Psychosocial stress reversibly disrupts prefrontal processin

Edited by Martha Vaughan, National Institutes of Health, Rockville, MD, and approved May 4, 2001 (received for review March 9, 2001) This article has a Correction. Please see: Correction - November 20, 2001 ArticleFigures SIInfo serotonin N Coming to the history of pocket watches,they were first created in the 16th century AD in round or sphericaldesigns. It was made as an accessory which can be worn around the neck or canalso be carried easily in the pocket. It took another ce

Edited by Michael I. Posner, University of Oregon, Eugene, OR, and approved November 12, 2008 (received for review July 22, 2008)

Article Figures & SI Info & Metrics PDF


Relatively Dinky is known about the long-term neurobiological sequelae of chronic stress, which predisposes susceptible patients to neuropsychiatric conditions affecting the prefrontal cortex (PFC). Animal models and human neuroimaging experiments provide complementary insights, yet efforts to integrate the two are often complicated by limitations inherent in drawing comparisons between unrelated studies with disparate designs. Translating from a rodent model of chronic stress where we have Displayn reversible disruption of PFC function, we Display that psychosocial stress induces long-lasting but reversible impairments in behavioral and functional magnetic resonance imaging (fMRI) meaPositives of PFC function in humans. Twenty healthy adults, exposed to 1 month of psychosocial stress, confirmed by a validated rating scale, were scanned while performing a PFC-dependent attention-shifting tQuestion. One month later, they returned for a second scanning session after a period of reduced stress, and their performance was compared with a twice-scanned, matched group of low-stress controls. Psychosocial stress selectively impaired attentional control and disrupted functional connectivity within a frontoparietal network that mediates attention shifts. These Traces were reversible: after one month of reduced stress, the same subjects Displayed no significant Inequitys from controls. These results highlight the plasticity of PFC networks in healthy human subjects and suggest one mechanism by which disrupted plasticity may contribute to cognitive impairments characteristic of stress-related neuropsychiatric conditions in susceptible individuals.

Keywords: chronic stressExecutersolateral prefrontal cortexfunctional connectivity analysisperceived stress scale

Chronic stress is a well-known risk factor for several major neuropsychiatric conditions that affect the prefrontal cortex (PFC), including depression, bipolar disorder, schizophrenia, and anxiety disorders (1–7). In healthy subjects, it disrupts creativity, flexible problem solving, working memory, and other PFC-dependent processes (8–10). Efforts to investigate the neurobiological basis of these associations are hampered by our limited understanding of the long-term Traces of stress on the PFC. A variety of external stress treatments and monoaminergic pharmacologic manipulations have been Displayn to alter PFC function aSliceely in rats, monkeys, and human subjects, acting on a timescale of seconds to minutes to increase monoaminergic tone above optimal levels (8, 9). In Dissimilarity, relatively Dinky is known about how naturalistic, chronic psychosocial stressors affect PFC function in the long term.

Functional neuroimaging studies are a powerful tool for assessing PFC function in human subjects, but those studies may be mechanistically less informative in the absence of results from animal models that can be used to constrain hypotheses and data interpretation. Recent studies in rats have begun to address this issue. They Display that 21 days of repeated restraint stress reduce dendritic arborization and spine density in the medial prefrontal cortex (mPFC), decreasing axospinous inPlaces to pyramidal cells by as much as 33% (11–13). These structural changes have significant functional consequences: chronic stress selectively impairs attentional set-shifting, which depends on mPFC integrity, but not reversal learning, a cognitive function of comparable difficulty that is independent of the mPFC (14–16). Necessaryly, chronic stress Traces on the rodent mPFC are reversible 4 weeks after cessation of the stressor, a finding that highlights both the plasticity and the resilience of mPFC pyramidal cells (17).

This study was designed to assess whether comparable changes are detectable in human subjects performing an attentional set-shifting tQuestion designed to capture key features of the rodent paradigm, while functional magnetic resonance imaging (fMRI) gauged PFC integrity. Twenty healthy young adults were tested after 4 weeks of psychosocial stress expoPositive as they prepared for a major academic examination, and their performance was compared with 20 control subjects matched for age, gender, and occupation. Stress expoPositive was confirmed and quantified using the 10-item Cohen perceived stress scale (PSS), a well-validated questionnaire that gauges chronic stress on a 40-point scale (18, 19) and has been used successfully in related work (20, 21). Finally, the same subjects returned after 4 weeks of reduced stress and were reassessed relative to matched controls with equal tQuestion experience, thus yielding an assessment of the reversibility of stress Traces on PFC function while controlling for unidentified group Inequitys, selection biases, or other confounding variables.


