significant decrease in rate of 11.01 was
found, t (21) = -6.60, p <,.0005. From Phase 2 to 3, a
decrease in rate of 0.46 was not significant. A nonsignif-
icant increase of 0.19 was found between Phase 3 and 4.
A nonsignificant decrease of 1.20 occurred between Phase 4
and 5. The level at Phase 5 was 0.55 and a decrease of
0.36 to Phase 6 was significant, t (9) = -2.54, p< .0005.
The behavior rating data (see Appendix H) followed a
0, 0, 1 model and at Phase 1 the level was estimated to
be 25. No significant change in level occurred from Phase
I to 5. From Phase 5 to 6, however, a significant decrease
in level of 3 was found, t (9) = -2.54, p< .025.
The academic performance measure, number of mathematics
problems answered correctly, was highly variabled. Conse-
quently, an appropriate model was not clearly identified.
A significant change in the level of this variable did not
occur between any successive phase of the program for this
subject.
The home activity data were Identified as following
a 0, 0, 1 model. The level at Phase 1 was 105.60, and
a decrease in level to Phase 3 of 40.82 was significant,
t (8) = -8.61, p< .0005. The changes in level of
14.94 from Phase 3 to 4, and of 13.99 from Phase 4 to 5
25
were not significant. From Phase 5 to 6, a drop in rate of
30.38 was significant, t (9) = -2.90, p < .01.
Subject 2
This subject's presession frontalis EMG data (see
Appendix D) assumed a 0, 0, 1 model for the data analysis,
and the level at Phase 1 was estimated to be 11.70 uV RMS.
A change in level at Phase 2 of -6.98 uV RMS was significant,
t (16):= -4.59, p< '.0005. The decrease in level at Phase 3
of 1.52 uV RMS was not significant. From Phase 3 to 4 a
significant increase of 7.41 uV RMS was found, t (12) = 10.91,
p < .0005. The level at Phase 5 decreased by 5.50 uV RMS
which was significant, t (12) = -6.08, p < .0005. The
decrease of 0.72 uV RMS in Phase 6 was not significant.
The frontalis EMG data for the mathematics session
(see Appendix E) followed a 0, 0, 1 model. At Phase 1, the
level was estimated at 7.20 uV RMS, and the decrease of
3.32 uV RMS in Phase 2 was significant, t (16) = -3.10,
p < .0005. The change in level from Phase 2 to 3 of -1.02 uV
RMS was not significant. A significant increase of 38.21 uV
RMS occurred from Phase 3 to 4, t (12) = 18.16, p < .0005.
A decrease from Phase 4 to 5 of 12.07 uV RMS was significant,
t (12) = -4.96, p< '.0005. From Phase 5 to 6, the level
decreased by 31.64 uV RMS which was significant, t (9) =
-13.71, p < .0005.
A 0, 0, 2 model was used to analyze the percentage of
time spent on task (see Appendix F). At Phase 1, the level
26
was 75% and an increase of 20% in Phase 2 was significant,
t (16) = 4.76, p< .0005. From Phase 2 to 3, an increase in
level of 4% was not significant. A decrease in level from
Phase 3 to 4 of 19% was found to be significant, t (12)
-14.68, p< .0005. From Phase 4 to 5, the level increased
by 10% which was significant, t (12) = 4.74, p< .0005.
Another increase in level of 16% between Phase 5 and 6 was
also significant, t (9) = 3,99, p K .0025.
The mathematics-session actometer rates (see Appendix G)
assumed a 0, 0, 1 model. At Phase 1, the level was estimated
to be 1.13 and a decrease in level of 0..77 in Phase 2 was
significant, t (16) =-3,37, p< .0025. Between Phase 2 and
3, the change of -0.08 was not significant. The level
increased by 5.53 from Phase 3 to 4 which was significant,
t (12) = 2.30, p< .025. A drop in level of 5.11 from
Phase 4 to 5 was not significant. At Phase 5, the level
was estimated at 5.14, and a decrease in level of 4.82 at
Phase 6 was significant, t (9) = 4,14, p< .0025.
The behavior rating data (see Appendix H) were analyzed
using a 1, 0, 1 model. The level at Phase 1 was 26, and
a decrease of 9 in Phase 2 was significant, t (16) = -3.62,
p< .0025. No significant differences were obtained between
Phase 2 and 5. The level at Phase 5 was 15, and a decrease
of 5 in Phase 6 was significant, t (9) = -3.39, p< .005.
The number of problems worked correctly by this subject
were highly variable, but were analyzed using a 0, 0, 1
model. At Phase 1, the level was 78. No significant change
27
was found from Phase 1 to 3. From Phase 3 to 4, a decrease
in level of 34 was significant, t (12) = -7.93, p < .0005.
No significant differences were found from Phase 4 to 6.
The home actometer data followed a 1, 0, 1 model.
The level at Phase 1 was estimated to be 68.64, and an
increase in Phase 3 of 11.58 was significant, t (8) = 2.51,
p< .025. From Phase 3 to 4, a significant change of 33.81
occurred, t (12) = 2.68, p< .01. No significant change
occurred from Phase 4 to 6.
Subject 3
The presession frontalis EMG data (see Appendix D)
were analyzed with a 0, 0, 1 model. At Phase 1, the level
was estimated to be 12.16 uV RMS, A change of -4.36 uV RMS
in Phase 2 approached significance, t (19) = -1.54, p< .10.
A decrease of 3.43 uV RMS in Phase 3 was not significant.
The level increased in Phase 4 by 7.54 uV RMS which was
significant, t (12) = 3.56, p< .0025. A nonsignificant
decrease of 1.22 uV RMS occurred between Phase 4 and 5.
From Phase 5 to 6, a decrease of 6.61 uV RMS was highly
significant, t (9) = -9.13, p< .0005.
