Skip to main content
NCBI home page
Human Brain Mapping logoLink to Human Brain Mapping
. 2012 Dec 20;35(2):712–722. doi: 10.1002/hbm.22187

Acute tryptophan depletion promotes an anterior‐to‐posterior fMRI activation shift during task switching in older adults

Melissa Lamar 1,2,3,✉, Michael Craig 2, Eileen M Daly 2, William J Cutter 2, Christine Tang 1, Michael Brammer 5, Katya Rubia 4, Declan GM Murphy 2
PMCID: PMC6868962  PMID: 23281064

Abstract

Studies have long reported that aging is associated with declines in several functions modulated by the prefrontal cortex, including executive functions like working memory, set shifting, and inhibitory control. The neurochemical basis to this is poorly understood, but may include the serotonergic system. We investigated the modulatory effect of serotonin using acute tryptophan depletion (ATD) during a cognitive switching task involving visual‐spatial set shifting modified for a functional MRI environment. Ten healthy women over 55 years were tested on two separate occasions in this within‐group double‐blind sham‐controlled crossover study to compare behavioral and physiological brain functioning following ATD and following a (“placebo”) sham depletion condition. ATD did not significantly affect task performance. It did modulate brain functional recruitment. During sham depletion women significantly activated the expected task‐relevant brain regions associated with the Switch task including prefrontal and anterior cingulate cortices. In contrast, following ATD participants activated posterior regions of brain more during switch than repeat trials. In addition to the main effects of depletion condition, a comparison of the ATD relative to the sham condition confirmed this anterior‐to‐posterior shift in activation. The posterior (increased) activation clusters significantly and negatively correlated with the reduced prefrontal activation clusters suggesting a compensation mechanism for reduced prefrontal activation during ATD. Thus, serotonin modulates an anterior‐to‐posterior shift of activation during cognitive switching in older adults. Neural adaptation to serotonin challenge during cognitive control may prove useful in determining cognitive vulnerability in older adults with a predisposition for serontonergic down‐regulation (e.g., in vascular or late life depression). Hum Brain Mapp 35:712–722, 2014. © 2012 Wiley Periodicals, Inc.

Keywords: acute tryptophan depletion, aging, fMRI, mental switching, prefrontal cortex, serotonin

INTRODUCTION

The prefrontal cortex (PFC) has been one of the main areas of focus for understanding normal age‐related cognitive decline. Studies consistently report that aging is associated with a decline in several functions modulated by the PFC, including executive functions like working memory, set shifting, and inhibitory control [Hedden and Gabrieli, 2004; Wecker et al., 2005]. In fact, the fronto‐striatal brain regions mediating these executive functions are more susceptible to age‐related structural alterations and neurotransmitter depletion than many other regions of brain [Raz et al., 2010, 1997; Resnick et al., 2003]. These vulnerabilities most likely all contribute to executive alterations in aging; however, the contribution of neurotransmitter depletion is poorly understood.

The PFC is richly supplied with serotonergic (5‐HT) neurons and previous studies have reported age‐related reductions of 5‐HT receptor density, ligand affinity, and binding potential in this region [Arranz et al., 1993; Goldberg et al., 2004; van Dyck et al., 2000]. Further, following a selective serotonin reuptake inhibitor (SSRI) challenge using PET [Goldberg et al., 2004], a double dissociation was reported such that increasing age was associated with decreasing metabolism in anterior brain regions including the PFC and anterior cingulate while increasing age was associated with increasing metabolism in posterior regions including the precuneus and temporal lobe. The authors concluded that age‐related loss of 5‐HT innervation led to different compensatory processes in anterior compared to posterior regions of the aging brain. Only one study to date [Lamar et al., 2009]; however, has examined the interaction between age‐related modulation of the serotonergic system and alterations in executive functions as it relates to brain recruitment and patterns of activation.

We previously reported [Lamar et al., 2009] that 5‐HT depletion following modulation of the serontonergic system through acute tryptophan depletion (ATD) altered regional brain function during an executive task in a double‐blind, placebo controlled, cross‐over functional magnetic resonance imaging (fMRI) study. Specifically, we observed an anterior‐to‐posterior shift in brain activation following ATD when compared to sham/placebo during an interference inhibition Simon task in women over 55 years of age. This shift in activation involved reduced PFC activation within fronto‐cingulo‐striatal regions and increased activity within neocerebellum and parietal regions. We hypothesized that this anterior‐to‐posterior shift in activation may be a compensatory mechanism of the aging brain in response to reduced 5‐HT in fronto‐striatal regions richly supplied with 5‐HT neurons made inert during ATD. However, it remains unclear whether our findings are generalizable to other aspects of executive functioning or restricted to cognitive interference inhibition.

Set shifting or switching—as assessed during a switch task—requires mental alternation from one stimulus‐response association set to another [Rubia et al., 2006; Smith et al., 2004]; cognitive flexibility associated with executive functioning. It differs from the previously investigated inhibitory control functions required by the Simon task in that the visuo‐spatial set shifting required during a switch task involves establishing and maintaining a mental set before an unexpected shift of set is demanded by task constraints. fMRI studies in adults 19–45 years of age, have reported that cognitive switching is associated with fronto‐striatal and fronto‐parietal brain activation with increased activity in lateral inferior and orbital prefrontal, cingulate, striatal, and parietal cortices during successful switch performance [Christakou et al., 2009; Dove et al., 2000; Gruber et al., 2010, 2009; Loose et al., 2006; Rubia et al., 2006; Smith et al., 2004; Yeung et al., 2006]. Similar, and at times greater activation within inferior frontal, orbitofrontal, and parietal regions has also been reported in adults over 60 years of age during task switching [Christakou et al., 2009; Smith et al., 2001; Townsend et al., 2006]. Older adults also activate posterior (i.e., temporo‐parietal and occipital) brain regions [DiGirolamo et al., 2001]. Similar fronto‐striatal and fronto‐parietal brain activation patterns are being recruited in adults for cognitive switching during the switch task as those seen in cognitive interference inhibition during the Simon task. This combined with the fact that declines in PFC during ATD often occur for tasks requiring greater inferior frontal and orbitofrontal involvement (see Evers et al., 2010 for review), we tested whether 5‐HT modulation during switching would lead to a similar anterior‐to‐posterior shift in patterns of activation in older adults like that seen for cognitive interference inhibition [Lamar et al., 2009].