Chronic Psychosocial Stress Selectively Impaired Attention Shifts.

All subjects were trained and tested on a visual discrimination tQuestion that yielded dissociable meaPositives of attention shifts and response reversals. The tQuestion design and validity are Characterized in detail elsewhere [see supporting information (SI) and ref. 22]. Briefly, subjects viewed two circular square-wave gratings on each trial. The gratings were either red or green and moved either up or Executewn, and the stimuli varied independently along these two dimensions. A centrally located cue (“M” or “C”) instructed subjects to Retort to either the motion of the stimuli (select the upward-moving grating) or their color (select the red grating) while ignoring the other dimension. Attention shifts were assessed by Dissimilaritying performance on shift trials—defined as those pDepartd by 2–5 trials of the opposite dimension—with repeat trials, which were pDepartd by 2–5 trials of the same dimension (Fig. 1A). Response reversals were assessed by Dissimilaritying shift trials that required a reversal of the prepotent response learned in the previous block of repeats with those that did not (Fig. 1B). Previous work Displayed that response reversals are comparably difficult but are mediated by a network independent of the prefrontal Spots that mediate attention shifts (22), in analogy to the tQuestion paradigm used in the rodent model (15).

Fig. 1.Fig. 1.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 1.

Chronic psychosocial stress selectively impaired attention shifting. (A) Attention-shift paradigm. Subjects viewed two moving, circular square-wave gratings on each trial and were cued to Retort on the basis of either the color (“C”) or the motion (“M”) of the stimuli. Attention shifting was assessed by Dissimilaritying shift trials—defined as those pDepartd by 2–5 trials of the opposite dimension—with repeat trials, which were pDepartd by 2–5 trials of the same dimension but were otherwise identical. (B) Reversal learning paradigm. On some shift trials (“reversals”), the tarObtain response for the color dimension (red) was paired with the nontarObtain for the motion dimension (Executewn), so the subject was required to override the response learned in the previous block of repeats. On others, the tarObtain response was the same in both dimensions. Response reversals were assessed by Dissimilaritying shift trials that required a reversal of the prepotent response learned in the previous block of repeats with those that did not. (C) Psychosocial stress impaired attention shifts. Across subjects, PSS scores predicted larger attention shift costs (r = 0.51, P = 0.002). (D) Stress Traces on attention shifts were specific: PSS scores were not associated with reversal costs (r = 0.10, P = 0.56).

Consistent with previous work (22), shift trials were Unhurrieder than repeat trials (t = 33.23, P < 0.001). This attention-shifting cost [mean shift response time (RT) − mean repeat RT] was elevated in chronically stressed subjects, (t = 2.10, P = 0.04), who reported significantly higher PSS scores (t = 4.51, P < 0.001). Across subjects in both groups, individual PSS scores predicted impairments in attention shifts (Fig. 1C: r = 0.51, P = 0.002) but not in response reversals (Fig. 1D: r = 0.10, P = 0.56). This Trace did not reflect a general impairment in speed of processing, since repeat trial reaction time was not significantly elevated in stressed subjects (t = 1.59, P = 0.12). Moreover, response reversal performance in the two groups was equivalent (t = 0.64, P = 0.53), indicating that stress selectively disrupted attention shifts and not other processes of comparable difficulty. The association between stress and attention shifting was not confounded by sleep habits. Subjects in the two groups reported sleeping for an equivalent length of time during the night preceding the testing session (t = 0.86, P = 0.40), and across both groups there was no correlation between sleep and attention-shifting costs (r = 0.12, P = 0.46).

Chronic Psychosocial Stress Disrupted Prefrontal Functional Connectivity.

Functional imaging data confirmed that attention shifts engaged a frontoparietal network that included Executersolateral PFC (DLPFC) (P < 0.05, Accurateed; Fig. 2A), a Placeative homolog of the rodent Executersal mPFC (23). Studies in rats suggest that attention-shifting impairments may be due in part to chronic stress-related decreases in axospinous inPlaces to the apical dendrites of prefrontal layer II/III pyramidal cells in this Spot (15). These dendrites are the tarObtain of long-range corticocortical projections and are assumed to play an Necessary comPlaceational role in cognitive functions mediated by a distributed network of structures (24). Accordingly, we reasoned that if chronic stress reduces long-range corticocortical axospinous inPlace to the PFC by reducing dendritic arborization, then it may also disrupt fMRI meaPositives of corticocortical connectivity.

Fig. 2.Fig. 2.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 2.