The mathematics session frontalis EMG data (see Appendix
E) followed a 0, 0, 2 model. The level at Phase I was
7.34 uV RMS. The level decreased by 2.76 uV RMS in Phase 2
which was significant, t (19) =-3.33, p< .0025. A
further decrease of 1.32 uV RMS in Phase 3 was not signif-
icant. From Phase 3 to 4, the Level increased significantly
by 34.47 uV RMS, t (12) 32.04, p< .0005. From Phase 4 to
28
5, the level decreased significantly by 11.58 uV RMS,
t (12) = -8.00, p < .0005. A further decrease of 22.94 uV
RMS from Phase 5 to 6 was significant, t (9) = -29.13,
p < .0005.
A 0, 0, 1 model was used to analyze the percentage
of time spent on task (see Appendix F). At Phase 1, the
level was 92%, and no significant change was evidenced
until Phase 4. in Phase 4, the level decreased by 18%,
which was significant, t (12) = -6.32, p< .0005. From
Phase 4 to 5, an increase in level of 5% was significant,
t (12) = 2.36, p< .025. The level again increased signif-
icantly by 11% from Phase 5 to 6, t (9) = 6.21, p< .0005.
The mathematics-session actometer data (see Appendix G)
were identified as follQwing a 1, 0, 1 model. No significant
change in level occurred from Phase 1 to 3. At Phase 3,
the level was 1.36 and an increase in rate of 38.02 in
Phase 4 was significant, t (12) = 5.74, p < .0005. No
significant change occurred from Phase 4 to 6.
The behavior rating data (see Appendix H) followed a
0, 0, 1 model. At Phase 1, the level was 19, and a decrease
of two in Phase 2 was significant, t (19) = -5.36, 2< .0005.
From Phase 2 to 4, no change occurred. The level decreased
by five in Phase 5 which was significant, t (12) = -1.89,
p< '.05. A further significant decrease of three occurred
in Phase 6, t (9) = -2.71, p,< .025.
The number of problems worked correctly were analyzed
with a 0, 1, 1 model. No significant change in level occurred
29
in any phase. Again, variability in the data made model
identification procedures inadequate.
The home actometer data assumed a 1, 0, 1 model. No
significant change was found from Phase I to 3. At Phase 3,
the level was estimated at 90.52 and a significant
increase in Phase 4 of 34.29 was found, t (12) = 2.56,
p < .025. From Phase 4 to 5, a significant decrease in
rate of 43.63 occurred, t (12) = -2.94, p .01. From
Phase 5 to 6, a significant increase of 35.45 was found,
t (9) 4.81, p< .0005.
Discussion
The measures of academic performance and home activity
were highly variable, and stable baselines were not attained.
They appeared to be influenced in an uncontrolled manner
by outside variables. The number of mathematics problems
correctly worked proved to be an unstable measure of academic
performance due to the variability in the difficulty level
in the problems presented (Ross, 1976). Because the level
of difficulty was generally increasing, improvements in
academic performance were difficult to obtain; if a subject
worked at the same rate each day, his mathematics performance
would appear to continually decrease. In order to use
mathematics problems correctly worked as a measure of
academic performance that reflected deficits or improvements
from day to day, the mathematics problems would need to be
of the same type, format, and order of difficulty throughout
30
the study. Such an approach would not closely approximate
the typical classroom situation, however.
The home actometer measures also appeared to be invalid
due to uncontrolled variation in the children's activities.
The trends in the weather tended to allow the children to
play outside more often as the study progressed which
produced increases in activity in most cases. Changes in
types of activity, choice of playmates, and location of
activity all produced much uncontrolled variation. It would
be necessary to standardize the location and activity of
children when measuring motoric activity to get an accurate
measure, and logistically this was beyond the capability of
this study. Both academic performance and home actometer
measures were judged to be invalid and are not considered
in the remainder of the discussion.
The Lupin's Child Behavior Rating Scale appeared to
be valid for measuring gross changes in a child's behavior,
and was concluded to be relevant in terms of consumer
attitudes of the parents. Since there was no reference
point against which to compare behaviors, apparently parents
rate their child according to his past few days' behavior.
Thus, there would appear to be no comparison of rates
between subjects, and slight daily changes within subjects
may not have been detected. Parent raters were not informed
as to which phase of treatment was occurring, but they may
have determined this to some extent by asking their children.
Thus, an expectancy bias cannot be ruled out. However,
31
expectancy during tension training may have been in the
wrong direction since "treatment" was ongoing. If that
were the case, expectancy bias could be assessed.
Because no reliability measures were established for
the on-task measures, they may have been subject to experi-
menter bias. The experimenter, knowing which phase of
treatment was occurring, would be subject to any expectancy
type of bias. However, both the responses of starting and
stopping the stopwatch, and the definitions of the behavior
being measured were simple. Therefore, the contamination
of the data from such bias may have been minimal. Kirby
and Shields (1972) used an identical on-task measurement
procedure with two raters, and obtained 98% agreement.
Subject 1
This subject exhibited desirable behavioral changes
concomitant with relaxation training. Changes occurred in
both the academic and the home settings.. Tension training
produced no significant lasting changes in his behavior,
however. As tension levels during mathematics decreased in
Phase 2, so did presession tension levels, and activity
levels during mathematics. He increased time spent on task
from 74% to 95%. These effects were maintained in Phase 3
with contingent reinforcement. As Lupin et al. (1974)
predicted, when tension levels drop, so do activity levels.
However, an improvement in behavior ratings by the parent
did not occur during this phase.
32
In Phase 4, presession tension levels increased along
with tension levels during mathematics. However, at Phase 1
the presession level was 16.37 uV RMS, and the increase in
Phase 4 was only to 12.51 uV RMS, below his initial
baseline. On-task behavior showed a significant reduction,
but remained well above baseline level. However, no change
in mathematics session activity level or home behavior
occurred. In Phase 5 both mathematics session and presession
tension levels dropped. The drop in mathematics session
tension was significant, and the drop in presession tension
approached significance. The percentage of time spent on
task, activity levels during mathematics, and behavior
ratings at home did not vary concomitantly with these
reductions in muscle-tension levels.