We therefore tested the effect of ATD on cognitive switching with a visual‐spatial set shifting task in women over 55 years of age during rapid event related fMRI. Given that women are more susceptible to the effects of ATD than men [Hood et al., 2005], we focused our recruitment on women for this study. Thus, while behavioral variability exists in terms of 5‐HT modulation on switch task performance [see Mendelsohn et al., 2009 for review], manipulation of 5‐HT modulates brain regions known to be involved with this task [Goldberg et al., 2004; Smith et al., 2011]. Thus, we hypothesized that ATD would be associated with an anterior‐to‐posterior shift in brain activation during the switching task in our sample of older adults. More specifically, ATD would decrease inferior/orbital prefrontal activation and increase more posterior (e.g., temporo‐parietal and neocerebellar) areas of activation regardless of its impact on behavioral output.

MATERIALS AND METHODS

Participants

Women over 55‐years‐old with no reported mental health problems were screened for study consideration via a telephone interview. Successfully screened individuals visited the Institute of Psychiatry (IOP) where they underwent the Structured Clinical Interview for DSM‐IV (SCID) defined Axis I disorders (SCID‐I) and Axis II: personality disorders [SCID‐II; APA, 1994]. Participants completed the Mini‐Mental State Examination [MMSE; Folstein et al., 1974], the Beck Depression Inventory [BDI; Beck and Steer, 1993] and the Beck Anxiety Inventory [BAI; Beck et al., 1988]. Routine blood and visual testing was also conducted. Exclusionary criteria included current or past personal history of comorbid medical or psychiatric diagnoses, substance/alcohol abuse, biochemical, or hematological abnormalities that may affect brain functioning and/or any foreign bodies that would preclude neuroimaging. We also excluded individuals with a past family history of clinical depression given this could exacerbate the effects of ATD [Young et al., 1985].

Ten right‐handed women met study criteria and averaged 63 years of age and 13 years of education. They displayed normal levels of cognitive (MMSE: 29.4 ± 0.8) and emotional (BDI: 3.9 ± 2.4; BAI: 3.9 ± 2.8) functioning. All women were right handed and 70% reported they had never smoked. All reported alcohol intake below the (British) Institute of Alcohol Studies' recommendation of 14 units (112 g)/week (average units: 6.2 ± 5.0). Participants had not received hormone therapy since completing menopause over 10 years prior (Table 1) nor were they on any other prescription medication at the time of testing. Biological measures of estradiol (mean levels 57.9 ± 15.9 pmol/l), follicle stimulating hormone (mean levels 72.0 ± 19.5 U/l) and leutinizing hormone (mean levels 32.1 ± 7.4 U/l) confirmed participants' postmenopausal state. All participants gave written informed consent in accordance with the Declaration of Helsinki. King's College Hospital and the IOP ethics committees approved this study prior to participant recruitment and testing.

Table 1.

Baseline participant characteristics (n = 10)

Mean (s.d.) Range
Age at initial scan 63.00 (5.3) 56–73
Years of education 13.60 (3.2) 10–18
Years postmenopausal 11.60 (4.7) 6–20
Mini‐Mental State Examination 29.40 (0.8) 28–30
WASI‐FSIQ 111.5 (14.2) 86–129
Anger questionnaire total score 55.20 (18.5) 37–97
Beck's Depression Inventory total score 3.60 (2.4) 1–9
Beck's Anxiety Inventory total score 3.90 (2.8) 0–9
Mood Rating Scale
Factor 1: Alertness 156.4 (137.17) 18–410
Factor 2: Contentedness 63.89 (60.86) 12–201
Factor 3: Calmness 28.33 (22.83) 7–71
Open in a new tab

s.d., standard deviation; WASI‐FSIQ, Wechsler Abbreviated Scale of Intelligence Full Scale Intelligence Quotient.

Design and procedures

This within‐group double‐blind sham‐controlled crossover study, described in detail elsewhere [Lamar et al., 2009], was counterbalanced for order of administration of drink type. Thus, five women were randomly assigned to receive the sham drink during their first scanning session and the tryptophan depleted drink during their second scanning session. The remaining five participants experienced the drinks in reverse order. Despite the small overall sample size (n = 10), the fact that each participant served as their own control by receiving two separate MRI studies adds to the strength of this study. Each session was separated by a 1‐week washout period bar two participants who completed their second sessions after 6‐ and 8‐day washout periods.

A rapid mixed trial event‐related fMRI design was used with jittered inter‐trial intervals and randomized presentation to optimize statistical efficiency [Dale, 1999]. Participants practiced the fMRI task once prior to scanning and task instructions were repeated once the participant was in the scanner; furthermore, exogenous cues were displayed at every trial of the Switch task visually instructing participants on their expected focus of attention. The task was projected through a mirror in the head coil onto a screen and participants made their response on a button box using their left or right thumb.

Depletion and monitoring

After having fasted from midnight, subjects drank a 100 g amino acid mixture 4.5 h prior to testing within the scanner. The exact 100 g tryptophan depletion mixture was composed of the following: l‐alanine 5.5 g, l‐arginine 4.9 g, l‐cysteine 2.7 g, glycine 3.2 g, l‐histidine 3.2 g, l‐isoleucine 8.0 g, l‐leucine 13.5 g, l‐lysine monohydrochloride 8.9 g, l‐methionine 3.0 g, l‐phenylalanine 5.7 g, l‐proline 12.2 g, l‐serine 6.9 g, l‐threonine 6.5 g, l‐tyrosine 6.9 g, and l‐valine 8.9 g. The balanced mixture (sham) contained the same amino acids plus 2.3 g tryptophan. The powder mixture was combined with cold water and a flavor of the participant's choice then drunk over 10 min.

Prior to drink ingestion, participants completed baseline questionnaires probing aspects of emotional functioning including mood and aggression (see descriptions in “Affective Measures”) and a blood sample was taken. These exact procedures were conducted again immediately prior to scanning (4.5 h after ingestion) with cognitive testing occurring after scanning to minimize fatigue during fMRI. This time frame is considered to be optimal for capitalizing on the effects of ATD on behavioral and plasma measures [Young et al., 1988, 1985] and brain 5‐HT synthesis [Nishizawa et al., 1997]. Total plasma tryptophan levels were determined from each blood sample using previously described methods [Keating et al., 1993].