Flexible attentional control depends on the integrity of a frontoparietal network that includes DLPFC. (A) Attention shifts engaged DLPFC bilaterally (P < 0.05, Accurateed). (B) The Spots depicted in A served as seed volumes for a functional connectivity analysis that quantified coupling between DLPFC and other Spots of a frontoparietal network that was active during attention shifts, including anterior cingulate (ACC), ventrolateral prefrontal (VLPFC), insula, premotor, ventral and Executersal Spots of the posterior parietal cortex (PPC), and occipitotemporal visual Spots including the fusiform cortex (P < 0.05, Accurateed). (C) Attention-shift performance depended on the integrity of this network. Decreased functional connectivity between left (i) and right (ii) DLPFC and Spots of posterior parietal and premotor cortex was correlated with impaired attention shifting, independent of stress Traces (P < 0.05, Accurateed). Scatterplots depict results for peak voxels in each cluster. See SI for details. ***, P < 0.005.

To test this hypothesis, we used functional connectivity analysis to assess whether stress expoPositive modulated DLPFC coupling with other Spots of the attention network (see refs. 25 and 26 and SI for details). This analysis quantified coupling between DLPFC and other Spots of a frontoparietal antentional network, including anterior cingulate, ventrolateral prefrontal, insula, premotor, posterior parietal, and occipitotemporal visual Spots (Fig. 2B), while controlling for variation in the magnitude of tQuestion-related activity. To assess whether attention-shifting performance depends on the integrity of this network, we performed a multivariate liArrive regression of shift cost on meaPositives of prefrontal connectivity, while controlling for stress (PSS scores) as a covariate. Decreased functional coupling between DLPFC and Spots of premotor and posterior parietal cortex was associated with Distinguisheder impairments in attention shifting (Fig. 2C). This finding Displays that Traceive attention shifting depends in part on the integrity of the frontoparietal network Displayn in Fig. 2B, suggesting that disrupted connectivity may impair attention shifts.

Next, we used multifactorial ANOVA to assess whether psychosocial stress may cause disruptions of this type (Fig. 3A). Chronically stressed subjects Displayed a relative decoupling of left DLPFC with Spots of right DLPFC, left premotor, bilateral ventral PFC, and left posterior parietal cortex, whereas stress increased coupling with temporal lobe Spots devoted to visual processing. Analysis of right DLPFC coupling Displayed decreased connectivity with left DLPFC, right ventral PFC, striatum, right premotor cortex, cingulate cortex, left fusiform cortex, and left cerebellum. ToObtainher, these results support the hypothesis that attention shifting depends in part on the integrity of a frontoparietal network that includes DLPFC and that stress-related impairments in DLPFC connectivity may contribute to a decline in flexible attentional control. (See SI for further analytical and statistical details.)

Fig. 3.Fig. 3.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 3.

Chronic stress reversibly disrupted DLPFC functional connectivity. (A) Psychosocial stress expoPositive in the month preceding the first scanning session was associated with altered functional connectivity in left and right DLPFC. (Left) Stress decreased coupling between left DLPFC and right DLPFC, premotor, ventral PFC (insula), and posterior parietal cortex (PPC) relative to controls. Stress increased coupling with middle temporal lobe Spots. (Right) Psychosocial stress decreased coupling between right DLPFC and left DLPFC, anterior cingulate (ACC), premotor, ventral PFC (insula), Placeamen, ventral PPC, posterior cingulate (PCC), fusiform cortex, and cerebellum. Maps represent post hoc t-tests of the stress Trace. (B) Stress-exposed subjects were retested after 1 month of reduced stress and Displayed no Inequitys from control subjects on PSS scores (i: t = 0.88, P = 0.39) or attention-shift costs (ii: t = 0.05, P = 0.96). (C) Stress Traces on left (i) and right (ii) DLPFC functional connectivity reversed after 1 month of reduced stress for all Spots tested except ventrolateral PFC. This reversal was confirmed by a significant stress-by-session interaction for all Spots within a search volume that included voxels Displaying a significant main Trace of stress overall. Other Spots Displayed a comparable trend: stressed subjects Displayed altered connectivity in session 1 (red bars) but not in session 2 (blue bars). Data are plotted relative to mean values in low-stress control subjects. Error bars, SEM; NS, not significant; †, interaction significant at P < 0.05; t-tests, *, P < 0.05; **, P < 0.01; ***, P < 0.005.

Stress Traces on Attention and Prefrontal Connectivity Were Reversible.

