Increased Rostral Anterior Cingulate Cortex Volume in Chronic Primary Insomnia

Abstract

Background:

Recent studies document alterations in cortical and subcortical volumes in patients with chronic primary insomnia (PI) in comparison with normal sleepers. We sought to confirm this observation in two previously studied PI cohorts.

Methods:

Two separate and independent groups of unmedicated patients who met Diagnostic and Statistical Manual for Mental Disorders, Fourth Edition (DSM-IV) criteria for PI were compared with two separate, healthy control groups (Study 1: PI = 20, controls = 15; Study 2: PI = 21, controls = 20). Both studies included 2 weeks of sleep diaries supplemented by wrist actigraphy. The 3.0 T MRI-derived rostral anterior cingulate cortex (rACC) volumes were measured with FreeSurfer image analysis suite (version 5.0) and results normalized to total intracranial volume (ICV). Unpaired t-tests (two-tailed) were used to compare rACC volumes between groups. Post hoc correlations of rACC volumes to insomnia severity measures were performed (uncorrected for multiplicity).

Results:

Both studies demonstrated increases in normalized rACC volume in PI compared with control patients (Study 1: right side P = 0.05, left side P = 0.03; Study 2: right side P = 0.03, left side P = 0.02). In PI patients from Study 1, right rACC volume was correlated with sleep onset latency (SOL) by both diary (r = 0.51, P = 0.02) and actigraphy (r = 0.50, P = 0.03), and with sleep efficiency by actigraphy (r = -0.57, P = 0.01); left rACC volume was correlated with SOL by diary (r = 0.48, P = 0.04), and wake after sleep onset (WASO) (r = 0.49, P = 0.03) and sleep efficiency (r = -0.49, P = 0.03) by actigraphy. In Study 2, right rACC volume was correlated with SOL by diary (r = 0.44, P = 0.05) in PI patients.

Conclusions:

Rostral ACC volumes are larger in patients with PI compared with control patients. Clinical severity measures in PI correlate with rACC volumes. These data may reflect a compensatory brain response to chronic insomnia and may represent a marker of resilience to depressive illness.

Citation:

Winkelman JW; Plante DT; Schoerning L; Benson K; Buxton OM; O'Connor SP; Jensen JE; Renshaw PF; Gonenc A. Increased rostral anterior cingulate cortex volume in chronic primary insomnia. SLEEP 2013;36(7):991-998.

INTRODUCTION

Insomnia is the most common sleep complaint in industrialized countries, affecting nearly one third of all adults in any given year, and chronically affecting 10-15% of the adult population.¹ Approximately 25% of those with chronic insomnia are considered to have primary insomnia (PI), which requires a sleep complaint exceeding 1 month, associated sequelae of daytime impairment or clinically significant distress, and the absence of coexisting sleep, medical, or psychiatric disorders. Although only a minority of patients with chronic insomnia meet criteria for PI, the scientific advantage of studying chronic PI is that the biology of insomnia can be investigated in a relatively pure condition, independent of influences attributable to any coexisting comorbid medical or psychiatric disorder.

Emerging data suggest that PI may be associated with neuroanatomical and neurophysiological alterations compared with healthy normal sleepers. Riemann et al.² found decreased bilateral hippocampal volumes (corrected for intracranial volume) in eight middle-aged patients with severe PI, although subsequent studies did not confirm this finding.^3–5 Another study, using whole brain voxel-based morphometry in 24 older patients with PI, found reduced volumes of the left orbitofrontal cortex (OFC), the bilateral anterior precuneus of the parietal cortex, and the bilateral posterior precuneus in the occipitoparietal cortex.³

Recent evidence suggests that the anterior cingulate cortex (ACC) may be an important area in which functional differences in PI are expressed. Using magnetic resonance spectroscopy (MRS) we demonstrated a 21% reduction in gamma-aminobutyric acid (GABA) levels specifically in the ACC.⁶ This finding is in agreement with other human studies in PI^7,8 and an animal model of insomnia,⁹ where abnormalities were observed in the ACC.