Results for Phase 6 were similar to those obtained in
Phase 2, and in addition, there was a significant drop in
occurrence of problem behaviors at home. For this subject,
relaxation training produced desirable changes in daily
tension levels, as seen in the decreases in the presession
and mathematics-session EMG readings, in academic performance,
as seen in the increase of percentage of time spent on task,
in activity levels in academic and home settings, and in
home behavior. Tension training may have aided the subject
to better distinguish between states of tension and relaxa-
tion, which would enhance his ability to control those states.
Tension training was associated with a significant reduction
33
in percentage of time spent on task in the laboratory
setting, however, no behavioral changes were effected in
the home situation.
Subject 2
As this subject learned to relax in Phase 2, presession
tension dropped significantly, on-task behavior increased
from 75% to 92%, activity level in the academic setting
decreased to nearly zero, and she exhibited a drop in the
occurrence of problem behaviors at home. In the academic
setting, the expected changes in her behavior occurred.
This subject reported that she practiced relaxation
techniques at home and in school after Phase 2 ended.
During Phase 3, tension levels in both presessions and
mathematics sessions remained low without the benefit of
biofeedback. The changes in percentage of time spent on
task, mathematics-session activity level, and problem
behavior at home were also maintained during this phase.
In Phase 4, as mathematics-session tension levels
increased, so did presession levels, but these remained
below the initial baseline level. This subject was parti-
cularly opposed to the tension requirements, and began to
show problem behaviors in the academic setting such as
crying, temper tantrums, and refusal to work. Time spent
on task declined significantly, and she exhibited increased
activity levels during the sessions. Behavior ratings at
home were highly variable, and no significant change occurred.
34
When tension training was ended in Phase 5, mathe-
matics-session tension level dropped significantly to just
above the reinforcement criterion level. Presession tension
showed a dramatic and significant drop back to the low
levels reached in Phase 3. Concomitant with these changes
percentage of time spent on task increased gradually, but
significantly, and mathematics-session activity levels
evidenced a strong trend toward improvement. The occurrence
of problem behaviors at home did not change significantly.
In Phase 6, mathematics-session tension levels were
reduced to below the levels reached in Phase 3. Presession
levels dropped to an extremely low point and remained there.
Percentage of time spent on task increased to 98% and mathe-
matics-session activity was significantly reduced. The
rate of problem behaviors at home dropped significantly, and
she ceased to be a problem in the laboratory setting.
Relaxation training reduced this subject's normal
tension levels, as seen by the decrease in presession EMG
readings. Desirable changes occurred in behavior at home
and in the academic setting. Tension training produced
the expected undesirable changes in behavior in the academic
setting--higher activity level, poor work performance,
tantrums, and other maladaptive and uncooperative responses.
Little effect was seen in home behavior coincident with
tension training.
35
Subject 3
Subject 3 was slow to respond to the relaxation training.
In Phase 2, mathematics-session tension levels dropped
significantly. Although there was a similar trend in
presession tension levels, the reduction approached, but
did not attain significance, At this point, she had not
learned to relax well enough for the effect to generalize
to other settings. There were no significant changes in
behavior in the laboratory setting, although on-task
behavior evidenced a desirable trend after an initial
reduction. Behavior ratings by the parents, although
variable, showed a significant decrease in the occurrence
of problem behavior during this period.
During Phase 3, presession tension levels remained low
suggesting that the relaxation training was slowly taking
effect. The improvements in behavior at home were maintained,
and time spent on task remained high.
This subject was the only one whose presession tension
levels increased to a higher point during Phase 4 than
were manifested in the initial baseline. A concomitant
drop in on-task behavior, and a dramatic increase in mathe-
matics'session activity resulted. However, there was no
significant change in home behavior.
In Phase 5, mathematics-session tension level dropped
significantly, and presession tension decreased to baseline
level. Mathematics-session activity level did not change
36
appreciably. On-task behavior increased, and the occur-
rence of problem behaviors at home decreased.
The second relaxation-training phase produced greater
gains in relaxation than attained previously. The subject's
presession readings dropped significantly, on-task behavior
greatly increased, and home behavior ratings showed improve-
ment. Although it did not attain a significant change in
level, activity during the mathematics session evidenced a
strong declining trend and variability reduced appreciably.
This subject was somewhat slow in responding to
relaxation training, and it was not until the second
training phase that the effects of reduced tension were
reflected in her behavior. She was quick to show behavioral
deficits following tension training, especially in the area
of overactivity. In the laboratory setting, when tension
level increased,activity level became very high--and as
tension was reduced,activity steadily decreased. Indeed,
tension and activity levels appeared to co-vary in this
subject. Little emotional change was noted following either
relaxation or tension training.
Summary
Combined results of all subjects indicate that EMG
biofeedback in conjunction with contingent reinforcement
was an effective procedure to either raise or lower EMG
levels of children engaged in an academic task. Such
training has typically been done with the participant
reclined and inactive. A review of the literature revealed
37
no studies in which relaxation was trained while a parti-
cipant was engaged in a task requiring motoric activity.
By training relaxation in an academic setting, the changes
in behavior occur where they are most needed, and with this
procedure no transfer of training is necessary.
The effects of EMG-biofeedback training while engaged
in an academic task (see Appendix E) were maintained by
contingent reinforcement for both tension reduction and
tension induction after biofeedback was withdrawn. While
working mathematics problems, it was found that high tension
levels were associated with high activity levels and
decreased academic performance. Low tension levels were
associated with low activity levels and increased academic
performance, as indicated by mathematics-session actometer-
rate data and percentage of time-spent-on-task data.