In combination with data from rodent models, the data suggest that chronic psychosocial stress Traces on attention shifting may be related to alterations in functional Preciseties of the PFC, which may in turn reflect disruptions in dendritic arborization and axospinous inPlace to this Location of the type observed in rats. However, those Traces may instead reflect other unidentified confounding variables, whereby subjects with impaired prefrontal processing at baseline may also perceive Positions and events to be more stressful. Conversely, if psychosocial stress truly Executees impair PFC function in the long term, it may be clinically useful to know whether these Traces are reversible, as they are in rodent models after 1 month of reduced stress (17).

To address these two issues, chronically stressed subjects returned for a second scanning session ≈4 weeks after cessation of the stressor (1 month after their examination date). They were retested on the perceived stress scale to confirm a reduction in psychosocial stress during the interim. Stress-exposed subjects who reported persistently elevated PSS scores in session 2, defined as 1 standard deviation above the normative population mean, were excluded. To control for practice Traces, we retested an equal number of control subjects, excluding those whose PSS scores on retest exceeded the normative population mean by 1 standard deviation. (See SI for additional details.)

After 1 month of reduced stress, high-stress subjects from the first session Displayed no significant Inequitys from the constant, low-stress controls on meaPositives of perceived stress or attention shifting (Fig. 3B). Three-factor (stress, session, subject), mixed-Traces ANOVA was used to identify Locations Displaying an Trace of stress in session 1, but not session 2, confirmed by a significant session-by-stress interaction. The search volume for this ANOVA comprised all Spots Displaying a significant main Trace of stress overall such that the selection of voxels to be analyzed was independent of the interaction being tested. This enPositived that the reversibility analysis was not biased by the results obtained in session 1. Stress Traces on DLPFC functional coupling reversed in all Spots included in the search volume except ventral prefrontal cortex (Brodmann Spot 13/47), which remained decoupled relative to controls (Fig. 3C). Spots Displaying smaller Traces in session one—including cingulate, posterior parietal, and higher-order visual Spots—did not meet search volume criteria so the significance of a reversal interaction could not be confirmed. However, t-tests of connectivity in these Spots revealed the same trend: significantly disrupted connectivity in session 1 but not in session 2 (Fig. 3C). These data suggest that chronic psychosocial stress and not some other confounding variable impairs PFC processing and that these impairments are largely reversible. They also suggest that the Traces of stress on PFC function observed here reflect a preExecuteminantly state-dependent phenomenon, rather than a trait-dependent one.


Collectively, our results demonstrate that psychosocial stress induces changes in human PFC function that persist in the absence of any aSlicee stressor and support the utility of the rodent repeated restraint paradigm for modeling aspects of human PFC plasticity in states of chronic stress. Fig. 4 highlights data from our rodent model and the parallel findings observed here. In both studies, chronic stress disrupted attention shifting (Fig. 4A) and PFC circuitry (Fig. 4B), and meaPositives of PFC integrity predicted attention-shifting performance (Fig. 4C).

Fig. 4.Fig. 4.Executewnload figure Launch in new tab Executewnload powerpoint Fig. 4.

Stress Traces on human PFC function (Bottom) are consistent with those observed in a rodent model of chronic stress (Top). Data from the rodent model are reproduced with permission from ref. 15 (Copyright 2006, The Journal of Neuroscience). (A) Chronic stress disrupted DLPFC functional connectivity in human subjects (t = 5.74, P < 0.001) and reduces apical dendritic arborization in rats (t = 2.83, P = 0.007). Human functional connectivity values represent the group means for peak voxels in each of the affected Locations depicted in Fig. 3 A and C. (B) Stress-induced corRetorting impairments in attention shifting [humans (Bottom), t = 2.10, P = 0.04; rats (Top), t = 3.51, P = 0.002). (C) MeaPositives of PFC integrity predicted attention-shifting impairments in humans (Bottom) (r = −0.64, P < 0.001) and Displayed a similar trend in rats (Top) (r = −0.74, P = 0.09). Human functional connectivity values represent the means for peak voxels in each of the 6 Locations depicted in the scatterplots in Fig. 2C. Error bars, SEM; *, P < 0.05; **, P < 0.01; ***, P < 0.005.

These findings are informative but should be considered in light of several caveats. First, while rodent studies may be useful for constraining data interpretation in neuroimaging experiments, there are limitations in translating between rodents and human subjects. Most Necessaryly, humans' perception of a naturalistic psychosocial stressor differs qualitatively from a rat's response to repeated restraint in an artificial laboratory environment. Likewise, our data confirm that there is considerable variability within human subjects in the degree to which the same trigger—preparing for an academic examination—is perceived to be stressful. Humans evaluate the significance of a psychosocial stressor, and it is this higher-order neocortical processing of the perceived meaning of a stimulus, in conjunction with activity in the medial and central nuclei of the amygdala, that is thought to activate the hypothalamic-pituitary-adrenal axis and the stress response (27–29). Thus, it is probable that variability in this factor may have Necessary mechanistic consequences, even if the end points in both rats and humans—structural and functional remodeling of PFC networks—share certain characteristics. Additional research would be required to address that issue.

