Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2

0006-8993/$ - see frohttp://dx.doi.org/10.

nCorresponding aAustria. Fax: þ43 66

E-mail address: B

Research Report

Hormonal contraceptives masculinize brainactivation patterns in the absence of behavioralchanges in two numerical tasks

Belinda Pletzera,b,c,n, Martin Kronbichlerb,c,d, Hans-Christoph Nuerke,f,Hubert Kerschbauma,c

aDepartment of Cell Biology, University of Salzburg, AustriabDepartment of Psychology, University of Salzburg, AustriacCenter for Neurocognitive Research, Salzburg, AustriadNeuroscience Institute, Christian Doppler Clinic, Paracelsus Medical University, Salzburg, AustriaeDepartment of Psychology, University of Tuebingen, GermanyfIWM-KMRC, Knowledge Media Research Center, Tuebingen, Germany

a r t i c l e i n f o

Article history:

Accepted 5 November 2013

The aim of the present study was to identify, whether and how oral hormonal contra-

ceptives (OCs) alter women's number processing. Behavioral performance and brain

Available online 11 November 2013

Keywords:

Number processing

Hormonal contraceptives

Synthetic steroids

Sex hormones

fMRI

nt matter & 2013 Elsevie1016/j.brainres.2013.11.00

uthor at: Department o2 8044 5126.elinda.Pletzer@sbg.ac.at

a b s t r a c t

activation patterns (BOLD-response) of 14 OC-users were evaluated during two distinct

numerical tasks (number comparison, number bisection) and compared to 16 men (high

testosterone), and 16 naturally cycling women, once during their follicular (low hormone

levels) and once during their luteal cycle phase (high progesterone). For both tasks, reliable

sex differences and menstrual cycle dependent modulation have previously been

described. If progestogenic effects of the synthetic progestins contained in OC play a

predominant role, OC-users should be comparable to luteal women. If androgenic effects of

the synthetic steroids exert the progestogenic actions, OC-users should be comparable to

men. Likewise, if neither of the above are the case, the reduction of endogenous steroids by

OCs should make OC-users comparable to follicular women. Our findings suggest that

OC-users resemble follicular women in their behavioral performance, but show male-like

brain activation patterns during both tasks. Analysis of brain–behavior relationships

suggests that OC-users differ from naturally cycling women in the way they recruit their

neural resources to deal with challenges of the tasks. We conclude that OCs, which are

used by 100 million women worldwide, may have profound effects on cognition that have

not been recognized so far.

& 2013 Elsevier B.V. All rights reserved.

r B.V. All rights reserved.7

f Psychology, Paris-Lodron University Salzburg, Hellbrunnerstrasse 34, A-5020 Salzburg,

(B. Pletzer).

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 129

1. Introduction

Accumulating evidence shows that endogenous hormonalfluctuations, as observed during the menstrual cycle, affect avariety of behaviors and cognitive abilities in women. Forexample, women score lower on spatial tasks, but higher onverbal tasks, when estrogen plasma levels are high, like in thelate follicular (2–3 days pre-ovulation) and subsequent lutealphase compared to the menstrual and early follicular phase,when estrogen and progesterone plasma levels are low(Hampson, 1990; Hausmann et al., 2000; McCormick andTeillon, 2001; Rosenberg and Park, 2002; Mordecai et al., 2008;Otero Dadín et al., 2009). These behavioral variations have beenlinked to cyclic changes in brain structure (Protopopescu et al.,2008; Pletzer et al., 2010) and function (Dietrich et al., 2001;Fernandez et al., 2003; Schoning et al., 2007; Craig et al., 2008;Konrad et al., 2008; Weis et al., 2008, 2010).

However, although the hormonal contraceptive pill – ameans to controlling endogenous hormone fluctuations – ison the market for over 50 years now and used by 100 millionwomen worldwide, little attention has been paid to how thecontained synthetic steroids affect cognitive abilities andeven less to their modulation of the related brain activationpatterns.

The most commonly used hormonal contraceptive is thecombined oral contraceptive pill (OC), which typically con-tains 0.02–0.04 mg ethinylestradiol and varying levels ofheterogenous synthetic progestins. Levonorgestrel, derivedfrom 19-nortestosterone, belongs to the so called secondgeneration progestins and is still widely used, not only incombined OCs, but also in subdermal implants, intra-uterinedevices and hormone replacement therapy. Desogestrel, gesto-den, dienogest and norgestimat are so called third generationprogestins, also derived from 19-nortestosterone. Drospirenoneis a so called “new” progestin (fourth generation progestin),derived from spirolactone. These progestins and their meta-bolites differ in their binding affinity to steroid receptors,their transactivational activity on these receptors, their bind-ing affinity to the sex hormone binding globuline, their effecton the enzymes relevant to the synthesis of endogenoussteroids and their impact on blood glucose levels and thelipid profile (e.g. Sitruk-Ware, 2006).

It seems plausible that these synthetic progestins femin-ize the brain due to their actions on progesterone receptors.Also the reduction of endogenous testosterone levels in OC-users (e.g. Jung-Hoffman and Kuhl, 1987; Graham et al., 2007;Hietala et al., 2007) may contribute to feminizing effects ofOC. On the contrary, synthetic progestins may via a 2-foldmechanism also exert androgenic effects, thereby masculi-nizing the brain. First, many synthetic progestins are derivedfrom testosterone and able to activate androgen receptors(Sitruk-Ware, 2006; Wiegratz and Kuhl, 2006). Binding affinityto the androgen receptor and transactivational activity arethe lower, the higher the generation of progestin. Never-theless, some metabolites of 19-nortestosterone derived pro-gestins are able to activate androgen receptors (Perez-Palacios et al., 1992). Importantly however, OC lead to areduction of endogenous estradiol and progesterone levels(e.g. Sahlberg et al., 1987), thereby facilitating the conversion

of testosterone into the physiologically more active dihydro-testosterone (Wright et al., 1983). Additionally, a variety ofmasculinizing effects in the brain are thought to be promotedby estrogen receptors following the local conversion of testos-terone to estrogen via the enzyme aromatase (Roselli, 2007).Consequently, estrogenic actions of the ethinylestradiol com-pound may contribute to possible masculinizing effects of OCon the brain.

Evidence for behavioral differences between oral contra-ceptive (OC) users and non-users comes from a small numberof solitary studies scattered over several decades and differentcountries (Sheldrake and Cormack, 1976; Garrett and Elder,1984; Wright and Badia,1999; Kuhlmann and Wolf, 2005;Mordecai et al., 2008; Wharton et al., 2008; Nielsen et al., 2011).