.The effects of EMG-biofeedback training to reduce
muscle tension in the laboratory setting transferred in
such a manner as to influence the ENG recordings taken
during the initial portion of subsequent sessions (see
Appendix D). This transfer of training was maintained by
reinforcement after biofeedback was withdrawn. However,
tension-induction training did not evidence this transfer
of training. Ratherthe EMG readings during the initial
portion of subsequent sessions returned to the levels
established during the initial baseline. Apparently,
tension induction did nothing more than negate effects of
prior relaxation training. Furthermore, when biofeedback
38
for tension induction was discontinued, the presession EMG
levels began to drop back to the levels reached during
tension-reduction training.
These findings are consistent with the data on the
occurrence of behavior problems in the home. Tension
reduction was associated with a concomitant reduction of
problem behaviors, whereas tension induction was not
accompanied by a change in the rate of problem behavior at
home. That tension induction showed less transfer than
tension reduction may have been due to the fact that the
subjects received a larger number of tension-reduction
trials than tension-induction trials,
Tension-induction training may have facilitated
relaxation training particularly with respect to the
latency of the onset of relaxation, and also possibly with
respect to the level of relaxation attained. Tension-
induction training may have aided the participant in
discriminating between states of tension and relaxation.
However, this improvement in relaxation during the final
phase may have resulted from the additional relaxation
training trials.
The changes in activity level, as indicated by the
mathematics-session actometer-rate data, and in academic
performance, as indicated by the percentage of time-spent-
on task data, were generally significant and in the predicted
direction.
39
Individual exceptions to these findings were present.
However, it is hypothesized that most of these exceptions
could be accounted for by a failure to extend a particular
experimental phase to the extent that is desirable in a
small N research design. In other cases, it may be that the
amount of uncontrolled variation was too great for parti-
cular effects to be validly assessed.
It should be noted that these findings were obtained
with children exhibiting both high activity levels and high
EMG levels, and may not hold for children exhibiting low
EMG levels. The results lend support to the efficacy of
using EMG-biofeedback training in conjunction with contingent
reinforcement of reductions in EMG-tension levels to reduce
overactivity, increase attention to an academic task, and
reduce problem behavior.
Further research using this procedure with extended
baseline and interventionperiods, a more controlled home
activity situation, and a uniform academic task would
illuminate the relationships between tension level, activity,
level, academic performance, and problem behavior. The
tendency of effects trained in the laboratory to transfer
to other settings could be more adequately measured. Future
researchers should withdraw contingent reinforcement to
determine if the effects of relaxation would be maintained.
It would be interesting to train problem students to relax
in a real classroom and measure the effects on behavior
there. Finally,if the effects of relaxation training on a
40
group of hyperactive children not receiving stimulant drugs
and a group of hyperactive children receiving the drugs were
compared, the information would be invaluable in assessment
of the possibility of using relaxation training as an alter-
native to drug treatment in the control of activity-level
problems.
4-4 0
U()4-1>
H -H
2 H
H
U)
0
U)Q0
a)
HQ) 4
0 re
> -H
fdl
HD a)H
Wa)
Hrd
IQ 0
p-H(Tirdl> K1
CN
Cr)
'0
10
(N
00
rn
IC
IC)
C)
(N
(N)
CY)
CNIC
H-
C)
00
(YC)
IC)
IC
(N
Ur)
co(N
01
00
03
r.
ILO
U)
U)4
LOLC)I) LO
0)C)C)C) ) C) H C
(N ( N H H-
U') 4 (
(N (N H H
Cr)
N131
CC)
0
H
N
NH
N
CC
0
(N
CY)
00
Nl-03
0
(N
'.0(N
03
IC) LO0
It!)
Hu)0 a)
-H c a)4J
EOco
-PU) C
CJ) iH
C) (n C)
H z (
(N N H
LC) HC) (Nq
I I
C) C)
H\ 00
LO
CD
0\0
CN
(N
CCl)
(2) U)
a) rdH0E-4
a4 P C)
41
(N
-
41
-4
H Y
SU)
E-4H
CC)
(N
(Y)'.0
C)!
CN
(N
CN
CC)
co En U) U) U)z z z z z
4-O rd
H-H() 4-4
4-40a~OU) 0
Q) rd
a)a)
4O0 4
r4 >
H1
Wa)
r -1
(d
-Ha
4
-Ha)
H
W )
0
U C
r-
r-4ro
H
rd
-H0
CO HH H-i
H .H r-4I
(N L o
L Lr)* C
00 0 00r- -i NCo a i o
U) a
CG OH o
H
H
N
H
(n
COr-A
LO C(N
m 00 00t- N k CO 00
(N COI IH(N)
4-40 En
~41
S0 -4O4 0
Z P4 C
10
1O
o Vo U) ) U) a
N N H H
C)
Ln
.>r
(N
0
r-4
Ha
H-iH
(N
H
(n
CO
a
I
(NCo
O
0
CO
o 1
(N
0(N
.
0
HLO0
r4I
HH-iH
C-O
10
N
(N
r-)HH
10
c%
0CoON
LOa
a
LOa
10L0
.