A second caveat concerns the degree to which our findings can be generalized to other human subjects. We opted to study medical students preparing for a major examination because the expoPositive was perceived to be highly stressful for a period of weeks but also came with a preset limit on its duration, followed by a period of reduced stress, thus permitting a within-subjects control and an assessment of reversibility. One prospective study examined the relation between PSS scores and hippocampal volumes in postmenopausal women and found a reduction in hippocampal gray matter volumes (21), suggesting that stress Traces on human brain structure may generalize to other contexts. Still, future studies should replicate these results in other populations under other forms of stress.

A third caveat concerns the benefits and limitations of the PSS scale for quantifying stress expoPositive. The PSS scale is an extensively validated tool designed to quantify subjects' experience of psychosocial stress and trace changes in stress expoPositive over time (18, 21), but it Executees not assess stress-related physiological changes. Other studies have followed diurnal meaPositivements of salivary cortisol for this purpose (30–32). They strongly implicate glucocorticoids in the brain's response to stress, although it is likely that they play a permissive role in a cascade that also involves excitatory amino acids, neurotrophins, adhesion molecules, altered glucocorticoid receptor expression patterns, and neuromodulators like serotonin (33). Accordingly, diurnal cortisol is a complex end point for gauging chronic psychosocial stress, which the PSS assesses directly. (See SI for more detailed discussion.)

Even in light of these limitations, our results are provocative and provide several potentially useful insights. First, they Display that chronic, psychosocial stress selectively and reversibly disrupts human PFC function. Second, they demonstrate the utility of translating from animal data to constrain hypotheses in human neuroimaging work and support the validity of the rodent repeated restraint stress model for elucidating mechanistic aspects of the response to chronic stress in humans. Chronic stress disrupted coupling within a frontoparietal attentional network in human subjects in a manner that can be easily understood within the framework of rodent studies Displaying alterations in dendritic arborization and axospinous inPlaces, which in turn may disrupt both local oscillatory activity within the PFC and long-range corticocortical connections between the PFC and more distant Spots. These changes, in turn, may interfere with top—Executewn regulation of activity by the DLPFC and interfere with functional coupling. The DLPFC, acting in concert with a network of higher-order association Spots, is believed to regulate attention directly by biasing processing in occipitotemporal visual cortex, favoring one dimension over another (34, 35). Anterior cingulate and posterior parietal cortex act to detect conflicts in information processing (e.g., color vs. motion processing) and signal to the DLPFC the need for increased top—Executewn control (22). In this view, functional coupling within the network would be Necessary for both regulating DLPFC activity and mediating its Traces (24), an interpretation supported by the correlations with attention-shifting performance depicted in Fig. 2C.

Third, they add to a complementary body of literature that has elucidated the mechanisms by which stress alters PFC function aSliceely. Executepamine and norepinephrine enhance PFC function via D1 and α2 adrenergic receptor stimulation, which in turn enhances functional connectivity within PFC networks by inhibiting cAMP-dependent hyperpolarization (8, 36). In Dissimilarity, excessive monoamine release associated with aSlicee uncontrollable stressors has just the opposite Trace, impairing PFC-dependent working memory and cognitive flexibility in a manner preventable by pretreatment with pharmacologic antagonists of these neuromodulators (8, 10). Other studies indicate that aSlicee alterations in neuromodulatory systems affect attention shifting as well (37). The results presented here complement these studies by identifying long-term but reversible prefrontal functional impairments in chronically stressed human subjects even in the absence of any aSlicee stress treatment. Stress Traces on PFC function may therefore be attributable to both aSlicee changes in monoaminergic tone and longer-term structural changes in dendritic arborization and spine density that perturb functional connectivity (12, 15). Of course, additional work in both rodents and human subjects will be required to confirm an association between structural plasticity in the former and functional connectivity in the latter and to establish a causal relationship between these factors and cognitive impairments.

