Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: Psychoneuroendocrinology. 2020 May 19;119:104710. doi: 10.1016/j.psyneuen.2020.104710

Impact of Childhood Adversity on Network Reconfiguration Dynamics During Working Memory in Hypogonadal Women

Sheila Shanmugan a,b,*, Wen Cao a, Theodore D Satterthwaite a, Mary D Sammel b,c,d, Arian Ashourvan e,f, Danielle S Bassett e,f,g,h, Kosha Ruparel a, Ruben C Gur a, C Neill Epperson i, James Loughead a
PMCID: PMC7745207  NIHMSID: NIHMS1605121  PMID: 32563173

Abstract

Many women with no history of cognitive difficulties experience executive dysfunction during menopause. Significant adversity during childhood negatively impacts executive function into adulthood and may be an indicator of women at risk of a mid-life cognitive decline. Previous studies have indicated that alterations in functional network connectivity underlie these negative effects of childhood adversity. There is growing evidence that functional brain networks are not static during executive tasks; instead, such networks reconfigure over time. Optimal dynamics are necessary for efficient executive function; while too little reconfiguration is insufficient for peak performance, too much reconfiguration (supra-optimal reconfiguration) is also maladaptive and associated with poorer performance. Here we examined the impact of adverse childhood experiences (ACEs) on network flexibility, a measure of dynamic reconfiguration, during a letter n-back task within three networks that support executive function: frontoparietal, salience, and default mode networks. Several animal and human subject studies have suggested that childhood adversity exerts lasting effects on executive function via serotonergic mechanisms. Tryptophan depletion (TD) was used to examine whether serotonin function drives ACE effects on network flexibility. We hypothesized that ACE would be associated with higher flexibility (supra-optimal flexibility) and that TD would further increase this measure. Forty women underwent functional imaging at two time points in this double-blind, placebo controlled, crossover study. Participants also completed the Penn Conditional Exclusion Test, a task assessing abstraction and mental flexibility. The effects of ACE and TD were evaluated using generalized estimating equations. ACE was associated with higher flexibility across networks (frontoparietal β=0.00748, D=2.79, p=0.005; salience β=0.00679, D=3.02, p=0.003; and default mode β=0.00910, D=3.53, p=0.0004). While there was no interaction between ACE and TD, active TD increased network flexibility in both ACE groups in comparison to sham depletion (frontoparietal β=0.00489, D=2.15, p=0.03; salience β=0.00393, D=1.91, p=0.06; default mode β=0.00334, D=1.73, p=0.08). These results suggest that childhood adversity has lasting impacts on dynamic reconfiguration of functional brain networks supporting executive function and that decreasing serotonin levels may exacerbate these effects.

Keywords: Adverse childhood experiences, executive function, dynamic connectivity, tryptophan depletion, menopause

1. Introduction

During the menopause transition, many women with no history of cognitive dysfunction experience new-onset difficulties with everyday tasks requiring sustained attention, motivation for work, working memory, organization, and planning (Epperson et al., 2011; Shanmugan and Epperson, 2014; Shanmugan et al., 2017b). Prior work has suggested that adverse childhood experiences (ACEs) may be a risk factor for executive difficulties during this period of waning estradiol (Shanmugan et al., 2017a; Shanmugan et al., 2017c). ACEs are highly prevalent in the general population and have repeatedly been associated with worse mental health-related outcomes in adults including substance dependence and depressive disorders (Chapman et al., 2004). Childhood adversity is also associated with poorer executive function in healthy adults (Philip et al., 2016). Surgically-menopausal women with a history of high levels of childhood adversity endorse more symptoms of executive dysfunction and perform worse on tasks requiring sustained attention and working memory than those with low levels of childhood adversity (Shanmugan et al., 2020). ACEs also negatively impact brain networks broadly engaged during executive tasks in menopausal women (Shanmugan et al., 2017c).

Recent evidence has suggested that functional brain networks are not static and, instead, dynamically reconfigure during executive tasks (Braun et al., 2016; Braun et al., 2015). Optimal dynamic reconfiguration of network modules is required for proper executive functioning (Braun et al., 2015). One measure that has been used to quantify this dynamic reconfiguration is network “flexibility,” a metric measuring the extent to which network modules vary across time. While too little flexibility is insufficient for optimal executive function, too much flexibility (or supra-optimal flexibility) is also maladaptive; supra-optimal increases in network flexibility during the n-back task have been associated with poorer set-shifting abilities and may be an indicator of genetic risk for psychiatric disorders with executive deficits such as schizophrenia (Braun et al., 2016).