Although sparse, these studies provide support for bothfeminizing and masculinizing effects of OC on behavioralperformance. On the one hand OC users show increasedverbal memory (Mordecai et al., 2008), increased recognitionworking memory during sleep deprivation (Wright andBadia,1999), a lack of memory impairment due to cortisol(Kuhlmann and Wolf, 2005) and better dream recall(Sheldrake and Cormack, 1976) compared to non-users. Onthe other hand, verbal reaction times are slower (Garrett andElder, 1984) and mental rotation performance is enhanced inOC users compared to non-users (Wright and Badia,1999;Wharton et al., 2008). Wharton et al. (2008) nicely demon-strated that mental rotation performance does not onlycorrelate with hormonal contraceptive use, but also withthe androgenicity of the progestin component . Users ofdrospirenone-containing contraceptives performed worse onthe mental rotation task than non-users. Furthermore, OCusers perform like men in an emotional memory paradigm(Nielsen et al., 2011).

Even less is known about how OC affect the neuralcorrelates of behavior, i.e. brain activation, and brain–beha-vior relationships. We recently demonstrated for the firsttime a profound difference in brain structure between hor-monal contraceptive users and non-users (Pletzer et al., 2010)and to our knowledge, there are only two functional imagingstudies, directly comparing brain activation patterns betweenoral contraceptive users and naturally cycling women duringcognitive tasks (Rumberg et al., 2010; Marecková et al., 2012).During verb generation, oral contraceptive users do not differfrom men in their brain activation patterns, but from natu-rally cycling women during both cycle phases (Rumberg et al.,2010). During face processing, oral contraceptive users showincreased activation of the fusiform face area, compared tonaturally cycling women (Marecková et al., 2012).

In the present study we seek to differentiate androgenic,progestogenic and endogenous hormone reduction effects ofOCs on behavior and brain activation during number proces-sing, by comparing OC-users separately to men (high testos-terone), women during their follicular phase (low hormones)and women during their luteal phase (high progesterone). Weaim to demonstrate the universality of these effects acrosstwo numerical tasks and several task modulations, for whichwe have recently demonstrated reliable sex differences andmenstrual cycle dependent modulation in behavioral perfor-mance and brain activation patterns (Pletzer et al., 2011,2013). In a number comparison task participants had to

Fofi

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

men follicular luteal OC

large distance

small distance ER #

non- mult.mult. mult. mult. mult.

non- mult.

non- mult.

non- mult.

Fig. 2 – Error rate (ER) patterns during number bisection. Oralcontraceptive (OC) users show a negative multiplicativityeffect (ER non-multiplicative – ER multiplicative items) likenaturally cycling women during their follicular phase. Whilethis reversal of the multiplicativity effect was observed insmall distance items in follicular women, it was observed inlarge distance items in OC-users. #Error rates were arcsinetransformed.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2130

decide which of two vertically aligned 2-digit numbers waslarger. In a number bisection task participants had to decidewhether the middle of three numbers in a row was thecorrect mean of the outer two numbers. In both tasksendogenous hormonal fluctuations influenced how perfor-mance was modulated by certain task factors, i.e. compat-ibility in number comparison and multiplicativity in numberbisection (see Experimental procedure for details). The result-ing sex and menstrual cycle dependent differences wereexplained by differences strategy choice. In the numbercomparison task sex differences in the compatibility effecthave been related to differences in the processing of globaland local stimulus aspects. We argued that men use a moreholistic strategy and compare whole number magnitudes,while women use a more decomposed strategy and comparedecade and unit digits separately (Pletzer et al., 2013). Innumber bisection sex differences in the multiplicativity effecthave been related to the use of math fact retrieval vs.magnitude processing (Pletzer et al., 2011).

While the specific underlying concepts are outside thescope of this paper and have been described in our previouspapers, we want to focus on how OC users relate to thepreviously described performance and brain activation pat-terns of men and naturally cycling women during differentcycle phases. If the effects of the synthetic progestins con-tained in OC strongly mimic the effects of progesterone, OC-users should be comparable to naturally cycling womenduring their luteal phase in their behavior and/or brainactivation patterns. If the synthetic steroids, however, arenot able to mimic the effects of the endogenous steroids onthe brain, we expect OC-users to be comparable to naturallycycling women during their follicular phase, due to the down-regulation of endogenous estrogen and progesterone levels.However, if the progestogenic effects of synthetic progestinsare even exerted by the possible androgenic mechanismsdescribed above, we hypothesize male-like performance andbrain activation patterns in OC-users. Most likely we are toexpect a combination of these effects, not only on brain

0

20

40

60

80

100

120

men follicular luteal OC

Decade Crossing

1

2

3

4

5

6

7

ig. 1 – Reaction times (RT) modulation during number comparisoral contraceptices (OC) showed a smaller decade crossing effect (ollicular phase. Like naturally cycling women, OC-users showedtems – RT compatible items) than men. WD¼within-decade.

activation and/or behavior, but also in particular on brain–behavior relationships.

2. Results

2.1. Behavioral data

2.1.1. Number comparisonRT of OC-users were compared to RT of naturally cyclingwomen during their follicular and luteal cycle phase and to RTof men in separate 2�2-ANOVAs, modulating either decadecrossing (WD vs. non-WD items) or compatibility (compatiblevs. incompatible items) as within-subjects-factor and group as

men follicular luteal OC 0

0

0

0

0

0

0

0

Compatibility

n. Like men and women during their luteal phase, users ofRT WD items – RT non-WD items) than women during theira significantly stronger compatibility effect (RT incompatible

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 131

between-subjects-factor (see Fig. 1). Since naturally cyclingwomen had been tested twice, session was used as a covariatewhen comparing OC-users to naturally cycling women. Over-all, RT of OC-users lay between those of men and naturallycycling women, thereby yielding no significant differences inresponse time between OC-users and men or OC-users andnaturally cycling women. Like men and naturally cyclingwomen during their luteal phase (Pletzer et al., 2013), OC-users showed a by trend weaker decade crossing effect thannaturally cycling women during their follicular phase (F¼3.28,p¼0.08). Like naturally cycling women (Pletzer et al., 2013), OC-users showed a significantly stronger compatibility effect in RTthan men (F¼5.55; po0.05).

2.1.2. Number bisectionER of OC-users were compared to ER of naturally cyclingwomen during their follicular and luteal phase and ER of menin separate 2�2�2 ANOVAs, modulating multiplicativity(multiplicative vs. non-multiplicative items) and distance(small distance vs. large distance items) as within-subjectsfactors and group as between-subject factors (Fig. 2). Sessionwas used as a covariate when comparing OC-users to natu-rally cycling women. The ER pattern of OC-users was com-parable to naturally cycling women during their follicularphase with the exception of a 3-fold interaction between OC-use, multiplicativity and distance (F¼5.96; po0.05). Likefollicular women (Pletzer et al., 2011), OC-users showed areversed multiplicativity effect compared to men (F¼5.44;po0.05). However, while naturally cycling women duringtheir follicular phase showed a reversed multiplicativityeffect in items with small distance, OC-users showed areversed multiplicativity effect in items with large distance.Like men (Pletzer et al., 2011), OC-users showed a reduceddistance effect compared to luteal women (F¼5.25; po0.05).