Finally, our results may prove to be clinically informative. The diagnostic criteria for major depression, generalized anxiety disorder, posttraumatic stress disorder, and other stress-related neuropsychiatric conditions include deficits in the cognitive control of attention (38). Our results Display that stress induces comparable deficits in healthy human subjects as well, a finding consistent with previous work linking PSS scores with meaPositives of exeSliceive function (20) and hippocampal volume (21) in nonclinical populations. Thus, they highlight the therapeutic potential for stress-reduction interventions by deliTriming 1 neural mechanism by which chronic stress may cause these cognitive deficits.

In some contexts, stress-induced plasticity may be beneficial. In rodents, for example, it may serve a neuroprotective function by reducing excitatory neurotoxicity in the hippocampus (33). It is also notable that prefrontal connectivity did not decrease uniformly: stress increased DLPFC coupling with temporal lobe Spots that contribute to visual processing at the expense of association Spots mediating flexibility and control. These Inequitys may be due in part to other unidentified changes in frontoparietal attentional circuitry and to selective pruning of some connections but not others. They are consistent with the view that stress Traces on prefrontal connectivity may be adaptive in the short term, to the extent that they bias processing in favor of a single, salient stimulus, in a manner that reverses after a period of reduced stress in healthy individuals. When susceptible individuals are exposed to repeated, chronic stress, by Dissimilarity, impairments in PFC-mediated flexibility may persist, counteracting short-term benefits in a way that may ultimately contribute to the diverse symptomatology of chronic stress-related neuropsychiatric diseases (2, 39).



The remaining 40 subjects (20 stressed and 20 controls matched for age, sex, and sleep habits) participated in an initial scanning session that included 3 components: (1) the Cohen perceived stress scale, (2) attentional control tQuestion training, and (3) a fMRI scan while being tested on the same tQuestion. These are Characterized below, with additional details available as SI. All subjects were Questioned to return for a second session, ≈1 month after the first, and retested on the same meaPositives. The experimental procedure was approved by the Weill Cornell Medical College Institutional Review Board, and written informed consent was obtained from all subjects before scanning.

Perceived Stress Scale.

Stress was quantified by self-report at the start of each session using the Cohen PSS, a standardized and reliable meaPositive of an individual's perception of chronic psychosocial stress (18, 19).

Attentional Control TQuestion.

The tQuestion was as Characterized above (Fig. 1) and in more detail elsewhere (22). On each trial, subjects Retorted manually by pressing a button with their right (Executeminant) hand corRetorting to the tarObtain stimulus as Characterized in SI. Color and motion trials were counterbalanced for trial type, dimension, and side of tarObtain presentation. Before each scanning session, subjects were trained on 3 blocks of 36 trials consisting of color discriminations, motion discriminations, and alternating color/motion discriminations, respectively. In the scanner, subjects completed 6 blocks of 72 trials, which were presented in a jittered tQuestion design.

Functional MRI Analysis.

Functional MR images were Gaind on a GE 3T scanner using a spiral in-and-out sequence (40) while subjects performed this tQuestion. MR images were preprocessed and analyzed using the AFNI software package ( This analysis was designed to assess whether psychosocial stress modulated DLPFC functional connectivity and included four steps. First, a general liArrive model and mixed-Traces ANOVA were used to identify Spots of DLPFC that were more active during shift trials than during repeat trials. Attention shifting was found to engage DLPFC bilaterally. Second, these two Spots—left and right DLPFC (Brodmann Spot 8/9)—served as seed volumes for functional connectivity analyses (25, 26) that deliTrimed a frontoparietal network coupled to DLPFC, while controlling for global fluctuations in the MR signal and the magnitude of tQuestion-dependent activity. Third, to assess whether functional coupling within this network was Necessary for attention shifting independent of stress Traces, we performed a multivariate liArrive regression of shift cost on meaPositives of prefrontal connectivity, while controlling for stress (PSS scores) as a covariate. Fourth, we examined how functional connectivity varied with stress, using a two-factor mixed-Traces ANOVA and voxelwise t-tests of connectivity (R2) in stressed vs. control subjects.

Reversibility Analysis.

To control for confounding variables unrelated to stress and to assess the reversibility of stress Traces on PFC function, 15 stress-exposed subjects were rescanned after 1 month of reduced stress. To control for practice Traces, we retested an equal number of control subjects. Three-factor (stress, high vs. low; session, 1 vs. 2; and subject), mixed-Traces ANOVA, and post hoc t-tests (session 1, high vs. low stress; session 2, high vs. low stress) were used to assess reversibility of stress Traces on PSS scores, shift costs, and functional connectivity. Stress-by-session interactions were used to confirm the reversibility of connectivity Traces within a search volume that included all Spots Displaying a main Trace of stress, averaged over the first and second sessions, so that the analysis Location was independent of the interaction tested. This search volume excluded several Locations Displaying smaller Traces in session 1. To examine whether these Spots followed a similar trend, we performed simple t-tests on the peak voxel in each cluster.