Based on this evidence of abnormal activity in the ACC in insomnia, we retrospectively examined the volumes of the rostral anterior cingulate cortex (rACC) bilaterally in two independent, but similarly defined, groups of patients with PI compared with healthy sleeper control patients (Study 1 and Study 2).

METHODS AND MATERIALS

Participants

Young adult and middle-aged patients (25-55 y in Study 1 and 18-60 y in Study 2) were recruited from clinical samples and community advertisements from June 2006 to May 2008 in Study 1 and from July 2009 to January 2011 in Study 2. Patients with PI met DSM-IV diagnostic criteria for PI, with a report of difficulty initiating or maintaining sleep or nonrestorative sleep with resulting daytime distress or dysfunction that was not attributable to another medical or psychiatric disorder. Consistent with recommended research diagnostic criteria,¹⁰ additional severity criteria included a self-report of typical sleep onset latency (SOL) plus wake after sleep onset (WASO) of at least 30 min, and a duration of insomnia ≥ 6 months. Severity criteria for inclusion were slightly more stringent in Study 1 in that patients had to report a habitual total sleep time ≤ 7 h and (1) SOL > 45 min or (2) WASO > 45 min or (3) total wake time during the sleep period (SOL + WASO) > 60 min. Age- and sex-matched healthy control patients without sleep complaints or a history of sleep disorders were also recruited. Patients were not permitted to use central nervous system (CNS) active agents for 2 weeks prior to data collection and for the duration of the study.

All patients were evaluated with an unstructured clinical interview for history of sleep and medical disorders, and interview for lifetime history of psychiatric disorders with the Structured Clinical Interview for DSM-IV (SCID). All patients were administered the Insomnia Severity Index (ISI),¹¹ Pittsburgh Sleep Quality Index (PSQI),¹² Dysfunctional Beliefs and Attitudes about Sleep (DBAS-16) (only Study 2),¹³ and Beck Depression Inventory (BDI-IA).¹⁴ Control patients in Study 1 were administered the PSQI only. Baseline laboratory studies included urine toxicology and pregnancy testing (for female patients) in both studies, and a thorough laboratory assessment in Study 1.¹⁵

Exclusion criteria for all patients included clinical evidence of any moderate to severe sleep disorder other than insomnia (e.g., obstructive sleep apnea, restless legs syndrome, etc.); current or past (within the preceding year) diagnosis of any Axis I disorder (including alcohol or drug dependence/abuse) other than PI (in the PI group); history of significant medical or neurologic illness including significant head trauma or loss of consciousness > 30 min; regular treatment (more than one time/ week) with CNS active agents within 3 months of the screening visit; body mass index (BMI) > 35; consumption of > 10 cigarettes/day, more than two caffeinated beverages/day, or more than two standard alcoholic drinks/day for a period > 1 month within the preceding year; history of shift work; perimenopausal symptoms that disrupted sleep; contraindications for magnetic resonance imaging (MRI); and women who were pregnant, lactating, or planning to become pregnant during the study.

The study was approved by the Institutional Review Boards of Partners Healthcare, the parent organization of Brigham and Women's Hospital, and McLean Hospital, and carried out in accordance with the Declaration of Helsinki. All patients received compensation for their participation in this study.

Sleep Diaries and Actigraphy

Following initial evaluation, patients completed sleep-wake diaries supplemented by wrist-worn actigraphy (Actiwatch AW-64, Minimitter Inc., Bend, OR) for a minimum of 2 weeks prior to MRI brain scans. Diaries included self-report of sleep-wake parameters (e.g., bedtime and wake time, and estimated SOL, number of awakenings, and WASO), caffeine/alcohol/medication consumption, and a visual analog scale (VAS) of subjective sleep quality (Study 2 only). Patients with PI were excluded if their sleep diaries did not demonstrate aforementioned severity criteria on the majority of days during the baseline period. Healthy sleeper control patients were excluded if they demonstrated < 7.5 or > 10 h of sleep on 10 of 14 nights during the screening period. Actigraphy data included six standard measures: time in bed, SOL, total sleep time, sleep efficiency, number of awakenings, and all WASO.¹⁵ Actigraphy was not used as an inclusionary/exclusionary measure per se but was used to corroborate sleep diary information.