We examined the impact of ACEs on dynamic reconfiguration of three networks involved in executive function: frontoparietal, salience, and default mode networks. We focused on these networks because prior work has demonstrated that executive functioning capacity depends primarily on the dynamics of nodes within these particular networks (Braun et al., 2015). Early adversity has been associated with lower within-network connectivity across networks (Shanmugan et al., 2017c). This lack of network segregation may make nodes more susceptible to switching network affiliation. We therefore hypothesized that flexibility would be greater in high ACE individuals in the three networks examined. Prior studies have suggested that childhood adversity exerts lasting effects on executive function through serotonergic mechanisms (Shanmugan et al., 2017a), so we sought to understand whether serotonin was also involved in ACE effects on network flexibility. We used tryptophan depletion (TD) to lower central serotonin levels and hypothesized that depleting tryptophan would increase flexibility in each network to a greater extent in the high ACE group whose network architecture may be more vulnerable from the effects of childhood adversity. To aid in interpretation of results, we then related network flexibility to performance on the Penn Conditional Exclusion Test, a task assessing abstraction and mental flexibility.

2. Materials and Methods

Methods regarding participant inclusion/exclusion criteria, study design, assessments of childhood adversity and mood, in scanner task paradigm, and image acquisition and processing were as previously described (Shanmugan et al., 2017c) and are summarized briefly below. Because previous work from our lab has suggested that the effects of ACE and TD on executive function may be greatest when female sex hormone levels are suppressed (Shanmugan et al., 2017a), our sample consisted of healthy hypogonadal women ages 48–60. Each participant underwent 2 imaging sessions: active TD or sham TD. The order of condition (active TD versus sham TD) was counter-balanced and double-blind. The Adverse Childhood Experiences (ACE) Questionnaire (Felitti et al., 1998) was used to assess history of abuse, neglect, and household dysfunction. Consistent with prior studies, subjects with an ACE score ≥2 were considered ‘high ACE’ while subjects with ACE score of < 2 were considered ‘low ACE’ (Chapman et al., 2004; Epperson et al., 2017; Shanmugan et al., 2017a; Shanmugan et al., 2017c). The final sample included in our analyses consisted of 24 low ACE participants (43 sessions) and 16 high ACE participants (30 sessions).

A letter version of the n-back task with 4 condition blocks (0-back, 1-back, 2-back, and 3-back) was used to probe working memory during fMRI scans. Prior to each fMRI scan, subjects completed the Penn Conditional Exclusion Test (CET) to assess abstraction and mental flexibility. In each trial, participants were presented with four figures and required to select the figure that does not belong based on one of three categories or sorting principles. Behavioral outcomes of interest were CET number of categories solved, number of trials, and total number of errors. Higher number of categories achieved, fewer trials, and fewer total errors are indicative of better performance. ACE and TD effects on n-back performance were previously reported (Shanmugan et al., 2017a).

2.1. Image Acquisition

Imaging data were acquired on a 3T Siemens Trio scanner. A magnetization-prepared, rapid acquisition gradient echo T1-weighted image (TR=1810ms, TE=3.51ms, FOV=180×241mm, matrix = 192 × 256, 160 slices, effective voxel resolution of 0.94×0.94×1mm) was acquired to aid spatial normalization to standard space. Functional images were acquired using a whole-brain, single-shot gradient-echo (GE) echoplanar sequence with the following parameters: TR/TE = 3000/32 ms, FOV = 192 × 192 mm, matrix = 64 × 64, slice thickness/gap = 3/0 mm, 46 slices, voxel resolution of 3 × 3 × 3 mm.

2.2. Network Construction and visualization

We examined the effects of ACE and TD on dynamic functional connectivity using a system of functional networks described by Power et al., in which networks are composed of 264 nodes and the connections between these nodes (Power et al., 2011). This node system provides good coverage of the whole brain (Power et al., 2011) and has been used to examine functional connectivity between and within brain networks during cognitive tasks and at rest (Dosenbach et al., 2010; Satterthwaite et al., 2013a). We focused specifically on the frontoparietal, salience, and default mode networks because prior work has demonstrated that working memory capacity depends on the flexibility of nodes within these particular networks (Braun et al., 2015).

2.3. Image registration

Subject-level functional and anatomical T1 images were co-registered using boundary-based registration (Greve and Fischl, 2009). The anatomical image was normalized to the MNI 152 T1 1 mm template using the top-performing diffeomorphic SyN registration in ANTs (Avants et al., 2011; Klein et al., 2009). Co-registration, normalization, and down-sampling of network nodes to subject space were concatenated so only one interpolation was performed.

2.4. Image processing

Time-series were despiked, motion-corrected (24 parameter), band pass filtered, spatially smoothed (6 mm FWHM), and mean-based intensity normalized. A band pass filter of 0.06–0.12 Hz was selected for consistency with prior work examining flexibility of functional brain networks during cognitive tasks (Bassett et al., 2011). This range represents a balance between capturing the task-related, low frequency components of fMRI signal that contribute to connectivity (Sun et al., 2004) while excluding higher frequency signals that are more susceptible to motion artifact (Satterthwaite et al., 2013a). Residuals of this analysis modeling task and motion served as time-series for the network analysis described below. Task regression ensures that connectivity matrices are not dominated by block-related coactivation (Braun et al., 2015), is consistent with prior literature (Braun et al., 2015; Cole et al., 2014), and improves test-retest reliability of connectivity measures (Cao et al., 2014). See Shanmugan et al., 2017b for further detail on image acquisition and preprocessing.