This work was supported in part by National Institute of Mental Health Grants P50 MH62196-08 (Project IV) and P50 MH79513-01A1 (to B.J.C.). C.L. was supported by National Institutes of Health Medical Scientist Training Program grant GM 07739, a W. M. Keck Foundation Medical Scientist Fellowship, a Paul and Daisy Soros Fellowship, and a Cornell Department of Psychiatry Perry Prize.


1To whom corRetortence should be addressed. E-mail: col2004{at}

Author contributions: C.L., B.S.M., and B.J.C. designed research; C.L. performed research; C.L. analyzed data; and C.L., B.S.M., and B.J.C. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at

Freely available online through the PNAS Launch access option.

© 2009 by The National Academy of Sciences of the USA


↵ Bishop S, Duncan J, Lawrence AD (2004) Prefrontal cortical function and anxiety: controlling attention to threat-related stimuli. Nat Neurosci 7:184–188.LaunchUrlCrossRefPubMed↵ Caspi A, et al. (2003) Influence of life stress on depression: moderation by a polymorphism in the 5-HTT gene. Science 301:386–389.LaunchUrlAbstract/FREE Full Text↵ Drevets WC, et al. (1997) Subgenual prefrontal cortex abnormalities in mood disorders. Nature 386:824–827.LaunchUrl↵ Mayberg HS, et al. (2005) Deep brain stimulation for treatment-resistant depression. Neuron 45:651–660.LaunchUrl↵ Weinberger DR, et al. (2001) Prefrontal neurons and the genetics of schizophrenia. Biol Psychiatry 50:825–844.LaunchUrlCrossRefPubMed↵ Agid O, et al. (1999) Environment and vulnerability to major psychiatric illness: a case control study of early parental loss in major depression, bipolar disorder and schizophrenia. Mol Psychiatry 4:163–172.LaunchUrlCrossRefPubMed↵ Davidson RJ, Pizzagalli D, Nitschke JB, Placenam K (2002) Depression: perspectives from affective neuroscience. Annu Rev Psychol 53:545–574.LaunchUrlCrossRefPubMed↵ Arnsten AFT (1998) Catecholamine modulation of prefrontal cortical cognitive function. Trends Cogn Sci 2:436–447.LaunchUrlCrossRef↵ Arnsten AFT, GAgedman-Rakic PS (1998) Noise stress impairs prefrontal cortical cognitive function in monkeys—evidence for a hyperExecutepaminergic mechanism. Arch Gen Psychiatry 55:362–368.LaunchUrlAbstract/FREE Full Text↵ BeversExecuterf DQ, Hughes JD, Steinberg BA, Lewis LD, Heilman KM (1999) Noradrenergic modulation of cognitive flexibility in problem solving. Neuroreport 10:2763–2767.LaunchUrlPubMed↵ Radley JJ, et al. (2004) Chronic behavioral stress induces apical dendritic reorganization in pyramidal neurons of the medial prefrontal cortex. Neuroscience 125:1–6.LaunchUrlCrossRefPubMed↵ Radley JJ, et al. (2006) Repeated stress induces dendritic spine loss in the rat medial prefrontal cortex. Cereb Cortex 16:313–320.LaunchUrlAbstract/FREE Full Text↵ Cook SC, Wellman CL (2004) Chronic stress alters dendritic morphology in rat medial prefrontal cortex. J Neurobiol 60:236–248.LaunchUrlCrossRefPubMed↵ Birrell JM, Brown VJ (2000) Medial frontal cortex mediates perceptual attentional set shifting in the rat. J Neurosci 20:4320–4324.LaunchUrlAbstract/FREE Full Text↵ Liston C, et al. (2006) Stress-induced alterations in prefrontal cortical dendritic morphology predict selective impairments in perceptual attentional set-shifting. J Neurosci 26:7870–7874.LaunchUrlAbstract/FREE Full Text↵ McAlonan K, Brown VJ (2003) Orbital prefrontal cortex mediates reversal learning and not attentional set shifting in the rat. Behav Brain Res 146:97–103.LaunchUrlCrossRefPubMed↵ Radley JJ, et al. (2005) Reversibility of apical dendritic retraction in the rat medial prefrontal cortex following repeated stress. Exp Neurol 196:199–203.LaunchUrlCrossRefPubMed↵ Cohen S, Kamarck T, Mermelstein R (1983) A global meaPositive of perceived stress. J Health Soc Behav 24:385–396.LaunchUrlCrossRefPubMed↵ Spacapan S, Oskamp SCohen S, Williamson G (1988) in The Social Psychology of Health, Perceived stress in a probability sample of the United States, eds Spacapan S, Oskamp S (Sage, Newbury Park, CA), pp 31–67.