MRI Acquisition

All participants underwent MRI on a 3.0 T Siemens Trio scanner (Siemens Medical Solutions, Erlangen, Germany) equipped with a standard 12-channel head coil. T1-weighted images with contiguous 1.3-mm slices in the sagittal plane were obtained using a magnetization prepared rapid gradient echo sequence. Imaging parameters for Study 1 were: echo time = 2.74 ms, repetition time = 2,100 ms, flip angle = 12°, acquisition matrix = 256 × 256, number of averages = 1, and field of view = 25.6 cm. Imaging parameters for Study 2 were as follows: echo time = 3.31 ms, repetition time = 2,530 ms, flip angle = 7°, acquisition matrix = 256 × 192, number of averages = 1, and field of view = 25.6 cm. Besides different acquisition parameters, the two studies had also different scanner operating systems (VA25A versus VB17A).

Imaging Processing and Analysis

The image files in DICOM (Digital Imaging and Communications in Medicine) format were transferred to a Mac OSX workstation for analysis. Individual scans were reviewed by a radiologist for clinical abnormalities and by a magnetic resonance physicist, and those with significant artifact or motion disturbance were excluded from analysis. Volumetric segmentation was performed automatically with the FreeSurfer image analysis suite (version 5), which is documented and freely available online (http://surfer.nmr.mgh.harvard.edu/). This software automatically provides segments and labels for many brain structures and assigns a neuroanatomic label to each voxel in MRI volume on the basis of probabilistic information estimated automatically from a manually labeled training set. The technical details of these procedures are described in prior publications.^16–27 Briefly, this processing includes motion correction, removal of nonbrain tissue using a hybrid watershed/ surface deformation procedure,²⁷ automated Talairach transformation, segmentation of the subcortical white matter and deep gray matter volumetric structures (including the hippocampus, amygdala, caudate, putamen, and ventricles),^20,21 intensity normalization,²⁸ tessellation of the gray-white matter boundary, automated topology correction,^19,29 and surface deformation following intensity gradients to optimally place the gray-white matter and gray matter/cerebrospinal fluid (CSF) borders at the location where the greatest shift in intensity defines the transition to the other tissue class.^16–18 Once the cortical models are complete, a number of deformable procedures are then performed for further data analysis including parcellation of the cerebral cortex into units based on gyral and sulcal structure.^24,30 FreeSurfer morphometric procedures have been demonstrated to show good test-retest reliability across scanner manufacturers and field strengths.²⁵ However, all cortical parcellations and subcortical segmentations were visually inspected for accuracy prior to inclusion in the group analysis.

Statistical Analysis

Statistical analysis was performed using the IBM SPSS software (version 19). Chi-square tests (for categorical variables) and t-tests (for continuous variables) were used to compare demographic and clinical characteristics across groups. The effects of diagnosis (patients with PI versus control patients) on volumes were investigated using a linear model with diagnosis as the between-subjects factor. Normality of distributions was tested with the Shapiro-Wilk test. Correlations between the clinical indices and volumes were carried out using a series of correlation coefficients. A P value < 0.05 was considered statistically significant.

The primary region of interest (ROI) was the bilateral rACC from Studies 1 and 2 (Figure 1). As is apparent in Figure 1, in FreeSurfer the rACC includes both rostral as well as ventral and subgenual aspects of the ACC. Secondary ROIs included the bilateral pars orbitalis, which is the region defined by FreeSurfer that most closely approximates the orbitofrontal cortex described by Altena, et al.,³ as well as the bilateral hippocampus. The hippocampus was not examined in Study 1 as our previous analysis of those data had demonstrated no significant differences in hippocampal volumes in this cohort relative to healthy control patients.⁴ Exploratory analyses examined all other cortical volumes reported with the FreeSurfer program. All volume results were reported normalized to total intracranial volume as calculated by FreeSurfer.