10
-H M4 -) (V3
E OW OW -H - r-I,-c00 rd4 U)40 r o
(1) 0 Q -U) F: r: 4 C)
42
H- H CNC\ C H
NH
N 10(N CO
0 LO
(N H
O 0
0\0 O\Or-i "i
0\0 0\0
N On
10
i IUT 10
0 (
14 4-)M
04)P (V r-H
C- (V E-
W-H %A E-' 0 H-I
-H
r-0
EdE-H
44Od
>H -H
C -
H0
0Qro
wrx
4 -Hrd
U04H
H
(dH-H 12)
0Mc
ON
00
CO
mO00
C)
0'
(N(N)
C
U)I0Nzi
C)
00
C4C)
Co
Y)
IS)
0u4 -)
o oC
-rd0 a) a
LOU U) -) U) (Nz z z
HCH4(N N a)(N (N H H--
H;I
r-4 CD
04
C;
0
C)C) N
CO (No C
LO
10
CN
00
C)
10q
43
1-
C) ) U) H
00 (N (N] ON-H Hr-
12)
-H4i
0
Cr) LO H C) C
(N (N (N (N (N
(N CO v L10 t'0
H N (N CO 10
U) 0
0 $:i o >r-
I
4-)
-H
-H
0
10 In IO0 0 C)0
Qo 1O. (N ( OH H H H
Ov
r-4Or
a)
H -H
> -H
a)
0 r
mu
-P o
rd4 E
C)Ord
4- 0
H
4 o
Scvc
r-v
rd H-|-H fc
H
Scv
Hr-d
-H
rdC4
H
C)0H
(NEn
r4
H
Ns
0
'0
0
Co
IO
H
InIn
0)InL(N
CO
H
(N
o10
N
rl-
C
0
L 10
H (N Vj vj In~i U
U>Ccv 5 H H
cv 0 %> 0P- W - 4 C0
I0 0 0 0
t o (N (N CnH H H H
0
C)H
00
In
(N
C)
N
In
Cr)
CNC)
(N
0
0
qcn
C12
r 0
) -H H H-
- U2C> oC~umC)
r-k0
CN
co
r-i
(N
(N)
00.
C)ro
N0
(N
H
CN
COr--q
4o
H
CO
In N -o0 CO
CO (N m
In D0I IT Ln)
o 0o U) 00 Z 0
. .(
H Ho H
t0
O\O
0C)
C>0
C)
O\O
0
coto
0\0(N
0
o\O cP L\O
(NH
0\0 0\0O 0\In N
(NHg
CO(N ro
cv) -r4 4-) r (N
-i C )P4 O Co 0C
44
C;)In
Co
r0
co
C-A
r-
(N
4-)
C)cv
04-4
>m
-H
cvmA H
-Hl
cvU)
cv
-H
I
0 CCN(NC)
S) o
S N .NH H H H
0)r4
> rH -H
0)L0)on -
-H
0
0rO
C)
-PH
0)
C)
H
Or
c-H
M Q
~0.r-4
H
I
10 1
(N
00 o
(N ((Y)
. .
C C
\O o\
0 0
0\ 0\C
H- H
o (N
4- d CN
a)1- 04
0 I
on 10
-p0)
Q0 00 C
lo m( V) Q
m 0 m * S
0 0 ND0L0
00 0 10 0HH
00 (N Om (N C
CN NV) 0 1
(N CO 0 k
~4-J(1) 0) 4
0)H 0Ha4 rQ-) c)
I
LO 0I I
q~z 10L
45
NCO
CO
10
CO
0
N
N
0
CO
H
H
a)
r-4
-Ati3NIi-ni.p.w.s.
CO
Co
C"
C4
(N
(N
COC0
CO
(N
H
CN(N
COq
10
N)
Ny
00
4 (N)
o r-rd 54-J H
rq O-q -E0 r0 .-H) CH)C)f~r
H
CN
CO
(N00
H
10
LnN H U U)
00 N N OH H
4-4Od
0- H
5' -HQ) 4-
44O
Q) rdW a)
0 (d4-
ara
rd
Q)Q)
U)
(d Q4
M -H
Crd,. )
-HG)H
4 rd
a)O
SW
o rd
>HO rO
rd E
00
NC
00Cn
00
IC)
HLC)
co
10LO
10
10n0
LO
0I
rHir-
CO 10
4- M C)
H O
o onM -H H
G)G rOE 4M cd C)o ( Da)
m r: NHf
LnC)
C)
NH
HH
Hs
'-
10
LnC) U) U) U) CDC> C)
1 ' N N O)H H H H
N e
I 1
N H ocO H CY
H H CO
I I
N N 10 Ns N. - Lf)- r.- .C)C) C) C C
mn C) I0 1 10
I I I
k-0 N r 1 C) 10 N Hs H H H
N
H
O k 10 O
N CO ) 10
$-4Cl) 04- - .:H - )
-- U r -r -1 04 H-4 4,.C>::: :: ,. ) r d
46
VC)
CN
H
H
N
koC4)
00
HH
a)
0
-aHArd
(N C)
m1 l C) CN) C
H H H H
Or)
4-H
0COto0
C)
0Ord4-H
fi~d
H
- 9(dH
U)
COr
-Q O
-H> rd
10
CN
410I
LN
I I
kD0H-
(NH-
(Y)0)
CO LO
C) C)
NC C
0 10
00
(NzH
0000
C)H--
H N (N Ci~1
0
i)iHH
0) C! >c)Z L
2C)
10 1 0 0 0(N C) ) CC) C) C ) C
C> 'Z C) C)
H H H Hr-
C)
10
00
CN
H
(N
C)
C)
I I
N
(N10
10r00
CO
(N CO
IE 0 >0-H H (N
4J ) C > C)
C)
C)
0
00
CN
00
(N)
0y)
(N
I)
C)
)
LO(N
1 10
10
C)
0\0
0)
C)
(N
0\0
(N 10 00
0\0(Nm) 00
LO
0)
(N COm
0)4JQ
U )E-4SC:)
P4)PH0 C
47
0
O
(N)
k.0
Co
-P
0)
CU)
rq
CO
E- -C
11"
CF)
LO
0O
CY)
a)
CO
CO
0)0)
H
N
(N
a- n (N (N onH H H H-I
Nr (N Ns(N ai W r-H
. 1 . .H1 C)1 H Hr4 C r-*4 r4i
I
N
H
C) ai Co (N. . 0
e n C) C
Ln Lr C(N H (N (N
HI
to
(-NH-
90
ci)H i)H
1 0
ZP4UC)
to
CN
r--U) U) C) U) U)r- Z-- --- Z--.-
.
