Reduced resting‐state functional connectivity and sleep impairment in abstinent male alcohol‐dependent patients

Abstract Alcohol dependence is associated with poor sleep quality, which has both been implicated with thalamocortical circuits function. To identify the possible roles of these circuits in the alcohol‐sleep association, we investigated the volume of both left and right thalamus and corresponding resting‐state functional connectivity (RSFC) differences between 15 alcohol‐dependent patients (AD) and 15 healthy controls (HC) male participants. The neuroimaging findings were then correlated with clinical variables, that is, Alcohol Use Disorders Identification Test (AUDIT) and Pittsburgh Sleep Quality Index (PSQI). Additionally, mediation analysis was carried out to test whether the thalamocortical RSFC mediates the relationship between drinking behavior and sleep impairments in AD when applicable. We observed a significant positive correlation between AUDIT score and PSQI score in AD. Compared with HC, AD showed reduced RSFC between the left thalamus and medial prefrontal cortex (mPFC), orbitofrontal cortex, anterior cingulate cortex (ACC), and right caudate. We also observed a negative correlation between RSFC of the left thalamus–mPFC and PSQI score in AD. More importantly, the left thalamus–mPFC RSFC strength mediated the relationship between AUDIT score and PSQI score in AD. No significant difference was detected in the normalized volume of both left and right thalamus, and volumes were not significantly correlated with clinical variables. Our results demonstrate that AD show abnormal interactions within thalamocortical circuits in association with drinking behaviors and sleep impairments. It is hoped that our study focusing on thalamocortical circuits could provide new information on potential novel therapeutic targets for treatment of sleep impairment in alcohol‐dependent patients.


| INTRODUCTION
Alcohol dependence is a psychiatric disorder characterized by a compulsive drive toward alcohol consumption and inability to inhibit its consumption despite negative consequences. Individuals with alcohol dependence report poor sleep and disturbances of sleep (Luc et al., 2006) and the prevalence of sleep problems has been reported to be as high as 91% in alcohol-dependent (AD) patients (Zhabenko, Wojnar, & Brower, 2012).
For instance, high alcohol consumption is associated with decreases in rapid-eye movement (REM) sleep, sleep continuity and sleep efficiency, and also increases in sleep latency (time to fall asleep) and sleep duration (Chaput, McNeil, Després, Bouchard, & Tremblay, 2012). These disturbances have caused serious impacts on the overall health and social wellbeing of the individual. Therefore, it is of great significance to investigate the underlying neural mechanisms of sleep impairments in AD patients.
Taken together, alcohol dependence is associated with significant sleep disturbances (Chakravorty, Chaudhary, & Brower, 2016), which are often accompanied by abnormal thalamus dysfunction . The thalamus plays an essential role in the dorsal relay of the brainstem reticular activating system, controlling cortical activation that occurs during wakefulness and REM sleep (Thakkar, Sharma, & Sahota, 2015). The thalamus is not a unitary structure but is composed of several nuclei which cooperate to complete various functions through strong connections with neocortex via radiating fibers (Sherman, 2016). Thalamocortical function has been intimately tied to both addiction and sleep in prior research. For instance, cigarette smokers showed reduced connectivity between the thalamus and the dorsolateral PFC, and the strength of the dorsolateral PFC-thalamus connectivity was negatively associated with severity of nicotine dependence . Similar results were also shown in ketamine-dependent individuals (Liao et al., 2016). In addition, thalamocortical neurons are essential for thalamocortical information transfer and regulation of arousal and sleep (Del Felice, Formaggio, Storti, Fiaschi, & Manganotti, 2012;Krone et al., 2017). These findings led us to hypothesize that thalamus centered RSFC mediates alcohol dependence and sleep impairment in AD patients. Therefore, the purposes of this study were to (a) assess the relationship between sleep quality (i.e., PSQI) and drinking behaviors in AD patients; (b) detect the volume and RSFC differences of thalamus between AD patients and controls and estimate their associations with drinking behaviors and sleep quality; (c) test the possible mediator role of the thalamus as the relationship between alcohol dependence and sleep impairment in drinkers. We hoped that by linking circuit-level interactions between brain regions (especially the thalamus and PFC) to particular dimensions of behavioral dysfunction, we could shed new light on the neural associations of sleep impairment in AD patients.