Affective measures

ATD has been reported to increase levels of aggression in susceptible individuals [Bjork et al., 2000]. Therefore, 2–4 weeks before neuroimaging sessions participants completed an anger questionnaire [Buss and Perry, 1992; Table 1]. Further, self‐report questionnaires were completed before (baseline) and 4.5 h after (follow‐up) drink ingestions. These included the Mood Rating Scale [MRS; Bond and Lader, 1974], an Aggression Rating Scale [ARS; Bond and Lader, 1986], which measured various aspects of mood and aggression. Other physical symptoms reported to be associated with ATD including nausea, dizziness, irritability, and anxiety were also measured using the Bodily Symptoms Scale [BSS; Cleare and Bond, 1995; Higgitt et al., 1986].

fMRI paradigm

A modified fMRI version of the Meiran Switch task was used [Meiran, 1996] that isolated brain activation areas related purely to switching behavior and independent of the effect of working memory [for details see Christakou et al., 2009; Rubia et al., 2006; Smith et al., 2004].

Subjects used a keypad with four buttons in a diamond configuration in order to make responses. Subjects were presented with a grid divided into four squares, in the center of which was a double‐headed arrow positioned either horizontally or vertically. The grid with the double‐headed arrow was presented for 1600 ms; 200 ms after presentation of the grid and arrows, a red dot appeared for 1400 ms in any one of the four squares of the grid. A horizontally pointing double‐headed arrow indicated that the subject had to confirm whether the circle was in either of the two left or the two right squares of the grid, by pressing the left or right button. After the 1600 ms of presentation time, there was a blank screen for 800 ms. Subjects were instructed to make their responses as soon as possible after stimulus presentation. This presentation was repeated for several repeat trials with a total ISI of 2.4 s. A minimum of four repeat trials were followed by a switch trial where the double‐headed arrows in the middle of the grid changed to a vertical position, and the subject had to indicate whether the circle was in either of the two upper or two lower squares of the grid by pressing the upper or lower button. This presentation pattern was maintained for several repeat trials followed by a switch trial where the arrow changed back to a horizontal position. Subjects thus had to switch their attention and their response between the horizontal dimension (i.e., is the dot on the left or right side of the grid?) and the vertical dimension (i.e., is the dot on the upper or lower part of the grid?). Switch trials were separated by a minimum of four repetition trials (TR ISI 2.4 s) in order to allow optimal separation of the hemodynamic response. Switch trials appeared pseudo randomly either after 4, 5, or 6 repeat trials (i.e., every 9.6, 12, or 14.4 s) to avoid predictability. The 6‐min task consisted of 152 trials with high‐frequency repeat trials (79%) interspersed with 32 low‐frequency switch trials (21%). Thus, on average, one in five trials was a switch trial [Rubia et al., 2006; Smith et al., 2004].

Behavioral measurements include separate totals for the number of errors committed during switch trials and repeat trials as well as a measure of the switching capacity. Switching capacity is measured in terms of switch costs (switch RT–repeat RT), which reflect the (usually) slower reaction time (RT) associated with trials where a switch has to be made relative to the faster RT associated with the easier repeat trials where no switch occurs [Meiran, 1996; Rubia et al., 2006]. The event‐related analysis contrasted activation associated with correct switch trials with that of correct repeat trials.

Imaging procedures

Images were acquired on a 1.5 T Horizon LX System (GE; Milwaukee, WI) at the IOP. Structural MRI acquisition included a sagittal T1‐weighted image for orientation along the AC/PC line. This was followed by a high‐resolution inversion recovery EPI of the whole brain acquired in the inter‐commissural plane with 3.3 mm slice thickness, matrix size of 128 × 128 and voxel size of 1.875 × 1.875 (TR = 1.6 s, TE = 40 mm) for a total of 43 slices. These images were used as the anatomical overlay for functional images and for normalization purposes. Functional MRI acquisition consisted of axial images of 7.7 mm thickness, matrix size of 64 × 64, voxel size of 3.75 × 3.75 and a 90° flip angle (TR = 2.4 s matched the mean ITI; TE = 40 mm) for a total of 25 slices. A total of 152 volumes were collected over ∼ 6 min of scan time. Additional tasks completed during this course of the fMRI session included a recognition memory paradigm and a Simon task published elsewhere [Lamar et al., 2009].

Image Processing and Analyses

Image preprocessing and analyses were conducted using X‐BAM v4.0 software developed at the IoP. Images were corrected for motion, intensity, and spin excitation history [Bullmore et al., 1999a]. A mean 3D image reflecting average intensity at each time point across the entire time series was then calculated. Each image in the series was realigned to the mean image and smoothed using a 7.2 mm FWHM Gaussian filter.

Individual first‐level analyses began by convolving each aspect of the experiment with two gamma variate functions. Repeat trials were used as implicit baseline trials. After determining the best model fit [Friman et al., 2003], a goodness of fit statistic (the ratio of the sum of squares of deviation from the mean image intensity over the entire time series due to the variables in the model to that due to the residuals) was computed (SSQ ratio) at each voxel and data were permuted by a wavelet‐based method [Bullmore et al., 2001]. During this process, a null distribution of the SSQ ratios is determined and used to calculate the critical value needed to threshold each individual brain activation map (IBAM) at P < 0.01. The BOLD effect size is also calculated at this time for use in second‐level group analyses. The statistical map used in second‐level group analyses reflected correct switch trials as compared to baseline repeat trials with errors modeled separately as variables of noninterest. Across both administrations, our sample had a minimum of 44 (∼ 70%) and a maximum of 64 (100%) correct switch trials for fMRI analyses.

Second‐level group analyses required the transformation of the derived SSQ ratio and the BOLD effect size IBAMS into Talairach space [Talairach and Tournoux, 1988] via a two‐stage warping procedure [Brammer et al., 1997]. Group brain activation maps (GBAMs) were then formulated based on the median SSQ ratio at each voxel with the overall intracerebral distribution used to derive the null distribution of SSQ ratios for thresholding purposes. By combining the SSQ ratio maps from the IBAM stage with the computation of median SSQ ratio maps at the GBAM stage, we account for intra‐ and inter‐subject variation separately creating a mixed effect analysis desirable in fMRI statistical methodology. GBAMs were spatially smoothed using a 5.0 FWHM Gaussian kernel. These analyses were thresholded at a voxel‐level P‐value of 0.05 and a cluster‐level P‐value of 0.007.