Figure 1.

Pial cortical representation of the rostral anterior cingulate cortex (rACC) (from Desikan et al.³⁰).

RESULTS

Demographic characteristics of the PI and control groups in both studies are presented in Table 1. In Study 1, 20 patients with PI and 15 control patients were available for morpho-metric analysis; in Study 2, 21 patients with PI and 20 control patients had evaluable morphometric data. In both studies, insomnia was long-standing, with continuous durations greater than 1 y in all patients with PI (except one patient in Study 1), and greater than 5 y for 12 of 20 patients with PI in Study 1 and 15 of 21 patients with PI in Study 2.

Table 1.

Demographic and sleep-wake variables in primary insomnia and control patients in Study 1 and Study 2

No patient with PI had a prior lifetime history of a mood or psychotic disorder; one patient with PI in Study 2 had a prior history of anxiety disorder not otherwise specified that had been in remission > 6 months. All patients with PI in both studies had not taken prescription sleep medication for at least 1 month prior to MRI; two patients in Study 1, and five in Study 2, had used such agents intermittently within the 6 months prior to MRI.

Patients in PI and control groups did not differ in age, sex, or BMI (see Table 1). Psychometric scales, sleep-wake diaries, and actigraphy corroborated the insomnia complaint in patients with PI (Table 1). A complete sleep diary was not available for one patient in Study 1. In both Study 1 and 2, sleep diaries demonstrate significantly longer sleep latencies (37.4 min and 31.5 min, respectively) and significantly shorter total sleep times (1.5 h in both) in the PI group compared with control patients. Actigraphy data were not available for two patients in Study 1 (1 PI, 1 control) and 1 patient with PI in Study 2. Actigraphy data demonstrated clear differences between the PI and control groups in Study 2, but not Study 1. The distinction between the two studies can largely be attributed to the prolonged actigraphic measure of WASO (and thus low sleep efficiency) in the control group in Study 1.

The BDI is reported with and without items related to sleep and fatigue (items 16 and 17) (Table 1). In both studies, low BDI scores confirmed the lack of active mood disorder in the patients with PI. In Study 2, though BDI scores were statistically higher in patients with PI versus control patients, they were still far below standard cutoffs for a clinical mood disorder, and elevated items were typical of daytime consequences of insomnia (e.g., irritability, effort, worry).

MRI measures of whole brain and normalized rACC volumes are shown in Table 2 for both Study 1 and Study 2. Shapiro-Wilk tests indicated that volumes were normally distributed, and group differences were evaluated with an independent-samples t-test. Both studies demonstrated increases in normalized rACC volume in patients with PI compared with control patients (Study 1: right side P = 0.05, left side P = 0.03; Study 2: right side P = 0.03, left side P = 0.02). No differences between PI and control patients were observed for intracranial volumes. Normalized rACC volumes did not correlate with age or duration of insomnia. Multivariate regression using age and sex as covariates did not change the results. Normalized rACC volumes correlated with BDI scores (without sleep items) in Study 2 on the right side only (r = 0.44, P = 0.05); these correlations were not significant in Study 1.

Table 2.

Cortical volumes (normalized to intracranial volume) in patients with primary insomnia and healthy control patients from morphometric analysis (mean ± standard deviation) in Study 1 and Study 2

Exploratory analyses examined the relationships of morphometric data with both subjective (sleep diary) and objective (actigraphy) indices of sleep in patients with PI in both Study 1 and Study 2 (Table 3 and Figure 2). In Study 1, ACC volumes were correlated with both sleep onset and maintenance measures by both diary and actigraphy. In Study 2, only right rACC volume was correlated with SOL by diary (r = 0.44, P = 0.05) but not by actigraphy.