-Ito
(Y)
CN
LO
on
L
CN
mo
to
N-
H
H3.~
(N
co
00Co
(NC)
Co
r4
H
rl-
r-
r4
00
N
H
H
(N
C)
C)
(NH
HN
H
(N m o q tO QO
o
0 o t
-H o - ,..4-) U ) ) C
48
U')LO 0N CDC) CD0 0
(N 0)H
00 O
0\0 0\0
O to
) C)
n ~\
H -H
> -H
4-40
i)0
r14
Od
6 -HH
-4 4Ici)
-
4 1
C d
PI0U
r-I4
-H 'i)
rci
IQ 0-d 12
Wro
>0
U C
0\0 4
(Nco
noQi)
:H4J
0u
Coci
LO tlo
I I
4-4 4J)0 f -4
(D rdr-l
4-H C)
P4 P 0 C)
LOC
z o oo
4-Od
CH -H r
-,4
4-40
n o
4-
Od
O -
rurd
H
H
H
4 J
.H
F-)
MOrd
-H
rd C
10(N
C)
(N)10
q1
CO
(NC)
00(N00
COjH--
0a
H-00
Cr)
(n
CN
CO
Co
co
C)
H
o
(Y)
0
a) )
(4 -Hg .i rd Co o w -O H
LO
U) r U 1 N
0) (N (N OH -H H Hr-
H-Nll
(N
C;
(NH
(N CO e 10 k
H~ (N CO 4.10
CO 00 rl l $H
-HH r- d -Hr: 04-H 14-)C)
CU) C) futZ
49
00 (N (NH Hr--
-H
-P
0
(Y)
ai)
rd-
10
-441
o Ha)
O N 0 co10 0 C) HD
CD U
H 1N1O0* r-4
50
Appendix ABiofeedback Training Release Form
I, , understand the general
procedure regarding the use of biofeedback training, and
am satisfied with the explanation given to me concerning the
function and safety of the equipment. I voluntarily agreed
to participate.
is currently in reasonably good
health, and is not being treated for any medical disorder
which would prohibit the use of biofeedback training.
Yes No
I, the undersigned, agree to hold North Texas State
University and/or their authorized representatives harmless
from liability for any injuries or damages resulting from
the intentional or unintentional use or misuse and/or
negligent use or misuse of any procedures included herein.
I understand that the data collected will be used
primarily for the participant's benefit and strict anonymity
will be adhered to if the data is utilized for other purposes
such as the advancement of knowledge in this area.
Participant Date
Parent or Guardian Witness
51
Appendix B
Rational of Biofeedback Training
Electromyograph biofeedback has been used effectively
in the treatment of tension related problems with adults,
and recently, has been used in conjunction with reinforce-
ment techniques to alleviate some symptoms of hyperactive
children. Several authors have suggest that a child's
activity level is related to his tension levels, and that
hyperactive children are also overly tense. Overactivity
is a means for them to allevaite some of their tension.
High tension also results in poor concentration on tasks and
many other problem behaviors.
The goal of this program is to make the child aware of
his tension level through biofeedback, and then to reinforce
him for raising and lowering that level when asked to do
so. In this way, the child can be taught to control his
tension level. The BFT EMG 401 system will be used to sense
electrical activity prodcued by muscles in the forehead
region, and to transpose this activity into a tone which
is fed back to the participant. The pitch of this varies
in proportion to the level of muscle activity present, and
this gives the participant information on this tension level.
The participant is sheilded from any electrical current by
optical isolation circuitry, so the equipment can in no
way result in electrical shock.
Though this procedure is experimental and no promises of
improvement can be made, it is believed that as the child
Appendix B--Continued 52
learns to control his tension level, he will also learn
to control his activity level, with a resultant improvement
in his behavioral and academic performance. Learning self-
control should also enhance the child's self-image allowing
him to react more positively and with less anxiety to other
persons and in new situations.
53
$4
010
.J rO
r-4 4..
.. C
cii - I
> >4-1
12-1 11 0 II II
-H H
rdd
<r4fn -'r4I I(I)
t1 u 1 1 1 1
0 En
4~)~ O 122I411 1M 1 1 1,. 4 )
0H 0
0 .p04' -r4 (0o Oed # --
4- 0 4-) -P Q) - o 4- 0r, 4J rdO r>, (D >0
0X0 > -$43)-M -4-
.r-4 u 4. a) rA 0 o > rCd 0 4j 4j >
04'40 :-0 0 A4 o 0 r-4
0 -P 0 4 44 U A4 p M 0) >1 0 - M M
04()Q- 4 44 U12)rd Q 0 4 H 4) r
: 04 4-J0 0-0 0 d Q H4- 44- I(dd UI U) IuI1-1 1 CO H 1)
0 to O 0 - 00 a
>1
H
r-i
Hr4
C)
-)
44
0
-rp
0
-)
ci)01
>ci)
Ir
-pU)
i4 )
0 41
o )
4- (l)0 7jo 4
9 (
C;
ri
d (1c) 4C
H O
H
ri
0HW
U)edr
HH
(0E-
1r-
54
rci
4-)
0C4
-irQ
55
Appendix D
Subj ect 1
I--I --- --- I It
Sujel
- % - - - 1
0
Subject 3
1 -4-----+ ---+I t- ----'
2 6 10 14 18 22 26 30 34 38 42 46
Sessions
Figure 1. Mean presession frontalis EMG levels for eachsubject.
56
Appendix E
Subject I
J4.4-~i-4-9 IT~ TTh
I'
-.ISubject 2
1~ -~k I
44-
ptt--4- - -----
tn AzzL. 7K11111Ii
-atI I I 1I L .~ .4 --- T-t-tlT -7
Subject 3
II
~ti~L~i --II - -- -- 'IIL It - - V -~
- "- -t- - -I- - -i - -+C
J1,
44111
L- -I--..