We also compared the experimental conditions of sham depletion and ATD by fitting the data using a linear model approach [Bullmore et al., 1999b]. By minimizing the sum of absolute deviation as opposed to using the SSQ during model fitting, we minimized the impact of outlying data. Difference maps were computed as described above with the exception that BOLD effect maps were used to compute significant differences not the SSQ ratio to avoid the negative impact of excess noise components that these standardized measures might bring to bear during analyses at this level. Results of these contrast analyses were thresholded P = 0.05 at the voxel‐level and a cluster‐level P‐value of 0.03, allowing for less than one error cluster.

In summary, our analysis method involved implementation of a mixed effects analysis at cluster level with testing by permutation and therefore no assumption of data normality. This was consistent with the recommendations of Thirion et al. 2007. Our choice of thresholds at voxel and cluster level were chosen to control the final expected rate of type I error clusters at <1 per volume. Thus, we achieve control of cluster‐level type I error rate at whole brain level with minimal assumptions. We report only those areas of activation with greater than 10 contiguous voxels. Localization of cluster‐level peak voxels of activation was determined using the Talairach Deamon [Lancaster et al., 1997, 2000] and verified using the Talairach Atlas [Talairach and Tournoux, 1988]. Cerebellar coordinates were determined using a 3D atlas of the human cerebellum warped into Talairach space [Schmahmann et al., 1999].

We extracted the statistical power of the BOLD response (SSQ) for the peak coordinate of activation for the largest cluster(s) of activation from each repeated measures ANOVA (i.e., sham depletion relative to 5HT depletion and 5HT depletion relative to Sham depletion). We extracted SSQ information per individual per drink condition for use in Pearson's product moment correlational analyses (two‐tailed) for use in correlational analyses of activation patterns across conditions. We set our P‐value to 0.005 for these analyses to correct for multiple comparisons.

RESULTS

Biochemical Assays

There was ∼ 6‐fold decrease in total plasma tryptophan levels after ATD. After sham depletion, plasma tryptophan levels increased 2.7‐fold (see Table 2). These rates of change in 5‐HT are comparable to those observed in other adult populations [Hood et al., 2005] for review.

Table 2.

Scan day participant characteristics and Simon behavioral performance (n = 10)

ATD Scan Day Sham Scan Day
0 h 4.5 h 0 h 4.5 h
Total plasma tryptophan levels (mmol/l) 63.0 (10.1) 11.8 (3.2) 66.7 (13.6) 180.7 (47.3)
Mood Rating Scale
Factor 1: Alertness 187.5 (148.8) 231.1 (157.5)a 211.7 (167.7) 303.8 (209.9)a
Factor 2: Contentedness 77.0 (75.5) 78.3 (73.8) 97.8 (77.0) 103.0 (91.2)
Factor 3: Calmness 32.1 (29.9) 35.0 (35.4) 46.9 (36.9) 45.9 (36.6)
Aggression rating scale mean score 16.2 (9.2) 12.8 (9.5)a 16.3 (9.5) 16.6 (14.9)
Bodily Symptoms Scale b
Anxiety 12.7 (11.3) 10.7 (10.0) 16.0 (21.3) 14.5 (11.5)
Nausea 7.5 (8.0) 11.4 (12.0) 10.0 (10.2) 21.1 (23.3)
Dizziness 7.3 (6.4) 8.7 (9.1) 10.4 (9.6) 9.7 (9.4)
Irritability 8.2 (8.3) 9.4 (10.5) 11.0 (10.4) 11.1 (11.4)
Physical Tiredness 18.4 (18.9) 24.4 (23.6) 21.0 (20.1) 23.2 (21.3)
Concentration 19.4 (20.5) 23.1 (21.2) 16.2 (16.0) 23.6 (19.0)
The Switch task
Switch trial errors (out of 32 trials) 4.2 (5.4) 1.5 (1.8)
Repeat trial errors (out of 120 trials) 13.2 (20.1) 5.6 (6.6)
Switch cost RT (ms) 81.5 (105.1) 70.3 (50.3)
Open in a new tab

ATD, acute tryptophan depletion; RT, reaction time; mean (standard deviation).

a

Within scan day analysis P ≤ 0.05.

b

Select symptoms most commonly impacted by ATD.

Behavioral Questionnaires

Alertness (as measured by the MRS) was significantly decreased from baseline when measured 4.5 h after both sham and ATD drink conditions [sham: t(9) = −2.6, P = 0.03; ATD: t(9) = −2.2, P = 0.05] making any influences of this variable on Switch performance or blood oxygenation levels equally present across both scan sessions. Mean state aggression scores as measured by the ARS were lower (meaning less aggression) 4.5 h after drink ingestion when compared to baseline scores during ATD only [t(9) = 3.2, P = 0.01]. Comparisons of other behavioral questionnaires including those related to mood variables such as depression and anxiety were not statistically significant (see Table 2).

Neuropsychological Performance on the Switch Task

Given that switch trials appeared pseudo randomly either after 4, 5, or 6 repeat trials, we compared switch RTs according to the number of repeat trials between switch trials (i.e., 4, 5, or 6) to ensure that there was enough variability in 4–6 events to diminish the predictability of the timing of the switch. That is, to ensure overall Switch RT cost was not influenced by shorter or longer intervals between switches. There were no significant differences between RTs by switch event number (all P‐values ≥0.18).

Although participants made more errors to repeat and switch trials during ATD, paired sample t‐tests comparing total errors to the sham depletion condition did not reach significance (both P‐values ≥0.11). There was also no statistically significant difference between drink conditions with respect to the Switch RT Cost (P‐value = 0.75; Table 2).

Functional Brain Activation

Sham depletion

The largest cluster of significant activation during sham depletion (Table 3; Fig. 1) centered on the right middle frontal gyrus (BA10) and extended to encompass the right inferior frontal gyrus (BA47). There were additional clusters of activation within bilateral superior frontal cortices (BA10 on the left; BA11 on the right) as well as the left middle frontal gyrus (BA10). Unilateral, left prefrontal regions of activation included the medial orbitofrontal cortex (BA11) and the anterior cingulate (BA 32).