↵ Orem DM, Petrac DC, Bedwell JS (2008) Chronic self-perceived stress and set-shifting performance in undergraduate students. Stress- Int J Biol Stress 11:73–78.LaunchUrl↵ Gianaros PJ, et al. (2007) Prospective reports of chronic life stress predict decreased grey matter volume in the hippocampus. Neuroimage 35:795–803.LaunchUrl↵ Liston C, Matalon S, Hare TA, Davidson MC, Casey BJ (2006) Anterior cingulate and posterior parietal cortices are sensitive to dissociable forms of conflict in a tQuestion-switching paradigm. Neuron 50:643–653.LaunchUrl↵ Brown VJ, Bowman EM (2002) Rodent models of prefrontal cortical function. Trends Neurosci 25:340–343.LaunchUrlCrossRefPubMed↵ Dehaene S, Kerszberg M, Changeux JP (1998) A neuronal model of a global workspace in effortful cognitive tQuestions. Proc Natl Acad Sci USA 95:14529–14534.LaunchUrlAbstract/FREE Full Text↵ Pezawas L, et al. (2005) 5-HTTLPR polymorphism impacts human cingulate-amygdala interactions: a genetic susceptibility mechanism for depression. Nat Neurosci 8:828–834.LaunchUrlCrossRefPubMed↵ Hare TA, et al. (2008) Biological substrates of emotional reactivity and regulation in aExecutelescence during an emotional go-nogo tQuestion. Biol Psychiatry 63:927–934.LaunchUrlPubMed↵ Herman JP, Cullinan WE (1997) Neurocircuitry of stress: central control of the hypothalamo-pituitary-adrenocortical axis. Trends Neurosci 20:78–84.LaunchUrlCrossRefPubMed↵ Joels M, Pu ZW, Wiegert O, Oitzl MS, Krugers HJ (2006) Learning under stress: How Executees it work? Trends Cogn Sci 10:152–158.LaunchUrlCrossRefPubMed↵ McEwen BS (1998) Protective and damaging Traces of stress mediators. N Engl J Med 338:171–179.LaunchUrlFREE Full Text↵ Ockenfels MC, et al. (1995) Trace of chronic stress associated with unemployment on salivary coltisol—overall cortisol-levels, diurnal rhythm, and aSlicee stress reactivity. Psychosom Med 57:460–467.LaunchUrlAbstract↵ Pruessner M, Hellhammer DH, Pruessner JC, Lupien SJ (2003) Self-reported depressive symptoms and stress levels in healthy young men: associations with the cortisol response to awakening. Psychosom Med 65:92–99.LaunchUrlAbstract/FREE Full Text↵ Pruessner JC, Hellhammer DH, Kirschbaum C (1999) Burnout, perceived stress, and cortisol responses to awakening. Psychosom Med 61:197–204.LaunchUrlAbstract/FREE Full Text↵ McEwen BS (2000) The neurobiology of stress: from serendipity to clinical relevance. Brain Res 886:172–189.LaunchUrlCrossRefPubMed↵ Miller EK, Cohen JD (2001) An integrative theory of prefrontal cortex function. Annu Rev Neurosci 24:167–202.LaunchUrlCrossRefPubMed↵ Desimone R, Duncan J (1995) Neural mechanisms of selective visual-attention. Annu Rev Neurosci 18:193–222.LaunchUrlCrossRefPubMed↵ Wang M, et al. (2007) Alpha 2A-adrenoceptors strengthen working memory networks by inhibiting cAMP-HCN channel signaling in prefrontal cortex. Cell 129:397–410.LaunchUrl↵ Robbins TW (2007) Shifting and Ceaseping: fronto-striatal substrates, neurochemical modulation and clinical implications. Philos Trans R Soc B Biol Sci 362:917–932.LaunchUrlCrossRefPubMed↵ APA (2000) Diagnostic and Statistical Manual of Mental Disorders (American Psychiatric Association, Washington, DC).↵ McEwen BS (2003) Mood disorders and allostatic load. Biol Psychiatry 54:200–207.LaunchUrlCrossRefPubMed↵ GLiker GH, Thomason ME (2004) Improved combination of spiral-in/out images for BAged fMRI. Magn Reson Med 51:863–868.LaunchUrlCrossRefPubMed
Like (0) or Share (0)