- ----- +t-- 1--
JItIEL10 14 18 22 26 30 34 38 42 46
Sessions
Figure 2. Mean mathematics frontalis EMG levels for each
subject.
7
i-iL':
In
CoI
co
(N
2 6
I
I
iI
.................. ...... ..
II I
u if--
I --- I I I I I I I I If I Io"Wo "mm- mom..-"
I t
®rww m
57
Appendix F
Subject 2
~F R--
C - -- ------ New
Subject 3
cc.o -
0-oI IF l
N MA~
6 10 14 18 22 26Li L 3L30 34 38 42 46
Sessions
Figure 3. Percentof time on-task during mathematics sessionfor each subject.
0)0o
C)
0l
Subject 1
7: III~III -~ liii--- -
- -" - - -1- - - - - - - - - - -
- - - - - I- -I
(d
0
N
-I
~i~L[i~IJI~ I L 2-1 L L1 V
21LVLJ I tJi.
I_
IL
58
Appendix G
Subject 1
Ell1
7jI
Subject 2
--- ------ --- --
I7 L u
["v'ljI'
{II
1 4 18 2
~1~~
K:1I.I-
L.
viVMN"
-4 -44--K-- --
bject 3
2 26 3034 38 42 46
Sessions.
Figure 4. Actometer rate during mathematics session for eachsubject.
(CNH-
00
I
H-
2'
Q)4-)
a)
04JFU
N
C)-I~: ~I
- I--1
Kz~-Y--t'ci' - -
2 6
...........i r
i
f--, 4- I-LL-V--4I
r - -
L
Appendix H
Subject 1
p L-
tz-iElz IzzzEmif-
zz: - 7:-?11pzEl i' - _ _ S
N\'I
subject
V9
JilL
IIIVKI~~~~~~ v F I ? A iI .f. . L~4I -+4 -* -
N I
'-u-I-I-I
A
I-II-uI
ilIMKI
4 & .. L
. .6. _ - a6 10 14 18 22
uL _
ect 3
26 30 34 38
7-.'
F- - - n
421 -6
42 46
Sessions
Figure 5. Daily sum of Lupin Child Behavior Behavior RatingScale for each subject.
59
0l
I)
0c:
L 1 .... _ F
4-)
0
rd
0
0(Y)
Q()
cq
I
------------
; I --I -4 rt--A ii -
-.1. to No a of 9- -
r-
"...I N- !, .. i At
2
1
IIIa
60
References
Ayllon, T.,, Layman, D., & Kandel, H. A behavioral-
educational alternative to drug control of hyperactive
children. Journal of Applied Behavioral Analysis,
1975, 8(2), 137-147.
Blanchard, E. B., & Young, L. D. Clinical applications of
biofeedback training: A review of evidence, Archives
of General Psychiatry, 1974, 30, 573-589.
Bower, C. & Glass, G. V. CORREL and TSX. Computer programs,
copyright June, 1974.
Bradley, C. The behavior of children receiving benzedrine.
American Journal of Psychiatry, 1937, 94, 577-585.
Braud, L. W. The effects of EMG biofeedback and progressive
relaxation upon hyperactivity and its behavioral concomi-
tants. Presented at the 22nd Annual Convention of the
Southwestern Psychological Association, April, 1975,
Houston, Texas.
Braud, L. W., Lupin, M. N., & Braud, W. G. The use of
EMG biofeedback in the control of hyperactivity.
Presented at the llth Annual Convention for Learning
Disabilities, February, 1974, Houston, Texas.
Budzynski, T. H. Biofeedback procedures in the clinic.
Seminars in Psychiatry, 1973, 5(4), 441-452.
Budzynski, T. H., & Stoyva, J. M. An instrument for
producing deep muscle relaxation by means of analog
information feedback. Journal of Applied Behavioral
Analysis, 1969, 2, 231-237.
61
Burks, H. The hyperkinetic child. Exceptional Child,
1960, 27, 18-26.
Comly, H. Cerebral stimulants for children with learning
disorders. Journal of Learning Disabilities, 1971,
4, 484-490.
Conners, K. C. Recent drug studies with hyperkinetic
children. Journal of Learning Disabilities, 1971,
4(9), 12-19.
Conners, K. C., & Eisenberg, L. The effects of methylpheni-
date on symptomatology and learning in disturbed children.
American Journal of Psychiatry, 1963, 120, 458-463.
Conners, K. C., Eisenberg, L., & Barcai, A. Effect of
dextroamphetamines in children: Studies on subject with
learning disabilities and school behavior problems.
Archives of General Psychiatry, 1967, 17, 478-485.
Conners, K. C., Eisenberg, L., & Sharpe, L. Effects of
methylphenidate (ritalin) on paired-associate learning
and porters maze performance in emotionally disturbed
children. Journal of Consulting Psychology, 1964, 28,
14-22.
Conners, K. C., & Rothschild, G. Drugs and learning in
children. In G. Helmuch (Ed.), Learning Disorders,
Vol. 3, Seattle: Special Child Publications, 1968.
Connoly, D. P., Besserman, R., & Kirschvink, J. Electromyo-
graphy biofeedback on hyperkinetic children, The Journal
of Biofeedback, 1974, 2. 24-30.
62
Denhoff, E., Davids, A., & Hawkins, R. Effects of dextro-
amphetamines on hyperkinetic children: A controlled
double blind study. Journal of Learning Disabilities,
1971, 4(9), 27-34.
Epstein, L., Lasagna, L., Conners, C., & Rodriquez, A.
Correlation of dextroamphetamine excretion and drug
response in hyperkinetic children. Journal of Nervous
and Mental Disease, 1968, 146, 136-146.
Ferman, S., & Feighner, A. Video feedback in treating
hyperkinetic children: A preliminary report. American
Journal of Psychiatry, 1973, 130(7), 792-796.