2.2. Neuroimaging data

2.2.1. Number comparisonFirst, lateralization indices (LIs) of OC-users in response toWD items and non-WD items were compared to LIs ofnaturally cycling women during their follicular and lutealphase and to LIs of men using separate 2�2-ANOVA designswith decade crossing (WD vs. non-WD) as within-subjectsfactor and group as between subjects factor. Second, BOLD-response to WD items, BOLD-response to non-WD items, thedecade crossing effect (WD items – non-WD items) in BOLD-response and the compatibility effect (incompatible items –

compatible items) in BOLD-response were compared betweenOC-users and naturally cycling women in their follicular andluteal cycle phase and between OC-users and men using fullfactorial designs. Session was used as a covariate whencomparing OC-users to naturally cycling women.

The decade crossing effect in lateralization indices of OCusers (0.065) lay between those of men (0.047) and womenduring their follicular phase (0.083) and did not differ sig-nificantly from men or naturally cycling women during theirfollicular or luteal phase (all Fo1.96, all p40.17). Conse-quently, we did not observe any differences in the BOLD-response decade crossing effect between OC-users and menor naturally cycling women. OC-users did furthermore not

differ from men in their BOLD-response to WD items, BOLD-response to non-WD items and the BOLD-response compat-ibility effect.

Differences in activation patterns to WD and non-WDitems between OC-users and naturally cycling women(Tables 1 and 2) Fig. 3 and Fig. 4 resembled the differencesbetween men and naturally cycling women (Pletzer et al.,2013). Furthermore, like men (Pletzer et al., 2013), OC-usersshowed a significantly weaker compatibility effect thannaturally cycling women during their follicular phase in theprecuneus ([0, �63, 51], k¼138, T26¼5.08, pFDRo0.001, Fig. 5).

2.2.2. Number bisectionThe multiplicativity effect (non-multiplicative items – multi-plicative items) and the distance effect (large distance items –

small distance items) in BOLD-response were comparedbetween OC-users and naturally cycling women in theirfollicular and luteal phase and between OC-users and menusing full factorial designs. Session was used as a covariatewhen comparing OC-users to naturally cycling women. OC-users did not differ from men in their BOLD-response multi-plicativity effect or the BOLD-response distance effect. Bytrend like men (Pletzer et al., 2011), OC-users showed a weakermultiplicativity effect than naturally cycling women duringtheir follicular phase and a weaker distance effect than natu-rally cycling women during their luteal phase in the mPFC(multiplicativity: [3, 33, 45], k¼40, T24¼3.71, pFDR¼0.104, dis-tance: [�6, 39, 36], k¼33, T24¼4.00, pFDR¼0.133) Fig. 6.

2.3. Brain–behavior relationships

Areas with significant brain–behavior relationships in men,naturally cycling women during their follicular or luteal cyclephase and OC users are displayed in Fig. 7. Correlationcoefficients and Fisher's z comparisons between groups aresummarized in Table 3.

2.3.1. Number comparisonMen show a negative relationship between RT and BOLD-response in left parietal and postcentral areas. The fastertheir reactions the stronger was the BOLD-response in theseareas. These relations were significantly stronger comparedto naturally cycling women. Contrarily, naturally cyclingwomen show a significantly positive relationship betweenRT and BOLD-response in the right postcentral gyrus duringtheir follicular phase. The slower their reactions the strongerwas the BOLD-response right postcentral. No apparent brain–behavior relations could be identified during the luteal phase.OC-users show a trend positive relationship between RT andBOLD-response in the right postcentral gyrus like womenduring their follicular phase. They show a significant negativerelationship between RT and BOLD-response in left latera-lized activation areas like men. However, while the negativerelationship was predominantly localized postcentral in men,it was located more posterior in the superior parietal lobule inOC-users. A bilateral negative relationship between RT andBOLD-response in the angular gyri was selectively present inmen and significantly stronger than in any female group.

Table 1 – Clusters with a significantly stronger BOLD-response to WD and non-WD number comparison items in naturally cycling women during their follicular phasecompared to men and OC users. g.¼gyrus, SMA¼supplementary motor area, PSPL¼posterior superior parietal lobule, inf.¼inferior.

Side Follicular4men Follicular4OC-users

Brain area MNI-coordinates (mm) #voxels T pFDR MNI-coordinates (mm) #voxels T pFDR

WD items X Y Z X Y Z

Lateral prefrontal cortex/precentral g./Postcentral g. Left �30 �3 63 168 5.39 o0.001 �27 �3 54 95 4.87 0.001�45 �18 39 92 4.21 0.001

Right 30 �3 51 35 4.39 0.044 21 �3 48 189 5.07 o0.00154 0 42 36 3.70 0.044

Middle frontal g. right 36 54 21 83 4.80 0.001SMA 0 15 51 103 4.43 o0.001 �3 3 54 68 4.28 0.003Postcentral g./Superior parietal lobule Left �33 �42 39 61 4.94 0.006 �27 �36 27 100 4.63 0.001

Right 27 �42 39 78 4.44 0.002 42 �33 42 39 4.09 0.03830 �54 60 34 4.54 0.047

Superior/middle occipital g. Right 30 �72 45 89 6.95 0.001 30 �69 42 151 7.18 o0.001Calcarine g./lingual g. Left �3 �75 9 46 4.23 0.021Fusiform g./inf. Occipital g./Cerebellum Left �27 �78 �12 37 5.10 0.041Cerebellum/fusiform g./lingual g. Right 30 �78 �21 36 4.07 0.044Inf./middle occipital g./inf./middle temporal g./fusiform g. Right 45 �75 �3 96 4.67 0.001

Non-WD itemsPSPL Right 30 �69 42 7.58 83 0.001 30 �69 42 240 7.71 o0.001

Left �24 �72 39 73 4.73 0.003SMA �3 0 63 4.87 82 0.001 �3 0 63 89 4.16 0.001Superior frontal gyrus Left �27 �6 54 4�79 111 o0.001 �27 �6 54 111 5.74 o0.001Middle frontal Right 36 45 30 48 4.29 0.016Precentral/inf. frontal g. Left �42 9 30 63 5.26 0.006Middle occipital g./middle temporal g. Right 45 �75 0 39 4.00 0.029

brain

research

1543

(2014)128–142

132

Tab

le2–Clusterswith

asign

ifica

ntlywea

ker

BOLD

-res

pon

seto

within-d

ecad

e(W

D)num

berco

mparison

item

sin

naturallycy

clingwom

enduringth

eirluteal

phas

eco

mpared

tom

enan

dor

alco

ntrac

eptive

(OC)use

rs.