Table 3.

Group brain activations for the sham depletion and the tryptophan depletion

Location BA X Y Z Cluster size Output p‐valuea
Sham depletion
Middle frontal gyrus R 10 29 55 −7 776 0.004
Inferior frontal gyrus 47 51 48 −7 101
L 10 −43 56 −2 35
Superior frontal gyrus R 11 25 63 −18 73
L 10 −29 63 15 38
Medial orbitofrontal gyrus L 11 −25 44 −13 11
Anterior cingulate gyrus L 32 −14 44 9 28
Acute tryptophan depletion
Precuneus extending into R 7 3 −55 53 698 0.005
Superior parietal lobule 11 −67 53 102
L −25 −44 42 20
Inferior parietal lobule incl post central gyrus L 40 −43 −48 37 30
−47 −26 48 14
Angular gyrus R 39 36 −67 26 62
Cuneus R 11 −70 31 32
Open in a new tab

X,Y,Z, Talairach coordinates; R, right; L, left; incl, including.

a

Results of all analyses were thresholded P = 0.05 at the voxel‐level and P ≤ 0.007 at the cluster‐level. The output P‐value column represents the P‐value for the most significant cluster of activation in each analysis; all other listed clusters had P‐values ≤0.007.

Figure 1.

Figure 1

Open in a new tab

Group brain activation patterns for sham depletion (top panel) and tryptophan depletion (bottom panel) at a voxel threshold of P < 0.05 and a cluster threshold of P < 0.007; only clusters of 10 or more activated voxels are reported in the manuscript.

5HT depletion

Switch performance during ATD was associated with significant activation of the right precuneus (BA7) extending into the superior parietal lobule. The left precuneus was also activated during ATD as was the left inferior parietal lobule (BA40). Additional clusters of activation included the right angular gyrus (BA39) and right cuneus (Table 3; Fig. 1).

Sham depletion relative to 5HT depletion

The largest cluster of activation during Switch performance under sham depletion relative to ATD involved the right medial orbitofrontal cortex (Table 4; Fig. 2). Right superior (BA11) and right middle frontal (BA10/47) gyri also showed significant activation during sham relative to ATD.

Table 4.

Repeated measures ANOVA between the two drink conditions for the Switch task

Location BA X Y Z Cluster size Output P‐valuea
Sham depletion > Acute tryptophan depletion
Medial orbitofrontal cortex R 36 52 −12 149 0.008
Superior frontal gyrus R 11 36 48 −18 42
Middle frontal gyrus R 10/47 43 48 −7 19
Acute tryptophan depletion > Sham depletion
Precuneus R 7 3 −70 42 928 0.001
L 19 −7 −81 42 71
Cuneus R 19 11 −78 26 43
L 19 −14 −78 31 104
Superior parietal lobule incl post central gyrus L −25 −48 59 24
−25 −48 64 26
Lingual gyrus R 18 4 −93 −7 96
17 0 −93 −2 62
Fusiform gyrus R 37 51 −56 −18 56
Cerebellum R
Crus I–Superior semilunar lobule 51 −56 −29 37
VI–Quadrangular Lobule
Posterior Portion: Declive 18 −67 −13 58
Tonsil 40 −44 −35 18
Open in a new tab

X,Y,Z, Talairach coordinates; R, right; L, left; incl, including.

a

Results of all analyses were thresholded P = 0.05 at the voxel‐level and P = 0.03 at the cluster‐level. The output P‐value column represents the P‐value for the most significant cluster of activation in each analysis; all other listed clusters had P‐values <0.03.

Figure 2.

Figure 2

Open in a new tab

Brain activation patterns resulting from repeated measures ANOVAs for Sham depletion > tryptophan depletion (top panel) and tryptophan depletion > sham depletion (bottom panel).

5HT depletion relative to Sham depletion

The majority of increased activation during ATD relative to sham depletion occurred in posterior brain regions (Table 4; Fig. 2). The largest cluster of activation involved the right precuneus (BA7) and there was a smaller cluster of activation in the left precuneus (BA19) and in the cuneus bilaterally (BA19). Regions of the left superior parietal lobule also contained a cluster of significant activation. Clusters of activation were also detected in the right lingual (BA17&18) and the right fusiform (BA37) gyri. Regions of activation within the cerebellum included a cluster of activation in Crus I encompassing the right superior region of the semilunar lobule and a cluster of activation within Quadrangular Lobule VI involving the right posterior declive. A smaller cluster of activation was also evident within the right cerebellar tonsil.

Associations of the Power of the BOLD Response as Measured by the Extracted SSQ Information

5HT and sham depletion peak voxels

We correlated the statistical power of the BOLD responses within the right medial orbitofrontal cortex (the largest cluster of activation for sham relative to 5HT depletion) with the statistical power of the BOLD responses within the right precuneus (the largest cluster of activation for 5HT relative to sham depletion) to determine if there was a negative association between these regions of activation during the ATD drink condition (Fig. 3). Results of this analysis revealed a significant negative association between the SSQ ratio for the right medial orbitofrontal cortex and the SSQ ratio for the right precuneus during ATD, r(10)= −0.92, P < 0.001. These regions were not significantly correlated during sham depletion, r(10) = 0.16, P = 0.64.

Figure 3.

Figure 3

Open in a new tab

Excerpts of the statistical power of BOLD response (SSQ; y‐axis) for the peak coordinate of activation for each contrast analysis by drink condition (top panel = sham depletion relative to ATD showing right medial orbitofrontal cortical activation; bottom panel = ATD relative to sham depletion showing right precuneus activation).

DISCUSSION

We observed that ATD during task switching performance is associated with a significant anterior‐to‐posterior shift in brain activity—but not a change in behavioral performance—in older women. Thus, ATD reduced prefrontal and anterior cingulate activation and increased posterior regions of activation including parietal and neocerebellar cortices during successful switches between stimulus‐response association sets. These findings support our a priori hypothesis of an anterior‐to‐posterior shift of activation for task switching in older adults. Although this should not be seen as an independent replication, the current study extends our previous ATD findings beyond a cognitive interference inhibition task [Lamar et al., 2009] to another aspect of executive function. Also, our correlational analyses provide preliminary evidence that this shift in activation (i.e., increases in posterior recruitment associated with decreases in anterior recruitment during ATD) may reflect a compensation for reduced prefrontal activation secondary to the modulation of 5‐HT in the aging brain.