| MATERIALS AND METHODS
All recruitment and testing procedures were reviewed and approved by the ethics committee of medical research in Xilinguole Meng Mongolian General Hospital, Xilinhaote, Inner Mongolian, China. Experimental procedures of this study (psychometric interviews and MRI) were explained and all participants provided written informed consent to participate. Participants were compensated $50.

| Participants
This study included 15 AD patients and 15 age-and gender-matched healthy controls (HC) male participants. We instructed participants to have a good night's sleep and not to drink alcohol 48 hr before the study days. Inclusion criteria for AD patients were as follows: (a) ages 21-55 years; (b) current alcohol dependence met the Diagnostic and Statistical Manual of Mental Disorders, 4th edition (DSM-IV); (c) no DSM-IV axis I disorder or biomedical condition that may have adversely affected brain neurobiology; (d) no current or history of substance use disorder (other than alcohol dependence in AD group); and (e) participants enrolled in the current study were nonsmokers. Healthy controls were identified as those who showed negative test results for the diagnostic criteria of alcohol abuse in DSM-IV. Participants in this study did not drink coffee at all, and so they were not instructed one way or the other. All of the participants were right-handed as measured by the Edinburgh Handedness Inventory. Demographic data and study characteristics of the sample are shown in Table 1.
Exclusion criteria for both groups were (a) history of neurological or other physical diseases such as respiratory, cardiac, renal, hepatic, and endocrinal diseases; (b) any penetrating head trauma and closed head injury with loss of consciousness >10 min; (c) any medications currently that may affect brain functioning within 2 weeks; and (d) any current use of illicit substances, verified by toxicology screening.
The neuropsychiatric interview was performed by psychiatrists on the day of the fMRI scan. The sleep measurement of participants in this study was assessed by the self-reported Pittsburgh Sleep Quality Index (PSQI) score, which combines seven component scores: Subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleeping medication, and daytime dysfunction. The PSQI is a well-validated 19-item self-report questionnaire that assesses sleep quality and quantity over the past month, which is considered to be a good measure of insomnia (Dietch et al., 2016). Scores range from 0 to 21 and higher scores represent worse sleep.
The measure of alcohol dependence severity was evaluated with the Alcohol Dependence Scale (ADS; Skinner & Allen, 1982); the frequency with which individuals engage in various indicators of alcohol abuse measured by the Alcohol Use Disorders Identification Test (AUDIT; Babor, Higgins-Biddle, Saunders, & Monteiro, 2001). The AUDIT developed by the World Health Organization (WHO) as a selfreport screening test to identify severity of alcohol use disorders and provide an overall measure of hazardous drinking. Hazardous use, dependence symptoms, and harmful use were the three symptom areas covered by the 10-item scale (Lundin, Hallgren, Balliu, & Forsell, 2015;World Health Organization, 2001). Total scores range from 0 to 40 and higher scores represent worse drinking. In addition, the participants also completed standardized questionnaires assessing depressive and anxiety symptomatology by Zung Self-Rating Anxiety Scale (SAS; Zung, 1971) and Zung Self-Rating Depression Scale (SDS; Zung, 1965). The impulsivity was measured by the Barratt Impulsiveness Scale (BIS-11; Patton, Stanford, & Barratt, 1995).

| MRI data acquisition
Functional imaging data were acquired using a 3T Philips scanner (Achieva; Philips Medical Systems, Best, The Netherlands). Prior to the MRI scanning, subjects were instructed to refrain from alcohol and other drugs consumption for the prior 48 hr, which was confirmed with a urine drug screen and a breath test for alcohol. For each participant, the highresolution three-dimensional T1-weighted image was acquired using a magnetization-prepared rapid gradient echo awake, and not to think about anything systematically. After the data acquisition, subjects were asked whether or not they remained awake during the whole procedure.

| Structural MRI data analysis
The intracranial volume (ICV) and regional brain volumes were obtained by applying the FreeSurfer 5.0 (http://surfer.nmr.mgh. harvard.edu) longitudinal processing pipeline on the T1-weighted MPRAGE images as described in our previous studies (Cai et al., 2016;Li et al., 2015;. The processing procedure consisted of (a) skull stripping; (b) automated Talairach transformation; (c) segmentation of the subcortical white matter and deep gray matter volumetric structures; (d) intensity normalization; (e) tessellation of the gray matter/white matter boundary; (f) automated topology correction; (g) surface deformation; (h) registration of the subjects' brains to a common spherical atlas.