2.5. Network Connectivity and Community Identification

We examined the effects of ACE and TD on dynamic functional connectivity using a system of functional networks described by Power et al., in which networks are composed of 264 nodes and the connections between these nodes (Power et al.,2011). We focused specifically on the frontoparietal, salience, and default mode networks because prior work has demonstrated that working memory capacity depends on the flexibility of nodes within these particular networks (Braun et al., 2015). For each participant, functional connectivity between each pair of network nodes was estimated using a wavelet coherence between time-series calculated within overlapping sliding windows (window length was 15 volumes, step size was 3 volumes) (Braun et al., 2015). These calculations yielded matrices describing the coherence between network nodes across windows. One hundred optimizations of the generalized Louvain multilayer community detection algorithm were used to partition these matrices into time-dependent communities (temporal and topological resolution parameter values were set to γ=1, ω=1) (Bassett et al., 2013; Jeub, 2011–2017). A time-dependent network flexibility matrix was calculated at each optimization, which indicated whether a node changed its community between two time windows (Mucha et al., 2010). The flexibility of each node was represented by the number of times that the node changed community assignment across the task, normalized by the total number of changes possible (Bassett et al., 2011). Averaging these measures of flexibility across 100 optimizations and across nodes within each network yielded measures of frontoparietal, salience, and DMN flexibility for each subject (Braun et al., 2016; Braun et al., 2015).

2.6. Statistical analysis

Generalized estimating equations implemented using the geepack package (Højsgaard et al., 2006) in R were employed to evaluate the effect of ACE, TD, and their interaction on CET performance and network flexibility. This method accommodates multiple assess- ments per participant and adjusts for non-independence among repeated measures. The interaction term was not significant and was removed from all models. Models controlled for age, estradiol level, and time since last menstrual period. Mean relative displacement (Satterthwaite et al., 2013b) was included as a covariate to minimize the impact of motion on flexibility. Given the potential bias of correlating brain signal with in-scanner task performance, generalized estimating equations were used to examine associations between network flexibility and CET performance. Prior literature has suggested that there may be a U shaped relationship between flexibility and executive function (Braun et al., 2016). We therefore evaluated both quadratic and linear fits for the relationship between flexibility and performance on the CET. Statistical tests were two-sided and we considered p = 0.05 to be significant. P-values reported have not been corrected for multiple comparisons. Regression coefficients (β) and effect sizes (D) are also reported.

3. Results

3.1. Behavioral results

The high ACE group required a high number of trials to solve the categorization on the CET in comparison to the low ACE group, indicating a trend for poorer performance in the high ACE group (β=7.18, D=1.61, p=0.1). Active tryptophan depletion was associated with worse performance in comparison to sham depletion, as indicated by fewer categories achieved (β=−0.38, D=−2.39, p=0.02) and more errors (β=5.88, D=1.82, p=0.07). There were no other significant or trend effects of ACE or TD on out-of-scanner task performance (Table 1).

Table 1.

ACE and TD Effects on Behavior

Beta Estimate Effect Size Mean (Low ACE or Sham TD) SD (Low ACE or Sham TD) Mean (High ACE or Active TD) SD (High ACE or Active TD) p
Total Trials
ACE group 7.18 1.61 54.5 17.3 62.0 22.9 0.1
TD status 1.36 0.25 56.3 20.0 58.7 20.0 0.8
Total Errors
ACE group 3.70 1.07 17.5 13.9 21.3 13.2 0.3
TD status 5.88 1.82 16.09 10.71 22.0 15.7 0.07
Categories Achieved
ACE group −0.28 −1.35 2.87 0.52 2.58 0.90 0.2
TD status −0.38 −2.39 2.94 0.24 2.56 0.95 0.02*
Open in a new tab

Unadjusted means are displayed. Tests of significance also controlled for age, time since last menstraul period, and estradiol level.

*

p≤0.05.