Side

Lutealomen

LutealoOC-u

sers

Brain

area

MNI-co

ordinates

(mm)

#vox

els

Tp F

DR

MNI-co

ordinates

(mm)

#vox

els

Tp F

DR

WD

item

sX

YZ

XY

Z

Superior/middle

frontalgy

rus

Left

�24

5718

185

5.43

o0.00

1Med

ialprefrontalco

rtex

(mPF

C)

�6

2460

835.16

0.00

1�3

3357

544.13

0.03

0Middle/anteriorcingu

late

gyru

s/mPF

C�6

3936

128

4.96

o0.00

1�6

3936

394.97

0.06

3Anteriorcingu

late

gyru

s/mPF

CRight

1536

1852

4.04

0.01

1Prec

entral

gyru

sLe

ft�42

�3

4538

4.79

0.03

6Prec

uneu

s�3

�51

5476

5.75

0.00

2Inferiorparietallobu

les(IPL

)Le

ft�60

�48

3092

4.68

0.00

3

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 133

2.4. Number bisection

Men show a negative relationship between ER and BOLD-response in task-related deactivation areas during both mul-tiplicative and non-multiplicative items. The less errors theymade the stronger was the BOLD-response in the defaultmode network (DMN), i.e. the less deactivation was observed.For follicular women and OC users the same relation wasconfirmed only for non-multiplicative items, but was signifi-cantly weaker than in men for multiplicative items. For lutealwomen, the relation was confirmed only for multiplicativeitems, but was significantly weaker than in men for non-multiplicative items. Naturally cycling women show a sig-nificantly positive relationship between ER and BOLD-response in left lateral frontal task-related activation areas(middle and inferior frontal) during non-multiplicative, butnot multiplicative items. The more errors they made thestronger was their BOLD-response in these areas. The middlefrontal relationship was confirmed during both follicular andluteal phase and only numerically stronger than in men andOC users. The inferior frontal relationship was strongerduring the luteal than during the follicular phase and couldalso be confirmed in men, but not in OC users.

3. Discussion

We evaluated performance and brain activation patterns ofOC-users in two distinct numerical tasks, for which we hadpreviously observed reliable sex differences and menstrualcycle dependent modulation. We were able to show that forboth tasks OC-users mostly resemble naturally cycling womenin their behavioral performance, but show more frequentlymale-like brain activation patterns, since OC-users (a) do notdiffer from men in their brain activation patterns (comparableto results of Rumberg et al., 2010) and (b) differ from naturallycycling women in their brain activation patterns consistentlythe same way as men.

Particularly, a strong resemblance of the previously reportedfronto-parietal differences between men and follicular womenwas observed for overall BOLD-response to non-WD items andWD items in the number comparison task. Apparently, naturallycycling women recruit these regions more strongly during theirfollicular phase. For WD items, but not for non-WD items, thesedifferences were also present when comparing women's BOLD-response between their follicular and luteal phase. Importantly,during their luteal phase, women deactivate midline areas, inparticular the mPFC, more strongly than during their follicularphase andmen. OC user's lack this strong deactivation and differfrom women during luteal phase in their BOLD-response in themPFC like men. Thus OC users resemble both, men and lutealwomen concerning activation areas, but differ from lutealwomen like men concerning deactivation areas.

More importantly, OC-use did not only masculinize overallbrain activation but also interacted with the modulated task-factors in the same way as sex during both tasks. In particularOC-use reduced the compatibility effect in number compar-ison in the precuneus and by trend the multiplicativity effectin number bisection in the mPFC in comparison to womenduring their follicular phase. Additionally OC-use by trend

follicular > luteal1

follicular > men1

follicular > OC

OC > luteal

men > luteal1

Fig. 3 – Effects of oral contraceptives (OCs) on BOLD-Response to within-decade (WD) number comparison items. OC-dependent effects (magenta) on the BOLD-response to WD items in comparison to the previously published (1Pletzer et al.,2013) sex differences (red) and menstrual cycle dependent effects (green).

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2134

reduced the distance-effect in number bisection in the mPFCin comparison to luteal women. Thus, like in men, in OC-users default mode network deactivation was more stableand less affected by modulations of the tasks than innaturally cycling women. This suggests that as men,

OC-users put relatively less effort in the processing of morecomplicated items than naturally cycling women. Theincreased deactivation of DMN regions during processing ofcomplicated items parallels the increased recruitment ofactivation areas in the follicular phase.

follicular > OC

follicular > men1

follicular > luteal1

Fig. 4 – Effects of oral contraceptives (OCs) on BOLD-Response to non-within-decade (non-WD) number comparison items.OC-dependent effects (magenta) on the BOLD-response to non-WD items in comparison to the previously published(1Pletzer et al., 2013) sex differences (red) and menstrual cycle dependent effects (green).

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 135

In both tasks, the strongest sex differences have pre-viously been observed between men and women during theirfollicular phase (Pletzer et al., 2011, 2013), suggesting differ-ential recruitment of brain areas during number processing inthe follicular phase. While women during their luteal phaseresembled men more strongly in several aspects and differedfrom women during their follicular phase in a similar way asmen, there were additional differences between men andwomen during their luteal phase. In number comparison,women during their luteal phase showed significantly weakerBOLD-response to WD items compared to men in the leftprecentral gyrus, lateral frontal areas deactivation areas(Pletzer et al., 2013). In number bisection, women during theirluteal phase showed a significantly stronger distance effect inthe mPFC and lateral prefrontal cortex than men (Pletzeret al., 2011).

The fact that OC-users do not only differ from follicularwomen in a way comparable to both men and luteal women,but also differ from luteal women in the same way as mensuggests that the OC-dependent effects on brain activationpatterns are rather attributable to androgenic than progesto-genic effects of the synthetic steroids involved. This resem-blance of male brain activation patterns is especially striking,since the majority of OC users participating in this study were

presumably using OCs containing third and fourth generationprogestins, which did not evoke any androgenic or even anti-androgenic effects in in-vitro and animal studies (Sitruk-Ware, 2006; Raudrant and Rabe, 2003). Furthermore, it isnoteworthy that these androgenic mechanisms appear tocompensate the OC-dependent reduction in endogenoustestosterone level (Jung-Hoffmann and Kuhl, 1987; Hietalaet al., 2007).