Table 3.

Correlation of rostral anterior cingulate cortex volumes to insomnia severity measures in patients with primary insomnia

Figure 2.

Correlations of left (A) and right (B) rostral anterior cingulate cortex (rACC) with diary-derived sleep latency in Study 1.

Exploratory analyses revealed a significant decrease in the right pars orbitalis volume in patients with PI compared with control patients (P = 0.03) in Study 2 and trends for decrease in Study 1 (P = 0.08 bilaterally) (Table 2). No differences between patients with PI and control patients in hippocampal volumes were found in Study 2. However, we did find significant correlations in this study between left hippocampal volume and both WASO (r = -0.45, P = 0.05) and sleep efficiency (r = 0.53, P = 0.02) derived from actigraphy. These correlations were not significant for the right hippocampus. No statistical differences were observed for any of the other cortical areas examined.

DISCUSSION

We have demonstrated larger rACC volumes bilaterally in two independent groups of patients with chronic PI compared with matched control patients. The fact that we had nearly identical findings in two separate groups of patients with insomnia and that we observed significant correlations of ACC volumes with subjective and objective measures of sleep in the PI group, such that larger volumes were associated with worse sleep, provide further confidence in our results. Furthermore, potential confounders such as age and sex did not explain differences between patients with PI and control groups, and all patients were medication free and carefully screened to be free of current and lifetime psychiatric illness.

Little is known about the role of the ACC in the regulation of normal sleep in humans. However, both cerebral blood flow using H₂¹⁵O positron emission tomography (PET) and glucose metabolism from fludeoxyglucose (FDG) suggest dramatic changes in ACC activity between sleep and wake,^31,32 during sleep deprivation,³³ and across sleep stages.³¹

In addition, there is increasing evidence of the importance of the ACC in insomnia. In our recent study using MRS,⁶ we documented a 21% reduction in GABA levels in the rostral-dorsal ACC in patients with PI (Study 2 in the current paper). Because GABA is the primary inhibitory transmitter in the CNS, such an imbalance of excitatory and inhibitory influences may represent the neurochemical basis for the hyperarousal observed in physiological and cognitive domains in PI.³⁴ Consistent with the role of hyperarousal in PI, an ¹⁸FDG-PET study in patients with PI found smaller decreases in glucose utilization in the entire ACC from waking to nonrapid eye movement sleep in comparison with healthy patients.⁷ In a follow-up PET study, ACC metabolism in the rostral ACC was correlated with both subjective and objective WASO in patients with insomnia.⁸ Finally, further support for the role of the ACC in insomnia comes from an animal model of insomnia,⁹ in which exposure to a stressful environment (cage exchange) was associated with both sleep disturbance and increased activity in the cingulate cortex as measured by Fos expression.

The increased size of the rACC observed in our Study 1 and 2 may be an expression of chronically increased activity in the ACC associated with PI. Activity-related changes in cortical thickness, density, and volume have been observed in adult humans associated with both physical training as well as cognitive activity.³⁵ For instance, mathematicians showed increases in parietal cortex gray matter density³⁶; ballet dancers,³⁷ experienced typists,³⁸ and those with improved dexterity in the nonimmobilized arm after a fracture³⁹ all had increases in relevant motor areas; and medical students demonstrated increases in hippocampal size after studying for exams.⁴⁰ Although such macrostructural changes are evident through neuroimaging, the underlying cellular mechanisms are still unknown. Potential processes include changes in the size of either the neuronal cell body/nucleus or capillaries, remodeling of dendridic spines and axon terminals, or glial hypertrophy.