ACE=Adverse Childhood Experieces. TD=Tryptophan Depletion. SD=Standard Deviation

3.2. Impact of ACE and TD on dynamic network reconfiguration

To test our hypothesis that ACE would be associated with greater reconfiguration of the frontoparietal, salience, and default mode networks, we used a model that evaluated the impact of ACE group,TD status, and ACE group × TD status on network flexibility. We used tryptophan depletion to evaluate whether differences in serotonin function underlie ACE effects on dynamic reconfiguration of brain networks underlying executive function. The interaction between ACE and TD was not significant and was removed from the model. The main effect of ACE in the reduced models was robust. Flexibility was significantly higher in the high ACE group than in the low ACE group across networks (frontoparietal β=0.00748, D=2.79, p=0.005; salience β=0.00679, D=3.02, p=0.003; and default mode β=0.00910, D=3.53, p=0.0004; Figure 2a–c; Table 2). We found that active TD increased flexibility in both ACE groups across networks (frontoparietal β=0.00489, D=2.15, p=0.03; salience β=0.00393, D=1.91, p=0.06; default mode β=0.00334, D=1.73, p=0.08). Using a similar model that evaluated the effect of total ACE score (continuous) rather than ACE group yielded similar results. Total ACE score was positively associated with flexibility in the frontoparietal network (β=0.00203, D= 2.96, p=0.003), salience network (β=0.00161, D=2.3, p=0.02), and DMN (β= 0.00259, D= 3.63, p=0.0003; Figure 2d–f ).

Figure 2. ACE Effects on Dynamic Network Reconfiguration.

Figure 2.

Open in a new tab

Flexibility was significantly higher in the high ACE group than in the low ACE group across networks (frontoparietal (a), salience (b), and default mode (c)). The total ACE score was also positively associated with flexibility in the frontoparietal network (d), salience network (e), and DMN (f).

Table 2.

ACE and TD effects on flexibility.

Beta Estimate Effect Size Mean (Low ACE or Sham TD) SD (Low ACE or Sham TD) Mean (High ACE or Active TD) SD (High ACE or Active TD) p
Frontoparietal Network
 ACE group 0.00748 2.79 0.0724 0.0110 0.0783 0.0125 0.005*
 TD status 0.00489 2.15 0.0726 0.0129 0.0772 0.0104 0.03
Salience Network
 ACE group 0.00679 3.02 0.0788 0.0095 0.0842 0.0103 0.003*
 TD status 0.00393 1.91 0.0793 0.0109 0.0829 0.0090 0.06
Default Mode Network
 ACE group 0.00910 3.53 0.0754 0.0103 0.0837 0.0097 0.0004*
 TD status 0.00334 1.73 0.0773 0.0119 0.0805 0.0094 0.08*
Open in a new tab

Unadjusted means are displayed. Tests of significance also controlled for age, time since last menstraul period, estradiol level, and motion.

*

p ≤ 0.05.

ACE = adverse childhood experiences; TD = tryptophan depletion; SD = standard deviation.

3.3. Associations between behavior and network reconfiguration

Prior literature has suggested that there may be a U shaped relationship between flexibility and executive function (Braun et al., 2016). We therefore evaluated both quadratic and linear fits for the relationship between flexibility and performance on the CET (Figure 3; Table 3). There was a significant quadratic relationship between frontoparietal network flexibility and two behavioral outcomes: total trials (quadratic term: β=61248.9, D=3.56, p=0.0004; linear term: β= −8958.6, D=−3.48, p=0.0005) and total errors (quadratic term: β=27869.5, D=2.85, p=0.004; linear term: β=−3929.7, D=−2.70, p=0.007). There was not a U-shaped relationship between frontoparietal network flexibility and categories achieved. Frontoparietal flexibility was linearly associated with fewer categories achieved (β=−9.9, D=−2.19, p=0.03), though the variability of this measure was small. There was not a U-shaped relationship between performance and flexibility of the DMN or salience network. For these networks, a linear function was the best fit for the data. Increased salience flexibility was associated with a greater number of errors (β=270.4, D=1.70, p=0.09) and greater number of trials (β=310.0, D=1.50, p=0.1) at a trend level. Increased flexibility in the DMN was similarly associated with a greater number of total errors (β=269.7, D=2.12, p=0.03), a greater number of trials (β=342.0, D=2.12, p=0.03), and fewer categories achieved (β=−11.7, D=−1.96, p=0.05).

Figure 3. Associations Between Network Reconfiguration and Behavior.

Figure 3.

Open in a new tab

There was a significant quadratic relationship between frontoparietal network flexibility and two behavioral outcomes: total trials and total errors (a). Frontoparietal flexibility was linearly associated with fewer categories achieved, though the variability of this measure was small. Increased salience flexibility was linearly associated with a greater number of errors and greater number of trials at a trend level (b). Increased flexibility in the default mode network was similarly associated with a greater number of total errors, a greater number of trials, and fewer categories achieved (c).

Table 3.

Associations between network reconfiguration and behavior.

Frontoparietal Network
 Salience Network
 Default Mode Network
Beta (Quadratic) Effect Size (Quadratic) p (Quadratic) Beta (Linear) Effect Size (Linear) p (Linear) Beta (Linear) Effect Size (Linear) p (Linear) Beta (Linear) Effect Size (Linear) p (Linear)
Total trials 61248.9 3.56 0.0004* −8958.6 −3.48 0.0005* 310.0 1.50 0.1 342.0 2.12 0.03*
Total errors 27869.5 2.85 0.004* −3929.7 −2.70 0.007* 270.4 1.70 0.09 269.7 2.12 0.03*
Categories achieved NA NA NA −9.9 −2.19 0.03* −6.5 −1.29 0.2 −11.7 −1.96 0.05*
Open in a new tab
*

p ≤ 0.05.