| Resting-state MRI data analysis
Analysis of Functional NeuroImages software (AFNI, http://afni.nimh.nih. gov/) and FMRIB's Software Library (FSL, http://www.fmrib.ox.ac.uk/fsl/) were used for fMRI resting-state images analysis. As described in our previous study , the workflow of resting-state MRI images preprocessing was divided into eight sections, dropping the first 5 volumes, slice-timing correction, rigid-body motion correction, removal skull, spatial smoothing, affine registration to the skullstripped structural image, spatially normalized into MNI152 template, and intensity normalization. It is inadequate to control the noise induced by head motion using nuisance regression and bandpass filtering alone (Patel et al., 2014;Power, Barnes, Snyder, Schlaggar, & Petersen, 2012). Therefore, wavelet despiking was applied for the T A B L E 1 Demographic characteristics of AD and HC in this study Values are mean ± SD unless otherwise indicated. Bold values refer to the significant differences.
resting-state functional connectivity (RSFC) analyses (Patel et al., 2014). The denoising steps were as follows: (a) time series despiking (wavelet domain); (b) nuisance signal regression (14-parameters regression); (c) a temporal Fourier filter (0.009-0.10 Hz). The regions of both left and right thalamus were chosen as our seeds, which were labeled according to Harvard-subcortical structural atlas (http://www. cma.mgh.harvard.edu/). The regional resting-state fMRI time series was generated by using the average functional time series of all voxels extracted for each ROI in the template atlas. To investigate the strength of RSFC between each ROI and the other voxel within the brain, Pearson correlation was carried out respectively for left thalamus and right thalamus, and then Fisher's r-to-z transformation was employed.

| Statistical analysis
The demographic characteristics and the volume of both left and right thalamus were extracted and imported into the SPSS 20.0 (SPSS Statistics, IBM, Armonk, NY). Demographic characteristics were compared by using independent t-tests between AD and HC. Pearson correlations were applied between the poor sleep quality measure (PSQI total score) and the alcohol dependence measure (AUDIT score). All tests were twotailed, and the level of significance was p < .05.
For each structure, the volumes for both the left and right thalamus were normalized by the brain volume and group analyses were conducted using independent two-sample t tests between AD and HC. Bonferroni procedure was employed to correct for multiple comparisons. All tests were two-tailed, and the level of significance was p < .025 (.05/2).
Then two-sample t tests were used to compare z value maps between AD and HC. Permutation-based nonparametric testing with 5,000 random permutations was used to investigate the group comparisons of functional connectivity. To control for multiple comparisons, threshold-free cluster-enhancement (TFCE) was used, and the significance threshold was set to p < .05, familywise-error (FWE) corrected in line with current reporting guidelines (Eklund, Nichols, & Knutsson, 2016). The correlations between the neuroimaging findings F I G U R E 1 (a-h) PSQI scores in alcohol-dependent patients and healthy controls. (i, j) The correlation between the PSQI score and the AUDIT score. Alcohol-dependent patients committed higher scores than controls in sleep efficacy and total score. Correlation analysis revealed significant positive correlations between the PSQI and AUDIT score in alcohol-dependent patients (i, r = .7695, p = .0008). With regard to the controls, no significant correlations were detected between PSQI score and AUDIT score [Color figure can be viewed at wileyonlinelibrary.com] (i.e., abnormal RSFC of thalamus with other regions) and behavioral data in AD (i.e., AUDIT, PSQI) were assessed by Pearson correlations.  (Kober et al., 2010), "X" refers to the AUDIT score, "Y" refers to the PSQI score, and "M" refers to left thalamus-mPFC RSFC strength in current study .

| Demographics
Detailed demographical analysis showed there were no significant differences for age between two groups. Significant group differences were observed for multiple drinking measures such that AD patients reported higher ADS score (t = 3.83, p = .001) and higher AUDIT score (t = 4.60, p < .001). Additionally, AD exhibited higher scores in selfreports of anxiety (t = 2.70, p = .012) and self-reports of depression (t = 2.18, p = .038) compared to controls. However, there were no significant group differences in BIS-11 (t = 2.00, p = .056; Table 1).
AD showed a significant positive correlation between the sleep quality (PSQI total scores) and alcohol use disorder measure (AUDIT score; r = .6975, p = .0008).