Alterations in rostral (and to a lesser extent, subgenual) ACC volume have also been consistently observed in patients with major depressive disorder (MDD).^41,42 The observation that rACC size is altered in both PI and MDD is not surprising given several factors: insomnia is an important risk factor for the development of MDD⁴³; hyperarousal is a core feature of both disorders⁴⁴; and sleep disturbance is among the diagnostic criteria of MDD, with insomnia being the predominant symptom, occurring in approximately 75% of patients with MDD.⁴⁵ Functional neuroimaging studies demonstrate abnormalities in the subgenual ACC⁴⁶ in mood disorders, and deep brain stimulation of the subgenual ACC may lead to remission of depressive symptoms.⁴⁷ However, most volumetric studies in MDD demonstrate smaller rostral and subgenual ACC volumes during an episode of depression.^41,42

If reduction in rACC volume is characteristic of MDD, which is commonly associated with sleep disturbance, what does our finding of increased rACC volume in PI (those with chronic sleep disturbance without psychiatric illness) suggest? In distinction to results in patients with active MDD, specific increases in ACC volume have been observed in patients with remitted MDD. In 27 unmedicated patients with long-standing remitted MDD,⁴⁸ rostral and ventral ACC volumes were larger compared with both control patients and currently depressed individuals. The authors interpreted the ACC volume enlargement as a marker of resilience against chronic or recurrent depression, either as a compensatory response to previous depression or as a trait marker. Similarly, Yucel⁴⁹ also found larger subgenual ACC volumes in remitted, but not currently depressed, medicated patients, compared with healthy control patients. Finally, an MRI study of 66 individuals with high familial risk of MDD⁵⁰ found thickening and greater volume of the right rostral and ventral ACC compared with those without familial risk. This increased volume was independent of current depressive symptoms, suggesting that it was a marker of risk, but not of disease, again suggesting resilience to depression.

Although speculative, we propose that the increased rACC volume in our patients with chronic insomnia without any history of psychiatric illness may be a marker of resilience to MDD. In this framework, the increased volume may be a compensatory response to chronic sleep disturbance, which then serves as a protective measure against the development of mood symptoms. An alternate scenario is that the increased rACC volume predates the onset of sleep disturbance and is present as a trait marker, rather than as a compensation, in those with resistance to the deleterious effects of insomnia on mood. Our cross-sectional data are unable to distinguish these two alternate interpretations, and longitudinal studies in such individuals would be necessary to understand this process more fully. Similarly, it is unclear that it is the sleep disturbance per se that produces the volumetric expansion; both the sleep disturbance and increased rACC size may reflect hyperarousal or some other underlying process that predisposes to MDD. Regardless of the mechanism of this association, our data suggest a potential reframing of PI, such that those with chronic insomnia are understood both as sufferers from a sleep disorder, but also as nonaffected individuals with both a risk for, and resilience to, mood disorders.

We were unable to confirm an earlier report from Riemann et al.² that demonstrated a 15% reduction in bilateral hippocampus volumes. That study used a 1.5 T magnet and included only eight patients with severe PI. In three previous publications,^3–5 and again in our current analysis of Study 2 data, loss of hippo-campal volume was not observed in PI. However, very similar to our previously published data from Study 1,⁴ we found correlations between left (but in distinction to Study 1, not right) hippocampal volumes and actigraphically derived sleep continuity measures. Thus, although the overall group differences in hippocampal volumes seen by Riemann et al.² could not be confirmed in either Study 1 or 2, our data suggest an association between smaller volumes in this structure and worse sleep in PI. One possible explanation for the discrepancies between the Riemann et al.² study and our two studies is that the sleep of patients with insomnia in the Reimann et al.² study was much more disturbed than in our PI groups (mean sleep efficiencies by sleep diary of 56.8% in Riemann et al.² and 77.9% and 80.5% in Study 1 and 2, respectively). Therefore, there may be a sleep disturbance severity threshold for hippocampal changes that patients with PI in our two studies did not reach. Consistent with this explanation, Noh et al.⁵ did not find group differences in hippocampal volumes between control and moderately poor-sleeping patients with PI (sleep diary-derived mean sleep efficiency of 74.1%) but did observe significant negative correlations between right and left hippocampal volumes and polysomnographically derived arousal index. Like Altena et al.,³ we did observe volume reduction in the orbitofrontal cortex in patients with PI in Study 1, though we found this on the left side whereas volume reduction in their patients was on the right side. Altena et al.³ used voxel-based morphometry and patients who were substantially older than ours (mean of 60 y versus 39 y), making the two studies difficult to compare. The Riemann et al.² data also quantified bilateral ACC volumes and observed no differences between patients with PI and control patients. In their study, the anatomical limits of the ACC ROI were not specified, but, in addition to the rostral and ventral ACC, which were analyzed in our current study, they probably included the larger caudal ACC, whose function is distinct from that of these more anterior portions of the ACC.⁵¹ In addition, as described previously, the patients evaluated in that study had more severe insomnia symptoms than those included in the current studies.

A number of limitations should be considered when assessing the results of this study. Screening polysomnography was only performed on patients with PI from Study 1 and not on patients from either control group. Thus, it is possible that the observed rACC volume differences between groups were based on an unequal distribution of noninsomnia sleep disorders assessable by polysomnography (e.g., sleep apnea, periodic limb movement disorder). We believe that this is unlikely because other predictors of such sleep disorders (e.g., BMI) were equivalent between groups and clinical evaluation was carefully performed to rule out such disorders. The rostral ACC ROI was identified by automated segmentation rather than by manual determination. This may have introduced a systematic bias into the anatomical region such that adjacent areas not in the ACC were included in the analysis. However, the automated segmentation package (FreeSurfer) is a widely used instrument, with excellent correlation established between manual and FreeSurfer determination of ACC size,³⁰ and preprocessing was thoroughly performed by a single experienced investigator (AG). In addition, if error was introduced by the automated software, it is more likely that a nonsystematic bias would be introduced, which should minimize any true difference between the control and PI groups. Because Study 1 and 2 had different data acquisition parameters, scanner operating systems, and patient characteristics (e.g., percentage of female patients) we did not combine the two datasets. Similarly, we did not test correlations between ACC GABA levels⁶ and ACC volume in Study 2 because the examined voxels only partially overlap using these two methods. It is possible that previous exposure to hypnotic medications was responsible for the rACC volume increases we observed in the PI group, as has been seen with lithium in patients with bipolar disorder.⁵² However, most of our patients with PI were hypnotic naive, and none had taken such medications on a regular basis. Finally, we performed multiple correlations of rACC volume and sleep variables, and thus these data should be considered preliminary, though the consistency of findings suggest that these associations are valid in these two samples of patients with PI.

CONCLUSIONS

We found larger rACC bilaterally in two independent groups with chronic PI patients compared with healthy sleeper controls. Moreover, in subjects with chronic PI, rACC volumes were positively correlated with self-reported and objective determinations of poor sleep quality. Chronic PI increases the risk of future depression and the larger rACC volumes seen in this study may represent a compensatory response to repetitive sleep disturbance and be a marker of resilience to developing a mood disorder. Future longitudinal studies examining ACC volumes in individuals with chronic insomnia, or studies using both sleep variables and ACC size as covariates in those with MDD, may provide a better understanding of the relationship of insomnia, ACC volume, and MDD.

DISCLOSURE STATEMENT

This study was supported in part by a research grant from Sepracor, Brigham and Women's Hospital General Clinical Research Center grant M01-RR02635, the American Sleep Medicine Foundation Physician Scientist Training Award, and NIH grant MH58681. Dr. Buxton is in part supported by NIH grant R01HL107240. Dr. Winkelman reports serving as a consultant or advisory board member for Pfizer, Sunovion, UCB, and Zeo; research support from GlaxoSmithKline, Impax Pharma, and Sepracor; and stock options in Zeo Inc. Dr. Plante reports having owned stock in Pfizer, and having received honoraria from Oakstone Medical Publishing and royalties from Cambridge University Press. The other authors have indicated no financial conflicts of interest.