4. Discussion

In this report, we examined the impact of childhood adversity and tryptophan depletion on network reconfiguration dynamics in the frontoparietal, salience, and default mode networks during a working memory task in menopausal women. Higher levels of childhood adversity were associated with greater flexibility across networks. While high and low ACE groups did not respond differently to TD, lowering serotonin levels increased network reconfiguration in both groups. Together, these results suggest that childhood adversity has lasting impacts on dynamic reconfiguration of networks underlying executive function and that decreasing serotonin levels can worsen these effects independent of adversity history.

We found a negative effect of ACE on network reconfiguration dynamics. These results are in agreement with a study that showed that higher overall flexibility during the n-back is associated with increased genetic risk for schizophrenia as well as poorer performance on an out-of-scanner task measuring set shifting in healthy individuals (Braun et al., 2016). Our results are also generally convergent with the existing literature demonstrating that childhood adversity negatively impacts several neural markers of executive function including brain activation (Philip et al., 2016; Shanmugan et al., 2017a) and functional connectivity (Philip et al., 2013; Shanmugan et al., 2017c). These results suggest that one mechanism by which childhood adversity exerts lasting negative effects on executive function involves alteration of functional brain network dynamics.

We found that decreasing central serotonin levels with TD increased flexibility of executive and default mode networks during performance of the n-back working memory task. This suggests that serotonin plays a role in network dynamics regardless of ACE status. Few studies have examined how this dynamic reconfiguration relates to mechanisms at the neurotransmitter level. While prior studies have related altered flexibility to NMDA hypofunction (Braun et al., 2016), heightened arousal (Betzel et al., 2017), and negative affect (Betzel et al., 2017), this study is the first to demonstrate that manipulating serotonin alters functional network dynamics. This finding provides additional evidence for the role of serotonin in executive function and extends our understanding of how the neurotransmitter environment impacts dynamic network reconfiguration.

Given our prior work suggesting that childhood adversity impacts serotonergic circuits underlying executive processes (Shanmugan et al., 2017a), we hypothesized that the mechanisms underlying the observed differences in ACE-associated network dynamics would involve alterations in serotonin function. However, there was not a significant ACE × TD interaction on network flexibility, suggesting that ACE effects on network flexibility may be primarily via other mechanisms that do not directly involve serotonin. These mechanisms may include ACE effects on dopaminergic or glutamatergic systems or ACE effects on functional or structural brain development. Animal studies have shown that early adversity is associated with increased dopamine transporter (DAT) expression (Novick et al., 2015), increased catechol-o-methyl transferase (COMT) expression (Grissom et al., 2015), and increased D2 autoreceptor expression (Lovic et al., 2013) and activity (Watt et al., 2014). Evidence from rodent studies indicates that early life stress alters the composition of N-methyl-D-aspartic acid (NMDA) receptor subunits in the prefrontal cortex (Kinnunen et al., 2003), which are important for the persistent firing of glutamatergic neurons during working memory and sustained attention (Arnsten and Wang, 2016). Early stress in mice has also been shown to decrease mGluR2 and mGluR3 mRNA and protein levels in the frontal cortex (Matrisciano et al., 2012). Adversity may also interfere with the segregation of functional brain networks that occurs during brain development (Vaidya and Gordon, 2013), potentially via structural changes in myelination or dendritic spine architecture. In female mice, peripubertal stress increases PFC expression of myelin basic protein and proteolipid protein, genes that are indicators of myelination (Morrison et al., 2016). Chronic postnatal stress has been shown to decrease dendritic spine density in the PFC of rats (Michelsen et al., 2007). This negative impact of early life stress on dendritic spine density is particularly intriguing given that executive functioning is dependent on the number of prefrontal small spines (Hara et al., 2012) and the substantial literature demonstrating negative effects of ovariectomy and positive effects of estradiol on dendritic spine morphology (Shanmugan and Epperson, 2014).

We also examined the relationship between flexibility and performance on the CET. Given that prior literature has suggested that there may be a U shaped relationship between flexibility and executive function (Braun et al., 2016), we evaluated both quadratic and linear fits. Variability in categories achieved was small, so the results regarding a quadratic versus linear fit are likely not meaningful for this outcome. Interestingly, the relationship between flexibility and the remaining two behavioral outcomes was quadratic for the frontoparietal network but linear for the DMN and salience network. This dissociation may be reflective of the task’s greater reliance on frontoparietal brain regions (Buchsbaum et al., 2005) such as the prefrontal cortex, where catecholamine concentration has an inverted U shaped dose-response influence on neuron firing and cognitive function (Arnsten et al., 2016).