Interestingly, however, OC-users largely resembled womenduring their follicular phase in their behavioral performance.Even though their BOLD-response compatibility effect in theprecuneus during number comparison was comparable to men,they did show a significant compatibility effect in RT, whichwas absent in men. Also, even though their BOLD-responsemultiplicativity effect in the mPFC during number bisection wascomparable to men, they did show a significantly reversedmultiplicativity effect in ER compared to men. However, thisreversion was only visible for the more difficult large distanceitems in OC-users, while it was more pronounced in the lessdifficult small distance items in follicular women.

Thus, while brain activation patterns of OC-users were moresimilar to men than to naturally cycling women, their beha-vioral performance was more similar to naturally cyclingwomen, particularly during the follicular phase, than to men.

Table 3 – Between-group comparison of brain–behavior relationships in number comparison and number bisection.

Men (n¼14) Follicular (n¼15) luteal (n¼15) OC (n¼14)

Number comparison RT_comp RT_incomp RT_comp RT_incomp RT_comp RT_incomp RT_comp RT_incomp

Postcentral_L �0.24 �0.39 0.48nnnn,a 0.49nnnn,a,c 0.13 �0.21 0.45a 0.48nnnn,a,c

Postcentral_R �0.49b �0.59n,b 0.25 0.15 �0.01 �0.19 �0.06 �0.03Parietal_inf_L 0.07 �0.20 �0.02 �0.09 0.15 �0.17 0.06 0.08Parietal_inf_R 0.01 �0.18 �0.09 �0.15 0.15 �0.03 �0.64n,a,b,c �0.51nnnn

Angular_L �0.26 �0.61n,b,c,d 0.05 �0.01 0.13 �0.04 0.21 0.30Angular_R �0.39 �0.69nnc,d 0.30 �0.45 0.09 �0.06 �0.31 �0.13

Number bisection ER_mult ER_nonmult ER_mult ER_nonmult ER_mult ER_nonmult ER_mult ER_nonmultFrontal_mid_L �0.30 0.13 0.17 0.55n �0.21 0.54n,e 0.21 0.19Frontal_mid_R �0.58n,b,c,d �0.11 0.34 0.15 0.17 0.44 0.13 0.17Frontal_inf_L 0.22 0.52nnnn,d �0.28 0.02 �0.40 0.69nn,b,d,e �0.08 �0.14Frontal_inf_R �0.27 0.03 0.25 0.10 �0.15 0.34 �0.20 0.07DMN �0.68nn,d �0.68nn,c �0.27 �0.52n �0.55n �0.13 0.06 �0.47nnnn

Correlation coefficients of mean reaction times (RT) in compatible (comp) and incompatible (incomp) number comparison items and error rates (ER) in multiplicative (mult) and non-multiplicative(nonmult) number bisection items with BOLD-response (principal eigenvariate) in selected regions of interest (ROIs). ROIs were identified from single group regression analyses (Fig. 7) as AAL areascontaining clusters with significant regression results. ROI maximum was used as seed voxel for extraction of eigenvariate.L¼ left, R¼right, parietal_inf¼ inferior part of superior parietal lobule, frontal_inf¼ inferior frontal gyrus (pars triangularis), frontal _mid¼middle frontal gyrus, DMN¼default mode network (includesbilateral medial prefrontal cortex, inferior parietal lobules, anterior cingulate gyrus and precuneus).n po0.05.nn po0.01.nnnn po0.1 (trend).a Significantly stronger than corresponding correlation in men according to Fisher's Z-transformation (po0.05).b Significantly stronger than corresponding correlation in follicular women according to Fisher's Z-transformation (po0.05).c Significantly stronger than corresponding correlation in luteal women according to Fisher's Z-transformation (po0.05).d Significantly stronger than corresponding correlation in OC users according to Fisher's Z-transformation (po0.05).e Significantly stronger than corresponding correlation in multiplicative items according to Fisher's Z-transformation (po0.05).

brain

research

1543

(2014)128–142

136

men > follicular1

OC > follicular

luteal > follicular1

Fig. 5 – Effects of oral contraceptives (OCs) on the BOLD-Response compatibility effect in number comparison. OC-dependenteffects (magenta) on the BOLD-response compatibility in comparison to the previously published (1Pletzer et al., 2013) sexdifferences (red) and menstrual cycle dependent effects (green).

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 137

Consequently, the behavioral results indicate OC-dependenteffects might be due to a reduction in endogenous sex hormonelevels, rather than actions of the synthetic progestins.

Behavioral differences in the absence of differences inBOLD-response have previously been reported for between-sex comparisons in a spatial and a verbal task (Halari et al.,2006). The authors argue that their findings provide supportfor sex differences in brain–behavior relationships. To testthis idea, we evaluated brain–behavior relationships in bothtasks for the two major conditions and compared thembetween groups.

In both tasks, brain–behavior relationships of OC-userswere comparable to those observed in women during theirfollicular phase with a few exceptions. In number comparisonOC-users as well as follicular women, but neither men norluteal women show RT modulation of BOLD-response in theright postcentral gyrus. However, OC-users additionally showa correlation between RT and BOLD-response in the inferiorpart of the left superior parietal lobule, which is absent inboth men and naturally cycling women. The strongest corre-lation within this area was observed roughly around theinferior parietal sulcus (IPS) a region implicated in numbermagnitude processing. Therefore, this relationship may beinvolved in the differential modulation of multiplicativity bydistance between OC users and women during their follicular

phase. In number bisection OC-users as well as follicularwomen do not show any brain–behavior relationships inmultiplicative items, and a significant relationship betweenER and DMN deactivation for non-multiplicative items. How-ever, OC-users like men lack brain–behavior relationships inmiddle frontal areas that were observed in naturally cyclingwomen during processing of non-multiplicative items. Thus,OC users appear to be more male-like in the processing ofnon-multiplicative items, but deal with multiplicative itemslike naturally cycling women. This modulation of brain–behavior relationships by multiplicativity might explainwhy the OC-dependent effects on BOLD-response patternswere not as strong during number bisection as duringnumber comparison.

It seems to be the case that the differences between OCusers and naturally cycling women are not so much routed indifferent strategy choices as has been suggested for sexdifferences, but in the recruitment of brain areas. It is along-lasting debate in the neuroimaging literature, whetherstronger activation of an area points to less efficient or moreefficient processing. Furthermore, increased default modenetwork deactivation has been related to increased taskdifficulty and RT on the one hand (McKiernan et al., 2003,Yarkoni et al., 2009), as well as enhanced performance on theother hand (Sambataro et al., 2010; Takeuchi et al., 2011).

Multiplicativity effect

Distance effect

follicular > men1

follicular > OC

luteal > men1

luteal > OC

follicular > luteal1

luteal > follicular1

Fig. 6 – Effects of oral contraceptives (OCs) on brain activation patterns in number bisection. OC-dependent effects (magenta)on the BOLD-response multiplicativity (upper half) and distance effect (lower half) in comparison to the previously published(1Pletzer et al., 2011) sex differences (red) and menstrual cycle dependent effects (green). Like men and naturally cyclingwomen during their luteal cycle phase, OC-users show a weaker multiplicativity effect in BOLD-response in the medialprefrontal cortex (mPFC) than naturally cycling women during their follicular phase. Like men, but unlike luteal women,OC-users show a significantly weaker distance effect in BOLD-response in the mPFC.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2138

Number Comparison Number Bisection

compatible incompatible multiplicative non-multiplicative

folli

cula

r m

en

lute

al

OC

Fig. 7 – Brain–behavior relationships during number comparison and number bisection. Results of regression analyses foreach group and item category. Clusters, where BOLD-response was significantly positively predicted by behavioralperformance (RT in number comparison, ER in number bisection), are displayed red. Clusters, where BOLD-response wassignificantly negatively predicted by behavioral performance are shown in green. FDR-corrected threshold was set at po0.05.See Results section for detailed description and labeling of clusters. See Table 3 for group comparisons. OC¼oralcontraceptives.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 139

Thus, OCs may alter the way women's brains deal with therequirements of a task and in how much they are able tocompensate for increased task difficulty by increased activa-tion or deactivation of an area.

In summary our data suggest that the actions of OC onbrain activation, behavior and brain–behavior relationshipsare not attributable to either androgenic or progestogeniceffects of the synthetic steroids or the reduction of endogen-ous steroids, but rather to a combination of these effects.Thus, these actions affect brain activation and behaviordifferentially, and may even affect different brain areas orcell types differentially depending on the receptor typesinvolved. Since the possible androgenic and progestogenicmechanisms are manifold and differ between progestins, it isof high importance to study the actions of different types ofprogestins in more detail and future studies on brain activa-tion patterns of OC users should assess and control the type

of progestin used in order to distinguish between theirandrogenic and progestogenic actions.

4. Experimental procedure

4.1. Participants

Fourteen healthy young women (mean age: 23.2273.51years), who were users of a combined oral contraceptive pill,participated in the study during their active pill phase. Thebrand of hormonal contraceptive was not recorded due to theexploratory nature of the study. A survey among 22 womenfrom the same population revealed that the progestin com-pound of OC was levonorgestrel in only 5% of participants, asecond generation progestin in 70% of participants anddrospirenone in 25% of participants. Sixteen naturally cycling

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2140

women (age: 26.5776.01 years) and 16 men (age: 25.1474.35years) had previously completed the same tasks (Pletzer et al.,2011, 2013). Naturally cycling women had a regular menstrualcycle (duration: 30.4673.37 days) and were tested twice, onceduring their early follicular (onset of menstruation to 5 daysbefore ovulation) and once during their mid-luteal phase (day3 post-ovulation to 5 days before onset of menstruation).Cycle phase was counterbalanced across scanning sessions.Cycle phase was determined by verbal reports (first day of lastperiod, cycle duration based on last 3 periods) prior to testingand confirmed by commercial ovulation tests and follow-upevaluation of the onset of the next menstruation. No subjectshowed brain tissue abnormalities on structural MRI.

4.2. Tasks

Stimuli were presented using Presentation Software (version0.71, 2009, Neurobehavioral Systems Inc., Albany, CA, USA) onan MR-compatible back-projection screen. Reaction times(RT) and error rates (ER) were evaluated.

4.2.1. Number comparison taskWe used the same number comparison stimuli as describedin Pletzer et al. (2012). In 150 items participants had toidentify the larger of two vertically aligned two-digit num-bers. In half of the items the upper number was larger and inthe other half the lower number was larger. We distinguishedwithin decade (WD) items from non-within decade (non-WD)items. In WD items both numbers contained the same decadedigit. Non-WD items were unit-decade compatible if the smallernumber contained the smaller unit digit (e.g. 42 vs. 67) and unit-decade incompatible otherwise (e.g. 47 vs. 62). Numbers rangedfrom 21 to 98. In non-WD items all four digits were different. InWD items unit digits were different from decade digits. Stimu-lus categories were matched for problem size, decades, unitsand parity. Additionally, participants passively viewed 30 con-trol items without responding, which consisted of hash marksinstead of numbers (null events, ## vs. ##). Order of stimuluscategories and control items was randomized. Each item waspresented for 2 s and followed by a one second inter-stimulusinterval. Probably due to the short presentation time and lowtask difficulty, we previously identified RT as the behavioralparameter most affected by endogenous hormone fluctuationsin number comparison. Thus, we focus the present behavioralanalysis of the number comparison task on RT.

4.2.2. Number bisection taskWe used the same number bisection task as described inPletzer et al. (2011). In 160 items, three two-digit numberswere displayed in a row (the smallest number on the left, thelargest number on the right) and separated by an underlinecharacter (e.g. 12_15_18). Participants had to decide, whetherthe middle of the three numbers was the correct mean of thenumbers on the left and on the right. All items werebisectable, i.e. the mean of left and right number was aninteger. Half of the items were correctly bisected (CB), theother half was not correctly bisected (NCB). CB items wereconsidered multiplicative, if the three numbers were part of amultiplication series (e.g. 12_15_18) and non-multiplicativeotherwise (e.g. 13_16_19). Distance was the difference

between the middle number and the left and right numberrespectively. It was considered large, if it was larger than five,and small, if it was smaller than 5. Stimulus categories werematched for problem size, distance and parity. Additionally,participants passively viewed 32 control items withoutresponding, which consisted of hash marks instead of num-bers (null events, ##_##_##). Order of stimulus categories andcontrol items was randomized. Each item was presented for5 s and followed by a 2.5 s inter-stimulus interval. Probablydue to the longer presentation time and higher task difficulty,we previously identified ER as the behavioral parameter mostaffected by endogenous hormone fluctuations in numberbisection. Thus, we focus the present behavioral analysis ofthe number bisection task on ER.

4.3. MRI data acquisition and analysis

Functional images were acquired using a T1-weighted singleshot echo planar (EPI) sequence (whole brain coverage, TE¼30ms, TR¼2100ms, flip angle 901, slice thickness 3.0 mm with0.6 mm gap, matrix 80�80, FOV 210mm, in-plane resolution2.6�2.6 mm) on a 3 T Philips Gyroscan NT scanner (PhilipsMedical System Inc., Maastricht, The Netherlands). Thirty-sixtransversal slices were taken oriented parallel to the AC-PC line.Furthermore, high resolution structural images were acquiredwith a T1-weighted 3D MPRAGE sequence (170 sagital slices,slice thickness¼1.2 mm, TE 3.3 ms, TR 6.8 ms, TI delay 854ms,FA 81, FOV 256�256, matrix 256�256).

SPM5 (http://www.fil.ion.ucl.ac.uk/spm) standard proceduresand templates were used for analysis of functional images. Thefirst five images of each session were discarded. Images wererealigned to correct for head movements, unwarped to correctfor interactions between head movements and EPI distortions(Andersson et al., 2001) and slice time corrected. One malesubject had to be removed from the sample due to excessivehead motion. The mean functional image was coregistered tothe high resolution structural image. The high resolutionstructural image was segmented and normalized to the MNIstandard stereotactic space and the resulting parameters wereused for normalization of functional images. Afterwards, functional images were resampled to isotropic 3�3�3mm voxelsand smoothed with a 6mm Gaussian kernel to enhanceactivation detection.

For statistical analysis we applied a two stage mixedeffects model. In the subject-dependent fixed-effects firstlevel analysis, each item category was modeled separatelyby a canonical hemodynamic response function. Data werehigh pass filtered with a cut-off of 128 s, and corrected forautocorrelation by an AR(1) model (Friston, 2002). The para-meter estimates of first-level contrasts were calculated in thecontext of a GLM. We defined first level contrasts for (i) allitems of a task compared to null events, (ii) each itemcategory of a task compared to null events, (iii) the compat-ibility effect in number comparison and the multiplicativityand distance effects in number bisection. The therebyobtained contrast images entered the group-based random-effects second level analysis. OC users were compared to theother groups using full factorial designs. For all analyses,we employed a cluster-level FDR-corrected threshold ofpo0.05. Primary thresholds were set at po0.005 (uncorrected).

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2 141

Task-related activation areas were defined as regions showinga significantly higher BOLD-response to a task compared tonull events. Deactivation areas were identified as regionsshowing a significantly lower BOLD-response to a task com-pared to null events. Regions of interest for brain–behaviorrelationships were identified using RT or ER as covariates inmultiple regression designs and defined via the WFU Pickatlas(http://www.nitrc.org/projects/wfu_pickatlas/) using AAL labelings (http://www.gin.cnrs.fr/spip.php?article217). Correlationcoefficients were calculated for each group after extraction ofthe principal eigenvariate within these ROIs and comparedbetween groups using Fisher's Z-transformation.

Acknowledgments

We acknowledge all participants for their time and will-ingness to contribute to this study. We thank GuilhermeWood and Korbinian Moeller for their help with stimuluspreparation, Markus Aichhorn, Juergen Bergmann and UlrikeKipman for their help with data acquisition and GuntherLadurner, posthoumously, for the possibility to conduct thisneuroimaging study.

r e f e r e n c e s

Andersson, J.L., Hutton, C., Ashburner, J., Turner, R., Friston, K.,2001. Modeling geometric deformations in EPI time series.Neuroimage 13, 903–919.

Craig, M.C., Fletcher, P.C., Daly, E.M., Rymer, J., Brammer, M.,Giampietro, V., Murphy, D.G., 2008. Physiological variation inestradiol and brain function: a functional magnetic resonanceimaging study of verbal memory across the follicular phase ofthe menstrual cycle. Horm. Behav. 53, 503–508.

Dietrich, T., Krings, T., Neulen, J., Willmes, K., Erberich, S., Thron,A., Sturm, W., 2001. Effects of blood estrogen level on corticalactivation patterns during cognitive activation as measuredby functional MRI. Neuroimage 13, 425–432.

Fernandez, G., Weis, S., Stoffel-Wagner, B., Tendolkar, I., Reuber,M., Beyenburg, S., Klaver, P., Fell, J., de Greiff, A., Ruhlmann, J.,Reul, J., Elger, C.E., 2003. Menstrual cycle-dependent neuralplasticity in the adult human brain is hormone, task, andregion specific. J. Neurosci. 23, 3790–3795.

Friston, K.J., 2002. Bayesian estimation of dynamical systems: anapplication to fMRI. Neuroimage 16, 513–530.

Garrett, K.F., Elder, S.T., 1984. The menstrual cycle from a bio-behavioral approach: a comparison of oral contraceptive andnon-contraceptive users. Int. J. Psychophysiol. 1, 209–214.

Graham, C.A, Bancroft, J., Doll, H.A., Greco, T., Tanner, A., 2007.Does oral contraceptive-induced reduction in freetestosterone adversely affect the sexuality or mood ofwomen?. Psychoneuroendocrinology 32, 246–255.

Halari, R., Sharma, T., Hines, M., Andrew, C., Simmons, A.,Kumari, V., 2006. Comparable fMRI activity with differentialbehavioral performance on mental rotation and overt verbalfluency tasks in healthy men and women. Exp. Brain Res. 169,1–14.

Hampson, E., 1990. Variations in sex-related cognitive abilitiesacross the menstrual cycle. Brain Cogn. 14, 26–43.

Hausmann, M., Slabbekoorn, D., Van Goozen, S.H., Cohen-Kettenis, P.T., Gunturkun, O., 2000. Sex hormones affectspatial abilities during the menstrual cycle. Behav. Neurosci.114, 1245–1250.

Hietala, M., Sandberg, T., Borg, A., Olsson, H., Jernstrom, H., 2007.Testosterone levels in relation to oral contraceptive use andthe androgen receptor CAG and GGC length polymorphisms inhealthy young women. Hum. Reprod. 22, 83–91.

Jung-Hoffman, C., Kuhl, H., 1987. Divergent effects of two low-dose oral contraceptives on sex hormone-binding globulinand free testosterone. Am. J. Obstet. Gynecol. 156, 199–203.

Konrad, C., Engelien, A., Schoning, S., Zwitserlood, P., Jansen, A.,Pletziger, E., Beizai, P., Kersting, A., Ohrmann, P., Luders, E.,Greb, R.R., Heindel, W., Arolt, V., Kugel, H., 2008. Thefunctional anatomy of semantic retrieval is influenced bygender, menstrual cycle, and sex hormones. J. Neural. Transm.115, 1327–1337.

Kuhlmann, S., Wolf, O.T., 2005. Cortisol and memory retrieval inwomen: influence of menstrual cycle and oral contraceptives.Psychopharmacology 183, 65–71.

Mareckova, K., Perrin, J.S., Nawaz Khan, I., Lawrence, C., Dickie,E., McQuiggan, D.A., Paus, T., The IMAGEN Consortium, 2012.Hormonal contraceptives, menstrual cycle and brain responseto faces. Soc. Cogn. Affect Neurosci.http://dxdoi.org/10.1093/scan/nss128 [epub ahead of print].

McCormick, C.M., Teillon, S.M., 2001. Menstrual cycle variation inspatial ability: relation to salivary cortisol levels. Horm. Behav.39, 29–38.

McKiernan, K.A., Kaufman, J.N., Kucera-Thompson, J., Binder, J.R.,2003. A parametric manipulation of factors affecting task-induced deactivation in functional neuroimaging. J. Cogn.Neurosci. 15, 394–408.

Mordecai, K.L., Rubin, L.H., Maki, P.M., 2008. Effects of menstrualcycle phase and oral contraceptive use on verbal memory.Horm. Behav. 54, 286–293.

Nielsen, S.E., Ertman, N., Lakhani, Y.S., Cahill, L., 2011. Hormonalcontraception usage is associated with altered memory for anemotional story. Neurobiol. Learn. Mem. 96, 378–384.

Otero Dadın, C., Rodrıguez Salgado, D., Andrade Fernandez, E.,2009. Natural sex hormone cycles and gender differences inmemory. Actas Esp. Psiquiatr. 37, 68–74.

Perez-Palacios, G., Cerbon, M.A., Pasapera, A.M., Castro, J.I.,Enrıquez, J., Vilchis, F., Garcıa, G.A., Moralı, G., Lemus, A.E.,1992. Mechanisms of hormonal and antihormonal action ofcontraceptive progestins at the molecular level. J. SteroidBiochem. Mol. Biol. 41, 479–485.

Pletzer, B., Kronbichler, M., Aichhorn, M., Bergmann, J., Ladurner,G., Kerschbaum, H.H., 2010. Menstrual cycle and hormonalcontraceptive use modulate human brain structure. Brain Res.1348, 55–62.

Pletzer, B., Kronbichler, M., Nuerk, H.-C., Kerschbaum, H.H., 2013.Sex differences in the processing of global vs. local stimulusaspects in a two-digit number comparison task – an fMRIstudy. PLoS One 8, e53824, http://dx.doi.org/10.1371/journal.pone.0053824.

Pletzer, B., Kronbichler, M., Ladurner, G., Nuerk, H.-C.,Kerschbaum, H.H., 2011. Menstrual cycle variations in theBOLD-response to a number bisection task: implications forresearch on sex differences. Brain Res. 1420, 37–47.

Protopopescu, X., Butler, T., Pan, H., Root, J., Altemus, M.,Polanecsky, M., McEwen, B., Silbersweig, D., Stern, E., 2008.Hippocampal structural changes across the menstrual cycle.Hippocampus 18, 985–988.

Raudrant, D., Rabe, T., 2003. Progestogens with antiandrogenicproperties. Drugs 63, 463–492.

Rosenberg, L., Park, S., 2002. Verbal and spatial functions acrossthe menstrual cycle in healthy young women.Psychoneuroendocrinology 27, 835–841.

Roselli, C.E., 2007. Brain aromatase: roles in reproduction andneuroprotection. J. Steroid Biochem. Mol. Biol. 106, 143–150.

Rumberg, B., Baars, A., Fiebach, J., Ladd, M.E., Forsting, M., Senf,W., Gizewski, E.R., 2010. Cycle and gender-specific cerebral

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 2 8 – 1 4 2142

activation during a verb generation task using fMRI:comparison of women in different cycle phases, under oralcontraception, and men. Neurosci. Res. 66, 366–371.

Sahlberg, B.-L., Landgren, B.M., Axelson, M., 1987. Metabolicprofiles of endogenous and ethynyl steroids in plasma andurine from women during administration of oralcontraceptives. J. Steroid Biochem. 26, 609–617.

Sambataro, F., Murty, V.P., Callicott, J.H., Tan, H.Y., Das, S.,Weinberger, D.R., Mattay, V.S., 2010. Age-related alterations indefault mode network: impact on working memoryperformance. Neurobiol. Aging 31, 839–852.

Schoning, S., Engelien, A., Kugel, H., Schafer, S., Schiffbauer, H.,Zwitserlood, P., Pletziger, E., Beizai, P., Kersting, A., Ohrmann,P., Greb, R.R., Lehmann, W., Heindel, W., Arolt, V., Konrad, C.,2007. Functional anatomy of visuo-spatial working memoryduring mental rotation is influenced by sex, menstrual cycle,and sex steroid hormones. Neuropsychologia 45, 3203–3214.

Sheldrake, P., Cormack, M., 1976. Dream recall and thecontraceptive pill. J. Nerv. Ment. Dis. 163, 59–60.

Sitruk-Ware, R., 2006. New progestogens for contraceptive use.Hum. Reprod. Update 12, 169–178.

Takeuchi, H., Taki, Y., Hashizume, H., Sassa, Y., Nagase, T.,Nouchi, R., Kawashima, R., 2011. Failing to deactivate: theassociation between brain activity during a working memorytask and creativity. Neuroimage 55, 681–687.

Weis, S., Hausmann, M., Stoffers, B., Vohn, R., Kellermann, T.,Sturm, W., 2008. Estradiol modulates functional brain

organization during the menstrual cycle: an analysis of

interhemispheric inhibition. J. Neurosci. 28, 13401–13410.Weis, S., Hausmann, M., Stoffers, B., Sturm, W., 2010. Dynamic

changes in functional cerebral connectivity of spatial

cognition during the menstrual cycle. Hum. Brain Mapp.http:

//dxdoi.org/10.1002/hbm.21126.Wharton, W., Hirshman, E., Merritt, P., Doyle, L., Paris, S., Gleason,

C., 2008. Oral contraceptives and androgenicity: influences on

visuospatial task performance in younger individuals. Exp.

Clin. Psychopharmacol. 16, 156–164.Wiegratz, I., Kuhl, H., 2006. Metabolic and clinical effects of

progestogens. Eur. J. Contracept. Reprod. Health Care 11,

153–161.Wright, F., Giacomini, M., Riahi, M., Mowszowicz, I., 1983.

Antihormone activity of progesterone and progestins

In: Bardin, C.W., Milgrom, E., Mauvais-Jarvis, P. (Eds.),

Progesterone and Progestins. Raven Press, pp. 121–134.Wright, K.P., Badia, P., 1999. Effects of menstrual cycle phase and

oral contraceptives on alertness, cognitive performance, and

circadian rhythms during sleep deprivation. Behav. Brain Res.

103, 185–194.Yarkoni, T., Barch, D.M., Gray, J.R., Conturo, T.E., Braver, T.S., 2009.

BOLD correlates of trial-by-trial reaction time variability in

gray and white matter: a multi-study fMRI analysis. PLoS One

4, e4257.