ATD was also accompanied by an increase in recruitment of more posterior brain regions in our older women. For example, during ATD relative to sham depletion posterior brain regions including the superior parietal lobule, lingual, and fusiform gyri as well as the neocerebellum showed increased activation. These brain regions share connections with the typically recruited prefrontal and anterior cingulate regions implicated in successful task switching under age‐appropriate levels of 5‐HT [DiGirolamo et al., 2001; Townsend et al., 2006]. Thus, task switching, mediated by complex fronto‐parietal, fronto‐striatal, and fronto‐cerebellar networks, relied on fronto‐striatal regions of activation during sham depletion in our aging cohort but became more dependent upon parietal/cerebellar regions after ATD. Taken together, this suggests an anterior‐to‐posterior shift in the recruitment of regions that form the larger fronto‐parietal, fronto‐striatal, and fronto‐cerebellar switching network when already low‐levels of 5‐HT in the aging brain are depleted further using ATD. We suggest that the increased recruitment of posterior parietal and neocerebellar regions during the Switch task allowed older adults to compensate for loss of activation in task‐related frontal regions of the brain. This increased reliance on alternative brain regions appears to have preserved switch task performance in our older adults.

Several issues must be taken into consideration when considering the results of this study. Age‐related reductions in 5‐HT within prefrontal and striatal regions [Goldberg et al., 2004] provide a lower baseline from which ATD works to decrease tryptophan levels; however, the generic impact of ATD on neurovasculature and comparisons to younger populations must also be considered. Future studies investigating these topic areas are needed; although the scant evidence on alterations in regional cerebral blood flow during tryptophan depletion [Talbot and Cooper, 2006] does not directly implicate regions found significantly altered by ATD in the current study. Additionally, differences in brain function following ATD may not just represent modulation of 5‐HT but downstream effects on other neurotransmitter systems [van Donkelaar et al., 2011]; however, the literature suggests that the overwhelming effect of ATD is most likely directly on the 5‐HT system [Young et al., 1989]. Our data regarding free tryptophan levels would suggest a certain degree of 5‐HT “loading” during the sham condition; however, this is unlikely the case given that ratio measures of tryptophan to other amino acids typically shows minimal increase during the sham condition [e.g., Sambeth et al., 2009]. While we lack requisite data to report ratio measures in our sample, previous studies provide clarification of the sham plasma data [Sambeth et al., 2009].

A limitation of this study may be its small sample size which only allows for detection of large effect sizes. While minimum numbers of 15–20 participants have been suggested for fMRI studies [Thirion et al., 2007], repeated measures designs like that used in the current within‐group double‐blind sham‐controlled crossover study are statistically more powerful than independent datasets, which makes the within‐subject ANOVA more robust. Despite marked changes in brain activation with ATD, we did not find effects on Switch task performance. This may be due to relatively small power for neuropsychological studies, although previous studies have not shown an effect of ATD on performance in switching tasks [see Mendelsohn et al., 2009 for review]. While we cannot rule out the possibility of a Type II error in this instance, our findings of marked brain changes despite a lack of performance effects are in line with the previously reported superior sensitivity of fMRI to capture adaptive physiological change not manifested in behavioral performance during pharmacological challenge studies [Honey and Bullmore, 2004; Lamar et al., 2009; Rubia et al., 2005].

In summary, our results suggest that in older adults ATD reduces task‐relevant prefrontal/cingulate activation but increases activation in other more posterior (parietal and neocerebellar) brain regions of the switching network to maintain successful performance during task switching. This anterior‐to‐posterior shift may reflect a functional compensation mechanism employed by older adults when portions of the network (e.g., fronto‐striatal) are disabled by the negative impact of ATD. This effect has previously been shown in the same sample of older adults during an interference inhibition task [Lamar et al., 2009], suggesting that the anterior‐to‐posterior activation shift with ATD may be an effect that transcends several executive function tasks. Future ATD studies in aging should assess whether this effect can also be observed for other cognitive domains (e.g., learning and memory) and other (larger) aging cohorts. In addition to extending these ATD findings to other aspects of cognition in normal aging, ATD may prove useful in determining cognitive vulnerability in older adults with a predisposition for serontonergic down‐regulation (i.e., individuals with vascular and/or late life depression) or cerebellar dysfunction.

Acknowledgments

The authors are grateful to Dr. Kate John and Dr. Roy Sherwood in the Department of Clinical Biochemistry at King's College Hospital, London for the analysis of the tryptophan levels. They would also like to thank the participants of this study and Dr. Anand Kumar in the Department of Psychiatry at the University of Illinois at Chicago for their support of this work.