4.1. Limitations

Certain limitations of the present study should be mentioned. First, our sample consisted of hypogonadal females, as this population is at risk for new onset executive dysfunction and may be particularly susceptible to the effects of TD and ACE (Shanmugan et al., 2017a). Given the effects of estradiol on executive function, these results cannot be generalized to males or pre-menopausal females. Second, the sample of healthy hypogonadal women used may not be representative of the general aging population who have experienced childhood adversity. However, using this sample allowed us to examine the effects of childhood adversity on dynamic network reconfiguration in a population where the effects of risk factors for cognitive dysfunction may be particularly acute, while simultaneously limiting the potentially confounding impacts of hormonal status and comorbid psychiatric symptomatology and unstable or chronic medical conditions. Third, although the impact of ACE group was strong, the impact of TD in this study is weak and would not survive correction for multiple comparisons. Further replication in a larger sample with greater power would provide stronger support for the relationship between serotonin and network dynamics.

4.2. Conclusions

In summary, these data suggest that one mechanism by which childhood adversity exerts a lasting negative impact on executive function during menopause involves altered reconfiguration dynamics of executive and default mode networks. Our results also provide preliminary evidence for the role of serotonin in dynamic reconfiguration of networks supporting executive function.

Figure 1. Calculation of Network Flexibility.

Figure 1.

Open in a new tab

We examined the effects of ACE and TD on dynamic functional connectivity of the frontoparietal, salience, and default mode networks as defined previously (a, b). BOLD time series were extracted from each node and functional connectivity between each pair of network nodes was estimated using a wavelet coherence between time-series calculated within overlapping sliding-windows (c). This calculation yielded a matrix describing the coherence between each pair of nodes during each of the 95 windows in the time-series (d). One hundred optimizations of the generalized Louvian multilayer community detection algorithm were used to partition these matrices into time-dependent communities (using resolution parameter values γ=1, ω=1). A time-dependent network flexibility matrix was calculated at each optimization that indicated whether a node changed its community between two time windows (e). The flexibility (F) of each node was defined as the number of times that the node changed community assignment across the task, normalized by the total number of changes possible. Averaging these measures of flexibility across the 100 optimizations and across nodes within each network yielded measures of frontoparietal, salience, and DMN flexibility for each subject.

Highlights.

  • Functional brain networks reconfigure during executive tasks

  • Optimal network dynamics are necessary for efficient executive function

  • Too much network reconfiguration may indicate poorer executive function

  • Early adversity is associated with greater network reconfiguration during menopause

  • Decreasing a precursor of serotonin increases functional network reconfiguration