| Thalamus volumes differences
Two-sample t tests indicated that there was no significant group difference in volume for the normalized left thalamus (t = 0.67, p = .50) and right thalamus (t = 0.37, p = .071; Figure 2 and Table 2).  Figure 3a). In addition, the right thalamus exhibited decreased RSFC with bilateral mPFC, OFC, and left caudate in AD group (Table 3 and Figure 3b).

| Brain-behavior correlation results
Pearson correlation was applied between the 8 and 5 regions that showed abnormal RSFC with left and right thalamus, respectively, and clinical variables (i.e., AUDIT, PSQI). All tests were two-tailed, and to correct for multiple comparisons we used Bonferroni correction (p < .0019 [.05/26]). According to these criteria, the PSQI score was negatively correlated with the RSFC strength between the left thalamus and ipsilateral medial prefrontal cortex in AD (r = .7479; p = .0013; Figure 3c). Correlation analysis revealed significant negative correlations between RSFC strength of left thalamus-mPFC and AUDIT score (r = −.64; p = .01

| Mediation analysis
Given the finding that PSQI score was correlated with left thalamus-mPFC RSFC strength in AD, together with the correlation with AUDIT score, mediation modeling was used to test whether the magnitude of alcohol dependence effect on sleep impairment depended on impaired left thalamus-mPFC RSFC (Hayes, 2012).
Mediation analysis was applied among AUDIT score, PSQI score, and left thalamus-mPFC RSFC strength in AD. As expected, we revealed that the RSFC strength of the left thalamus-mPFC pathway partly mediated the relationship between AUDIT score and PSQI score

| DISCUSSION
We observed that the RSFC of left thalamus-mPFC with PSQI score, which provided novel insights into the alcohol dependence-sleep impairments associations in abstinent males. We also for the first time Values are mean ± SD unless otherwise indicated.
to date, formally demonstrate that the RSFC of left thalamus-mPFC played a mediating role in the association of alcohol dependence and poor sleep quality. We computed the connectivity strength and showed in mediation analyses that problem drinking is related to sleep impairment via decreased thalamocortical connectivity. This study may provide new insight for sleep impairment in male AD patients, and this F I G U R E 3 (a, b) The resting-state functional connectivity (RSFC) of thalamic differences between alcohol-dependent patients and controls. (c, d) The correlation between functional connectivity (medial prefrontal cortex-thalamus) and PSQI score. Significantly lower resting-state functional connectivity (RSFC) showed between left thalamus and several regions in alcohol-dependent patients (a, FWE corrected, p < .05), that is, bilateral medial prefrontal cortex (mPFC), orbitofrontal cortex (OFC), anterior cingulate cortex (ACC), left angular gyrus, and right caudate. In addition, the right thalamus exhibited decreased RSFC with bilateral mPFC, OFC, and left caudate in alcohol-dependent patients (b, FWE corrected, p < .05). Significant negative correlation (r = .7479; p = .0013) was found between the left medial prefrontal cortex-thalamus RSFC strength and PSQI score in alcohol-dependent patients ( in turn has implications for treatment because of the brain areas identified.

| Sleep impairments in AD patients
In this study, we revealed the poor subjective sleep quality in AD patients ( Figure 1). The association between AUDIT score and PSQI score in AD patients was also found the chronic effect of alcohol on the sleep impairment among drinkers. The relationship between alcohol dependence and sleep impairments is of great complexity. While acute alcohol intake in nonalcoholic social drinkers reduces sleep onset latency and improves NREM sleep (Roehrs & Roth, 2012), alcohol dependence is generally associated with sleep impairments (Thakkar et al., 2015) although the data vary widely across studies (Chakravorty et al., 2016;Hartwell, Bujarski, Glasner-Edwards, & Ray, 2015;Zhabenko et al., 2012). On the other hand, alcohol causes adenosine buildup to disrupt sleep homeostasis. Adenosine can inhibit wake-active neurons via adenosine A 1 receptor thus disinhibiting sleep-active neurons, which also stimulates sleep-active neurons via A 2A receptors (Alam & McGinty, 2017). In addition, adenosine is a key mediator of neuronal responses to alcohol (Newton & Messing, 2006). This property permits alcohol to have a secondary effect on virtually all neurotransmitter/neuromodulator systems related to homeostatic regulation of sleep within the brain. Therefore, it follows that alcohol has great impact on sleep. Taken together, by revealing AD patients had a greater association between AUDIT score and PSQI score than controls, we provided partial evidence for the sleep impairment in AD patients.