REFERENCES

  1. American Psychiatric Association (1994):DSM‐IV: Diagnostic and statistical manual of mental disorders, 4th ed. Washington, D.C.:American Psychiatric Press. [Google Scholar]
  2. Arranz B, Eriksson A, Mellerup E, Plenge P, Marcusson J (1993): Effect of aging in human cortical pre‐ and postsynaptic serotonin binding sites. Brain Res 620:163–166. [DOI] [PubMed] [Google Scholar]
  3. Beck AT, Epstein N, Brown G, Steer RA (1988): An inventory for measuring clinical anxiety: Psychometric properties. J Consult Clin Psychol 56:893–897. [DOI] [PubMed] [Google Scholar]
  4. Beck AT, Steer RA (1993):Manual for the beck depression inventory.San Antonio, Texas:Psychological Corporation. [Google Scholar]
  5. Bjork J M, Dougherty DM, Moeller FG, Cherek DR, Swann AC. (2000): Differential behavioral effects of plasma tryptophan depletion and loading in aggressive and nonaggressive men. Neuropsychopharmacology 22:357–369. [DOI] [PubMed] [Google Scholar]
  6. Bond A, Lader M (1974): The use of analogue scales in rating subjective feelings. Br J Med Psychol 47:211–218. [Google Scholar]
  7. Bond A, Lader M (1986): A method to elicit aggressive feelings and behavior via provocation. Biol Psychol 22:67–79. [DOI] [PubMed] [Google Scholar]
  8. Brammer MJ, Bullmore ET, Simmons A, Williams SC, Grasby PM, Howard RJ, Woodruff PW, Rabe‐Hesketh S (1997): Generic brain activation mapping in functional magnetic resonance imaging: A nonparametric approach. Magn Reson Imaging 15:763–770. [DOI] [PubMed] [Google Scholar]
  9. Bullmore ET, Brammer MJ, Rabe‐Hesketh S, Curtis VA, Morris RG, Williams SC, Sharma T, McGuire PK (1999a).Methods for diagnosis and treatment of stimulus‐correlated motion in generic brain activation studies using fMRI. Hum Brain Mapp 7:38–48. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bullmore ET, Long C, Suckling J, Fadili J, Calvert G, Zelaya F, et al. (2001): Colored noise and computational inference in neurophysiological (fMRI) time series analysis: Resampling methods in time and wavelet domains. Hum Brain Mapp 12:61–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Bullmore ET, Suckling J, Overmeyer S, Rabe‐Hesketh S, Taylor E, Brammer MJ (1999b): Global, voxel, and cluster tests, by theory and permutation, for a difference between two groups of structural MR images of the brain. IEEE Trans Med Imaging 18:32–42. [DOI] [PubMed] [Google Scholar]
  12. Buss AH, Perry M (1992): The aggression questionnaire. J Pers Soc Psychol 63:452–459. [DOI] [PubMed] [Google Scholar]
  13. Christakou A, Halari R, Smith AB, Ifkovits E, Brammer M, Rubia K (2009): Sex‐dependent age modulation of frontostriatal and temporo‐parietal activation during cognitive control. Neuroimage 48:223–236. [DOI] [PubMed] [Google Scholar]
  14. Cleare AJ, Bond AJ (1995): The effect of tryptophan depletion on subjective and behavioral agression in normal male subjects. Psychopharmacology 118:72–81. [DOI] [PubMed] [Google Scholar]
  15. Dale AM (1999): Optimal experimental design for event‐related fMRI. Hum Brain Mapp 8:109–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. DiGirolamo GJ, Kramer AF, Barad V, Cepeda NJ, Weissman DH, Milham MP, Wszalek TM, Cohen NJ, Banich MT, Webb A, Belopolsky AV, McAuley E (2001): General and task‐specific frontal lobe recruitment in older adults during executive processes: A fMRI investigation of task‐switching. Neuroreport 12:2065–2071. [DOI] [PubMed] [Google Scholar]
  17. Dove A, Pollmann S, Schubert T, Wiggins CJ, von Cramon DY (2000): Prefrontal cortex activation in task switching: An event‐related fMRI study. Brain Res Cogn Brain Res 9:103–109. [DOI] [PubMed] [Google Scholar]
  18. Evers EA, Sambeth A, Ramaekers JG, Riedel WJ, van der Veen FM (2010): The effects of acute tryptophan depletion on brain activation during cognition and emotional processing in healthy volunteers. Curr Pharm Des 16:1998–2011. [DOI] [PubMed] [Google Scholar]
  19. Folstein MR, Folstein SE, McHugh PR (1974): Mini‐mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12:189–198. [DOI] [PubMed] [Google Scholar]
  20. Friman O, Borga M, Lundberg P, Knutsson H (2003): Adaptive analysis of fMRI data. Neuroimage 19:837–845. [DOI] [PubMed] [Google Scholar]
  21. Goldberg S, Smith GS, Barnes A, Ma Y, Kramer E, Robeson K, Kirshner M, Pollock BG, Eidelberg D (2004): Serotonin modulation of cerebral glucose metabolism in normal aging. Neurobiol Aging 25:167–174. [DOI] [PubMed] [Google Scholar]
  22. Gruber O, Diekhof EK, Kirchenbauer L, Goschke T (2010): A neural system for evaluating the behavioural relevance of salient events outside the current focus of attention. Brain Res 1351:212–221. [DOI] [PubMed] [Google Scholar]
  23. Gruber O, Melcher T, Diekhof EK, Karch S, Falkai P, Goschke T (2009): Brain mechanisms associated with background monitoring of the environment for potentially significant sensory events. Brain Cogn 69:559–564. [DOI] [PubMed] [Google Scholar]
  24. Hedden T, Gabrieli JD (2004): Insights into the ageing mind: A view from cognitive neuroscience. Nat Rev Neurosci 5:87–96. [DOI] [PubMed] [Google Scholar]
  25. Higgitt A, Lader M, Fonagy P (1986): The effects of the benzodiazepine antagonist Ro 15‐1788 on psychophysiological performance and subjective measures in normal subjects. Psychopharmacology (Berlin) 89:395–403. [DOI] [PubMed] [Google Scholar]
  26. Honey G, Bullmore E (2004): Human pharmacological MRI. Trends Pharmacol Sci 25:366–374. [DOI] [PubMed] [Google Scholar]
  27. Hood SD, Bell CJ, Nutt DJ (2005): Acute tryptophan depletion. I. Rationale and methodology. Aust N Z J Psychiatry 39:558–564. [DOI] [PubMed] [Google Scholar]
  28. Keating J, Dratcu L, Lader M, Sherwood RA (1993): Measurement of plasma serotonin by high‐performance liquid chromatography with electrochemical detection as an index of the in vivo activity of fluvoxamine. J Chromatogr 615:237–242. [DOI] [PubMed] [Google Scholar]
  29. Lamar M, Cutter WJ, Rubia K, Brammer M, Daly EM, Craig MC, Cleare AJ, Murphy DG (2009): 5‐HT, prefrontal function and aging: fMRI of inhibition and acute tryptophan depletion. Neurobiol Aging 30:1135–1146. [DOI] [PubMed] [Google Scholar]
  30. Lancaster JL, Summerlin JL, Rainey L, Freitas CS, Fox PT (1997): The talairach daemon: A database server for Talairach Atlas labels. Neuroimage 5:s633. [Google Scholar]
  31. Lancaster JL, Woldorff MG, Parsons LM, Liotti M, Freitas CS, Rainey L, Kochunov PV, Nickerson D, Mikiten SA, Fox PT (2000): Automated Talairach atlas labels for functional brain mapping. Hum Brain Mapp 10:120–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Loose R, Kaufmann C, Tucha O, Auer DP, Lange KW (2006)Neural networks of response shifting: Influence of task speed and stimulus material. Brain Res 1090:146–155. [DOI] [PubMed] [Google Scholar]
  33. Meiran N (1996)Reconfiguration of processing mode prior to task performance. J Exp Psychol: Learn Mem Cogn 22:1423–1442. [Google Scholar]
  34. Mendelsohn D, Riedel WJ, Sambeth A (2009): Effects of acute tryptophan depletion on memory, attention and executive functions: A systematic review. Neurosci Biobehav Rev 33:926–952. [DOI] [PubMed] [Google Scholar]
  35. Nishizawa S, Benkelfat C, Young SN, Leyton M, Mzengeza S, de Montigny C, Blier P, Diksic M (1997): Differences between males and females in rates of serotonin synthesis in human brain. Proc Natl Acad Sci U S A 94:5308–5313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Raz N, Ghisletta P, Rodrigue KM, Kennedy KM, Lindenberger U (2010): Trajectories of brain aging in middle‐aged and older adults: Regional and individual differences. Neuroimage 51:501–511. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Raz N, Gunning FM, Head D, Dupuis JH, McQuain J, Briggs SD, Loken WJ, Thornton AE, Acker JD (1997): Selective aging of the human cerebral cortex observed in vivo: Differential vulnerability of the prefrontal gray matter. Cereb Cortex 7:268–282. [DOI] [PubMed] [Google Scholar]
  38. Resnick SM, Pham DL, Kraut MA, Zonderman AB, Davatzikos C (2003): Longitudinal magnetic resonance imaging studies of older adults: A shrinking brain. J Neurosci 23:3295–3301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Rubia K, Lee F, Cleare AJ, Tunstall N, Fu CH, Brammer M, McGuire P (2005): Tryptophan depletion reduces right inferior prefrontal activation during response inhibition in fast, event‐related fMRI. Psychopharmacology (Berlin) 179:791–803. [DOI] [PubMed] [Google Scholar]
  40. Rubia K, Smith AB, Woolley J, Nosarti C, Heyman I, Taylor E, Brammer M (2006): Progressive increase of frontostriatal brain activation from childhood to adulthood during event‐related tasks of cognitive control. Hum Brain Mapp 27:973–993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Sambeth A, Riedel WJ, Tillie DE, Blokland A, Postma A, Schmitt JA (2009): Memory impairments in humans after acute tryptophan depletion using a novel gelatin‐based protein drink. J Psychopharmacol 23:56–64. [DOI] [PubMed] [Google Scholar]
  42. Schmahmann JD, Doyon J, McDonald D, Holmes C, Lavoie K, Hurwitz AS, Kabani N, Toga A, Evans A, Petrides M (1999): Three‐dimensional MRI atlas of the human cerebellum in proportional stereotaxic space. Neuroimage 10(3 Pt 1):233–260. [DOI] [PubMed] [Google Scholar]
  43. Smith AB, Taylor E, Brammer M, Rubia K (2004): Neural correlates of switching set as measured in fast, event‐related functional magnetic resonance imaging. Hum Brain Mapp 21:247–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Smith EE, Geva A, Jonides J, Miller A, Reuter‐Lorenz P, Koeppe RA (2001): The neural basis of task‐switching in working memory: Effects of performance and aging. Proc Natl Acad Sci U S A 98:2095–2100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Smith GS, Workman CI, Kramer E, Hermann CR, Ginsberg R, Ma Y, Dhawan V, Chaly T, Eidelberg D (2011): The relationship between the acute cerebral metabolic response to citalopram and chronic citalopram treatment outcome. Am J Geriatr Psychiatry 19:53–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Talairach J, Tournoux P (1988):Co‐Planar Stereotaxic Atlas of the Human Brain. 3‐Dimensional Proportional System: An Approach to Cerebral Imaging (M. Rayport, Trans.).New York:Thieme Medical Publishers. [Google Scholar]
  47. Talbot PS, Cooper SJ (2006): Anterior cingulate and subgenual prefrontal blood flow changes following tryptophan depletion in healthy males. Neuropsychopharmacology 31:1757–1767. [DOI] [PubMed] [Google Scholar]
  48. Thirion B, Pinel P, Meriaux S, Roche A, Dehaene S, Poline JB. (2007): Analysis of a large fMRI cohort: Statistical and methodological issues for group analyses. Neuroimage 35:105–120. [DOI] [PubMed] [Google Scholar]
  49. Townsend J, Adamo M, Haist F. (2006): Changing channels: An fMRI study of aging and cross‐modal attention shifts. Neuroimage 31:1682–1692. [DOI] [PubMed] [Google Scholar]
  50. van Donkelaar EL, Blokland A, Ferrington L, Kelly PA, Steinbusch HW, Prickaerts J (2011): Mechanism of acute tryptophan depletion: Is it only serotonin? Mol Psychiatry 16:695–713. [DOI] [PubMed] [Google Scholar]
  51. van Dyck CH, Malison RT, Seibyl JP, Laruelle M, Klumpp H, Zoghbi SS, Baldwin RM, Innis RB (2000): Age‐related decline in central serotonin transporter availability with [(123)I]beta‐CIT SPECT. Neurobiol Aging 21:497–501. [DOI] [PubMed] [Google Scholar]
  52. Wecker NS, Kramer JH, Hallam BJ, Delis DC (2005): Mental flexibility: Age effects on switching. Neuropsychology 19:345–352. [DOI] [PubMed] [Google Scholar]
  53. Yeung N, Nystrom LE, Aronson JA, Cohen JD (2006): Between‐task competition and cognitive control in task switching. J Neurosci 26:1429–1438. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Young SN, Ervin FR, Pihl RO, Finn P (1989): Biochemical aspects of tryptophan depletion in primates. Psychopharmacology (Berlin) 98:508–511. [DOI] [PubMed] [Google Scholar]
  55. Young SN, Pihl RO, Ervin FR (1988)The effect of altered tryptophan levels on mood and behavior in normal human males. Clin Neuropharmacol 11 (Suppl 1):S207–S215. [PubMed] [Google Scholar]
  56. Young SN, Smith SE, Pihl RO, Ervin FR (1985)Tryptophan depletion causes a rapid lowering of mood in normal males. Psychopharmacology (Berlin) 87:173–177. [DOI] [PubMed] [Google Scholar]

Articles from Human Brain Mapping are provided here courtesy of Wiley

RESOURCES