Acknowledgments

Funding and Disclosure

This research was supported by P50 MH099910 (Epperson), Penn PROMOTES Research on Sex and Gender in Health (Epperson), K12 HD085848 (Epperson), R01 AG030641 and R01 AG048839 (Epperson), R01MH107703 (Satterthwaite), R01MH113550 (Satterthwaite and Bassett), R01MH107235 (Gur), F30AG055256 (Shanmugan) and R25MH119043. Bassett acknowledges support from the Alfred P. Sloan Foundation, the ISI Foundation, and the John D. and Catherine T. MacArthur Foundation. Dr. Epperson reports that she has received funding from Shire Pharmaceuticals for investigator initiated research and is site investigator for a multi-site, randomized clinical trial funded by Sage Therapeutics. Dr. Epperson consults to Sage Therapeutics and Shire and discloses personal and/or family investments in the following companies; BMS, Johnson and Johnson, Merck, Abbott, and Abbvie.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Arnsten AF, Wang M, 2016. Targeting Prefrontal Cortical Systems for Drug Development: Potential Therapies for Cognitive Disorders. Annual review of pharmacology and toxicology 56, 339–360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Arnsten AFT, Wang MJ, Paspalas CD, 2012. Neuromodulation of thought: flexibilities and vulnerabilities in prefrontal cortical network synapses. Neuron 76, 223–239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Avants BB, Tustison NJ, Song G, Cook PA, Klein A, Gee JC, 2011. A reproducible evaluation of ANTs similarity metric performance in brain image registration. NeuroImage 54, 2033–2044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bassett DS, Porter MA, Wymbs NF, Grafton ST, Carlson JM, Mucha PJ, 2013. Robust detection of dynamic community structure in networks. Chaos (Woodbury, N.Y.) 23, 013142. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bassett DS, Wymbs NF, Porter MA, Mucha PJ, Carlson JM, Grafton ST, 2011. Dynamic reconfiguration of human brain networks during learning. Proceedings of the National Academy of Sciences of the United States of America 108, 7641–7646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Betzel RF, Satterthwaite TD, Gold JI, Bassett DS, 2017. Positive affect, surprise, and fatigue are correlates of network flexibility. Scientific reports 7, 520. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Braun U, Schafer A, Bassett DS, 2016. Dynamic brain network reconfiguration as a potential schizophrenia genetic risk mechanism modulated by NMDA receptor function 113,12568–12573. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Braun U, Schafer A, Walter H, Erk S, Romanczuk-Seiferth N, Haddad L, Schweiger JI,Grimm O,Heinz A,Tost H,Meyer-Lindenberg A,Bassett DS, 2015Dynamicreconfigurationof frontalbrainPre-proofnetworksduring executivecognition inhumans.ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesof America 112,11678–11683. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Buchsbaum BR,Greer S,Chang WL,Berman KF,2005Meta-analysisof neuroimagingstudies oftheWisconsinCard-Sortingtaskandcomponentprocesses. Humanbrainmapping 25,35–45. [Google Scholar]
  10. Cao H, Plichta MM,Schafer A, Haddad L,Grimm O,Schneider M,Esslinger C, Kirsch P,Meyer-Lindenberg A,Tost H,2014Test-retestreliabilityoffMRI-based graphtheoreticalpropertiesduringworkingmemory,emotionprocessing,and restingstate. NeuroImage 84,888–900. [DOI] [PubMed] [Google Scholar]
  11. Chapman DP,Whitfield CL,Felitti VJ,Dube SR,Edwards VJ,Anda RF,2004. Adverse childhoodexperiencesandthe riskofdepressivedisordersinadulthood. Journalofaffectivedisorders 82,217–225. [Google Scholar]
  12. Cole MW,Bassett DS,Power JD,Braver TS, Petersen SE,2014Intrinsicand task-evokednetworkarchitecturesofthehuman brain.Neuron 83,238–251. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Dosenbach NU, Nardos B,Cohen AL,Fair DA,Power JD,Church JA,Nelson SM,Wig GS, Vogel AC,Lessov-Schlaggar CN,Barnes KA,Dubis JW,Feczko E, Coalson RS, Pruett JR Jr., Barch DM,Petersen SE,Schlaggar BL,2010. Predictionofindividualbrain maturityusing fMRI. Science (NewYork,N.Y.)329, 1358–1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Epperson CN, Pittman B, Czarkowski KA, Bradley J, Quinlan DM, Brown TE, 2011. Impact of atomoxetine on subjective attention and memory difficulties in perimenopausal and postmenopausal women. Menopause (New York, N.Y.) 18, 542–548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Epperson CN, Sammel MD, Bale TL, Kim DR, Conlin S, Scalice S, Freeman K, Freeman EW, 2017. Adverse Childhood Experiences and Risk for First-Episode Major Depression During the Menopause Transition. The Journal of clinical psychiatry 78, e298–e307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Felitti VJ, Anda RF, Nordenberg D, Williamson DF, Spitz AM, Edwards V, Koss MP, Marks JS, 1998. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The Adverse Childhood Experiences (ACE) Study. American journal of preventive medicine 14, 245–258. [DOI] [PubMed] [Google Scholar]
  17. Greve DN, Fischl B, 2009. Accurate and robust brain image alignment using boundary-based registration. NeuroImage 48, 63–72. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Grissom NM, Herdt CT, Desilets J, Lidsky-Everson J, Reyes TM, 2015. Dissociable deficits of executive function caused by gestational adversity are linked to specific transcriptional changes in the prefrontal cortex. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology 40, 1353–1363. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Højsgaard S, Halekoh U, J Y., 2006. The R Package geepack for Generalized Estimating Equations. Journal of Statistical Software 15/2, 1–11. [Google Scholar]
  20. Hara Y, Rapp PR, Morrison JH, 2012. Neuronal and morphological bases of cognitive decline in aged rhesus monkeys. Age (Dordrecht, Netherlands) 34, 1051–1073. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Kinnunen AK, Koenig JI, Bilbe G, 2003. Repeated variable prenatal stress alters pre- and postsynaptic gene expression in the rat frontal pole. Journal of neurochemistry 86, 736–748.Pre-proof [DOI] [PubMed] [Google Scholar]