| RSFC of the left thalamus to mPFC mediates the relationship between alcohol dependence and sleep impairments in AD patients
We noticed that AUDIT score was correlated with PSQI score only in AD patients ( Figure 1). Moreover, the left thalamus-mPFC RSFC strength was correlated with PSQI score as well as AUDIT score in AD patients (Figure 3). We suggested that thalamocortical circuits RSFC strength could be a biomarker of sleep quality and alcohol addiction severity. We proposed that thalamocortical RSFC (i.e., thalamus-mPFC) mediated the relationship between AUDIT score and PSQI score, which was confirmed by our mediation analysis results (Figure 4). We suggest that alcohol drinking regulated the sleep quality via the reduced thalamocortical RSFC in AD patients.
The thalamus is critical in sleep homeostasis through its regulation by adenosine to its wake-promoting cholinergic neurons innervating the basal forebrain (Brown, Basheer, McKenna, Strecker, & McCarley, 2012). Adenosine A1 receptors are upregulated in the PFC, which may consequently produce inhibition of PFC activity (Elmenhorst et al., 2007). These findings may support the theory that alcohol, via

| Reduced RSFC with cortico-striato-thalamocortical loop circuits in AD patients
We also provided evidence for the role of the thalamus-centered circuits in alcohol dependence. Anatomically, thalamus receives projections from multiple cortical and midbrain regions, which is integral to the circuits that underlie the reward (Haber & Calzavara, 2009) and response inhibition processes mediating control processes in goal-oriented behaviors (de Bourbon-Teles et al., 2014;Phillips, Kambi, & Saalmann, 2016). Previous neuroimaging studies observed structural and functional differences of thalamus between substance use disorder and healthy controls (Feil, Sheppard, Fitzgerald, Yücel, & Lubman, 2010). Several studies showed that reduced thalamic connectivity was associated with increased craving and other measures of severity of drug dependence (Liao et al., 2016;Wang et al., 2017). However, the implication of thalamocortical circuits in neural mechanisms of AD patients remains unclear. Our findings contribute toward filling this gap by showing the reduced RSFC between thalamus and a series of regions in AD patients (Figure 3), such as mPFC (involved in cognitive control), OFC (involved in goal-directed behaviors), ACC (involved in inhibitory control and awareness), and caudate (involved in reward; Beck et al., 2012;Bruno & Sakmann, 2006;Cai et al., 2016;Li et al., 2015;Yuan et al., , 2016Yuan, Yu, Cai, et al., 2017). Tract tracing studies in nonhuman primates have determined that the striatum projects via the globus pallidus to the thalamus, which then projects back to the frontal cortex (Haber & McFarland, 2001 (Klumpp et al., 2017;Klumpp, Hosseini, & Phan, 2018). Objective measures, such as polysomnography should be used to further investigate the association of alcohol dependence and sleep in future study. Second, our sample was relatively small, and therefore our group comparisons were not powered to detect small effects.
Third, the thalamus is a very complex brain region and has distinctive subregions with different functions; accordingly, RSFC in distinctive subregions of thalamus in AD could be taken into consideration in future studies. Additionally, we cannot extend our findings to females with AD as our sample included only males. Due to the lack of longitudinal experiment design, we failed to show whether the sleep disorder came before alcohol dependence or the other way around. In the current study, we focused on the sleep impairments in AD patients. Still, we agree that the associations of PSQI and anxiety and depression in the AD patients should be investigated in future work. Recently, Dopamine D2 receptors has been implicated in sleep deprivation, and in alcohol and other substance use disorders (Volkow et al., 2008(Volkow et al., , 2012Wiers et al., 2016). In the current study, we did not collect objective sleep measurements for the night before the scanning. Thus, we cannot exclude the possible acute effects of sleep on our resting state results.

| CONCLUSION
We revealed reduced RSFC within cortico-striato-thalamo-cortical loops that were correlated with drinking behavior and sleep impair-

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.