Fish, B. The "one child, one drug" myth of stimulants in
hyperkinesis: Importance of diagnostic categories in
evaluating treatment. Archives of General Psychiatry,
1971, 25, 193-203.
Glass, G. V., Willson, V. L., & Gottman, J. M. Design and
Analysis of Time-Series Experiments. Colorado Associated
University Press, Boulder, Colorado, 1975.
Hentoff, N. The drugged classroom. Evergreen Review,
December, 1970, 6-11.
Hollis, J., & St. Omer, V. Direct measurement of psycho-
pharmacologic response: Effects of chlorpromazine on
motor behavior of retarded children. American Journal
of Mental Deficiency, 1972, 76, 397-407.
Jacobson, E. Progressive Relaxation. Chicago: University
of Chicago Press, 1938.
63
Keogh, B. Hyperactivity and learning disorders: Review and
speculation. Exceptional Children, 1971, 38, 101-109.
Kinsley D. E. The effects of operant conditioning techni-
ques upon selected behavior of neurologically handicapped
children. Dissertation Abstracts, 1970, 30(9-A), 3792.
Kirby, F. D., & Shields, F. Modification of arithmetic
response rate and attending behiavor in a seventh grade
student. Journal of Applied Behavioral Analysis, 1972,
5, 79-84.
Knights, R., & Hinton, G. Minimal brain dysfunction:
Clinical and psychological test characteristics.
Academic Therapy, 1968, 4, 265-273.
Krippner, S., Silverman, R., Cavallo, M., & Healy,, A. A
study of hyperkinetic children receiving stimulant drugs.
Academic Therapy, 1973, 8, 261-269.
Ladd, E. Pills for classroom peace? Saturday Review,
November, 1970, 66-83.
Lupin, M., Braud, L. W., Braud, W. G,, & Duer, W. E. Effects
of relaxation upon hyperactivity using relaxation tapes
for children and parents. Presented at the llth Annual
Convention for Learning Disabilities, February, 1974,
Houston, Texas.
Morrison, J., & Stewart, M. Evidence for polygenetic inheri-
tance in the hyperactive child syndrome. American
Journal of Psychiatry, 130(7), 1973, 791-792.
Nall, A. Alpha training and the hyperkinetic child--Is it
effective? Academic Therapy, 1973, 9, 5-19.
64
Patterson, G. R. An application of conditioning techniques
to the control of a hyperactive child. In P. Ullman &
L. Krasner (Eds.), Case Studies in Behavior Modification,
New York: Holt, Rinehart, and Winston, 1965, 370-375.
Patterson, G. R., Jones, R., Whittier, J., & Wright, M.
A behavior modification technique for the hyperactive
child. Behavior Research and Therapy, 1965, 2, 217-226.
Pihl, R. D. Conditioning procedures with hyperactive
children. Neurology, 1967, 17, 421-423.
Ross, R. M. The effect of a free-time contingency of arith-
metic and problem behavior in the classroom. North
Texas State University Dissertation, August, 1976.
Sargent, J. D., Green E. E., & Walter, E. D. Preliminary
report on the use of autogenic feedback training in the
treatment of migrain and tension headaches. Psychosomatic
Medicine, 1973, 35, 129-135.
Schulman, J. L.,, & Resiman, J. M. An objective measure of
hyperactivity. American Journal of Mental Deficiency,
1959, 64, 455-456.
Sprague, R., Barnes, B., & Werry, J. Methylphenidate and
thoridazine: Learning, reaction time, activity, and
classroom behavior in disturbed children. American
Journal of Orthopsychiatry, 1970, 40, 615-628.
Sprague, R., & Toppe, L. Relationship between activity
level and delay of reinforcement. Experimental Child
Psychology, 1966, 3, 390-397.
65
Sroufe, L. A., & Stewart, M. A. Treating problem children
with stimulant drugs. The New England Journal of Medicine,
1973, 289(8), 407-412.
Stewart, M. A., Pitts, F. N., Craig, A. G., & Dieruf, W.
The hyperactive child syndrome. American Journal of
Orthopsychiatry, 1966, 36, 861-867.
Sullivan Mathematics Program. Palo Alto, California:
Behavioral Research Laboratories, 1970.
Sulzbacher, S. Behavior analysis of drug effects in the
classroom. In G. Semb (Ed.), Behavior Analysis and
Education. University of Kansas, 1972.
Sykes, D.,, Douglas, V., Weiss, G., & Minde, K. Attention
in hyperactive children and the effects of methylpheni-
date (ritalin). Journal of Child Psychology and
Psychiatry, 1971, 12, 129-139.
Twardosz, S., & Sajivaj, T. Multiple effects of a procedure
to increase sitting in a hyperactive retarded boy.
Journal of Applied Behavioral Analysis, 1972, 5, 73-78.
Weiss, G., Werry, J., Minde, K., Douglas, V., & Sykes, D.
Studies on the hyperactive child: The effects of dextro-
amphetamine and chlorpromazine on behavior and intel-
lectual functioning. Journal of Child Psychology and
Psychiatry, 1968, 9, 145-156.
Werry, J. S., Weiss, G., Douglas, V., & Martin, J. Studies
on the hyperactive child: The effects of chlorpromazine
upon behavior and learning ability. Journal of the
American Academy of Child Psychiatry, 1966, 5, 292-312.
66
Wiltz, N. A., & Gordon, S. B. Parental modification of
a child's behavior in an experimental residence. Journal
of Behavior Therapy and Experimental Psychiatry, 1974,
5 (1), 107-109.
Wolpe, J. Psychotherapy by Reciprocal Inhibition. Palo
Alto, California: Stanford University Press, 1958.
Zrull, J. P. Discussion of the effects of methylphenidate
on distrubed children. American Journal of Psychiatry,
1963, 120, 463-464.