  22. Klein A, Andersson J, Ardekani BA, Ashburner J, Avants B, Chiang MC, Christensen GE, Collins DL, Gee J, Hellier P, Song JH, Jenkinson M, Lepage C, Rueckert D,Thompson P,Vercauteren T,Woods RP,Mann JJ, Parsey RV, 2009. Evaluation of 14 nonlinear deformation algorithms applied to human brain MRI registration. NeuroImage 46, 786–802. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lovic V, Belay H, Walker CD, Burton CL, Meaney MJ, Sokolowski M, Fleming AS, 2013. Early postnatal experience and DRD2 genotype affect dopamine receptor expression in the rat ventral striatum. Behavioural brain research 237, 278–282. [DOI] [PubMed] [Google Scholar]
  24. Matrisciano F, Tueting P, Maccari S, Nicoletti F, Guidotti A, 2012. Pharmacological activation of group-II metabotropic glutamate receptors corrects a schizophrenia-like phenotype induced by prenatal stress in mice. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology 37, 929–938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Michelsen KA, van den Hove DL, Schmitz C, Segers O, Prickaerts J, Steinbusch HW, 2007. Prenatal stress and subsequent exposure to chronic mild stress influence dendritic spine density and morphology in the rat medial prefrontal cortex. BMC neuroscience 8, 107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Morrison KE, Narasimhan S, Fein E, Bale TL, 2016. Peripubertal Stress With Social Support Promotes Resilience in the Face of Aging. Endocrinology 157, 2002–2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jeub GS, Bazzi M, Jutla IS, Mucha PJ, 2011–2017. A generalized Louvain method for community detection implemented in MATLAB. [Google Scholar]
  28. Mucha PJ, Richardson T, Macon K, Porter MA, Onnela JP, 2010. Community structure in time-dependent, multiscale, and multiplex networks. Science (New York, N.Y.) 328, 876–878. [DOI] [PubMed] [Google Scholar]
  29. Novick AM, Forster GL, Hassell JE, Davies DR, Scholl JL, Renner KJ, Watt MJ, 2015. Increased dopamine transporter function as a mechanism for dopamine hypoactivity in the adult infralimbic medial prefrontal cortex following adolescent social stress. Neuropharmacology 97, 194–200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Philip NS, Sweet LH, Tyrka AR, Carpenter SL, Albright SE, Price LH, Carpenter LL, 2016. Exposure to childhood trauma is associated with altered n-back activation and performance in healthy adults: implications for a commonly used working memory task. Brain imaging and behavior 10, 124–135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Philip NS, Sweet LH, Tyrka AR, Price LH, Bloom RF, Carpenter LL, 2013. Decreased default network connectivity is associated with early life stress in medication-free healthy adults. European neuropsychopharmacology: the journal of the European College of Neuropsychopharmacology 23, 24–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Power JD, Cohen AL, Nelson SM, Wig GS, Barnes KA, Church JA, Vogel AC, Laumann TO, Miezin FM, Schlaggar BL, Petersen SE, 2011. Functional network organization of the human brain. Neuron 72, 665–678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Satterthwaite TD, Elliott MA, Gerraty RT, Ruparel K, Loughead J, Calkins ME, Eickhoff SB, Hakonarson H, Gur RC, Gur RE, Wolf DH, 2013a. An improved framework for confound regression and filtering for control of motion artifact in the preprocessing of resting-state functional connectivity Pre-proofdata. NeuroImage 64,240–256. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Satterthwaite TD, Wolf DH, Ruparel K, Erus G, Elliott MA, Eickhoff SB, Gennatas ED,Jackson C,Prabhakaran K,Smith A,Hakonarson H, Verma R,Davatzikos C., Gur RE, Gur RC, 2013b. Heterogeneous impact of motion on fundamental patterns of developmental changes in functional connectivity during youth. NeuroImage 83, 45–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Shanmugan S, Epperson CN, 2014. Estrogen and the prefrontal cortex: towards a new understanding of estrogen's effects on executive functions in the menopause transition. Human brain mapping 35, 847–865. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Shanmugan S, Loughead J, Cao W, Sammel MD, Satterthwaite TD, Ruparel K, Gur RC, Epperson CN, 2017a. Impact of Tryptophan Depletion on Executive System Function during Menopause is Moderated by Childhood Adversity. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Shanmugan S, Loughead J, Nanga RP, Elliott M, Hariharan H, Appleby D, Kim D, Ruparel K, Reddy R, Brown TE, Epperson CN, 2017b. Lisdexamfetamine Effects on Executive Activation and Neurochemistry in Menopausal Women with Executive Function Difficulties. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology 42, 437–445. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Shanmugan S, Sammel MD, Loughead J, Ruparel K, Gur RC, Brown TE, Faust J, Domchek S, Epperson CN, 2020. Executive function after risk-reducing salpingo-oophorectomy in BRCA1 and BRCA2 mutation carriers: does current mood and early life adversity matter? Menopause (New York, N.Y.). [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Shanmugan S, Satterthwaite TD, Sammel MD, Cao W, Ruparel K, Gur RC, Epperson CN, Loughead J, 2017c. Impact of early life adversity and tryptophan depletion on functional connectivity in menopausal women: A double-blind, placebo-controlled crossover study. Psychoneuroendocrinology 84, 197–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Sun FT, Miller LM, D'Esposito M, 2004. Measuring interregional functional connectivity using coherence and partial coherence analyses of fMRI data. NeuroImage 21, 647–658. [DOI] [PubMed] [Google Scholar]
  41. Vaidya CJ, Gordon EM, 2013. Phenotypic variability in resting-state functional connectivity: current status. Brain connectivity 3, 99–120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Watt MJ, Roberts CL, Scholl JL, Meyer DL, Miiller LC, Barr JL, Novick AM, Renner KJ, Forster GL, 2014. Decreased prefrontal cortex dopamine activity following adolescent social defeat in male rats: role of dopamine D2 receptors. Psychopharmacology 231, 1627–1636. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES