Resting state synchrony in long-term abstinent alcoholics with versus without comorbid drug dependence

Abstract

Background

We previously reported that when long-term abstinent alcoholics (LTAA; with no drug comorbidity) are compared to controls, they show increased resting state synchrony (RSS) in the executive control network and reduced RSS in the appetitive drive network suggestive of compensatory mechanisms that may facilitate abstinence. The aim of the present study was to investigate whether long-term abstinent alcoholics with comorbid stimulants dependence (LTAAS) show similar RSS mechanisms.

Methods

Resting-state functional MRI data were collected on 36 LTAAS (20 females, age: 47.85 ± 7.30), 23 LTAA (8 females, age: M = 47.91 ± 6.76), and 23 non-substance abusing controls (NSAC; 8 females, age: M = 47.99 ± 6.70). Using seed-based measures, we examined RSS with the nucleus accumbens (NAcc) and the subgenual anterior cingulate cortex (sgACC).

Results

Results showed commonalities in LTAA and LTAAS RSS (similar enhanced executive control RSS and left insula RSS) as well as differences (no attenuation of appetitive drive RSS in LTAAS and no enhancement of RSS in right insula in LTAA).

Conclusions

We believe these differences are adaptive mechanisms that support abstinence. These findings suggest common as well as specific targets for treatment in chronic alcoholics with vs without comorbid stimulant dependence.

1. Introduction

Alcoholism is a disorder with an overwhelming impulsive and compulsive “drive” toward alcohol consumption (Kamarajan et al., 2005), and an inability to “hit the brakes” or inhibit alcohol consumption. It is important to know whether brain functional organization is different in alcohol dependence after long-term abstinence. Our group has reported that when long-term abstinent alcoholics (LTAA) are compared to controls, they show increased synchrony in the executive control network and reduced synchrony in the appetitive drive network (Camchong et al., 2013a). These findings suggest that there are adaptive mechanisms in functional organization at rest in long-term abstinence that may facilitate the behavioral control required to maintain abstinence (i.e., enhanced inhibitory control and reduced appetitive drive). We also identified these adaptive mechanisms in a short-term abstinent sample, but not to the same extent, suggesting that synchrony within the appetitive drive network progressively decreases and the synchrony within the executive control network progressively increases during abstinence (Camchong et al., 2013b).

In our previous study, LTAA showed lower resting state synchrony (RSS) than non-substance abusing controls (NSAC) within the appetitive drive network between regions such as nucleus accumbens, thalamus and caudate (Camchong et al., 2013a). These regions have specific but interactive roles in reward processing such that NAcc first mediates the definition of a reward, thalamus relays this information to cortex to guide behavior toward the reward, and, if the reward is processed numerous times, caudate facilitates habit formation (Everitt and Robbins, 2005; Vollstadt-Klein et al., 2010). We also reported enhancement of RSS within an executive control network, particularly between subgenual anterior cingulate cortex (sgACC) and dorsolateral prefrontal cortex (DLPFC). The level of synchronization between these regions increases or decreases in relation to cognitive demand (Hare et al., 2009; Medalla and Barbas, 2009). This suggests that LTAA have an adaptive mechanism evident during rest characterized by attenuated interaction of appetitive drive regions (to avert behavior toward alcohol) and enhanced interaction of executive control regions (to enhance behavioral regulation), a mechanism that may support abstinence maintenance.

While above findings are based on abstinent individuals with alcohol dependence only, it is increasingly more likely to find alcoholics with comorbid dependence on stimulants (e.g., methamphetamine or cocaine (SAMHSA, 2010)). Alcohol and stimulants increase the levels of extracellular dopamine in nucleus accumbens; however, they have different mechanisms of action. Alcohol indirectly increases extracellular dopamine levels by binding to dopamine receptors, whereas stimulants directly trigger dopamine release. The high incidence and morbidity of alcohol and stimulant dependence in the U.S. population highlights the need to understand whether the brain mechanisms and adaptations that support LTAA without comorbid stimulant dependence (Camchong et al., 2013a) are also found in long-term abstinent alcoholics with comorbid stimulant dependence.

While it is difficult to achieve long-term abstinence from alcohol only, it is especially difficult to achieve abstinence when alcoholics have comorbid stimulant use. The current study investigated whether long-term abstinent individuals who were dependent on both alcohol and stimulants (methamphetamine and cocaine) have similar patterns suggestive of adaptive mechanisms as previously found in LTAA with history of alcohol dependence only (Camchong et al., 2013a).

We hypothesize that individuals with long-term abstinence from alcohol and stimulants (LTAAS) will either (a) have similar RSS to LTAA, suggestive of compensatory mechanisms with higher RSS of executive control regions and lower RSS of appetitive regions than healthy controls, or (b) due to the added dopaminergic effects of stimulants in appetitive drive regions, will show less attenuation of RSS in the appetitive drive network. Given the relationship between RSS and behavior (Mennes et al., 2010, 2011; Zhu et al., 2011), we hypothesized that RSS in the: (1) executive control network in the LTAAS group will be positively correlated with performance on a task that assesses executive control, (2) appetitive drive network in the LTAAS group will be positively correlated with behavior ruled by appetitive drive, such as antisocial symptoms, and (3) appetitive drive network within the LTAAS group will be positively correlated with alcohol, methamphetamine, and cocaine use measures.

An increasingly number of neuroimaging studies point to the important role of insula in addiction in behavioral aspects such as stress coping, decision-making or cue-induced tasks (Naqvi and Bechara, 2010). More importantly, reduced insula activity in stimulant addicts during a decision-making task (Paulus et al., 2005) or an attention task (Clark et al., 2012) may predict subsequent relapse. Because insula has reciprocal connections with both of our executive control and appetitive drive seeds (sgACC and NAcc) (Craig, 2009; Kelly et al., 2012), we expect to find differences in insular functional organization in LTAAS (perhaps increased RSS) when compared to LTAA or NSAC.

2. Methods

2.1. Participants

All subjects were recruited from the island of Oahu. We compared 36 long-term abstinent alcoholics with stimulant (methamphetamine/cocaine) dependence (LTAAS) with the age and gender comparable samples of 23 long-term abstinent alcoholics (LTAA) and the 23 non-substance abusing controls (NSAC) from our previous paper (Camchong et al., 2013a) (Table 1). All participants provided written informed consent and received monetary compensation. The consent process and all procedures were reviewed and approved by the institutional review board (IRB) at Queens Medical Center. Long-term abstinent individuals were required to have at least 18 months of abstinence at study entry and met DSM-IV lifetime criteria for alcohol dependence. LTAAS met lifetime criteria for methamphetamine or cocaine dependence (Table 2). Subjects completed the computerized Diagnostic Interview Schedule (cDIS; Levitan et al., 1991) to ascertain externalizing, anxiety or mood disorder diagnoses and symptom counts (Table 3). Antisocial personality disorder measures, inclusion/exclusion criteria and family history of substance use problems can be found in Supplementary Materials.

Table 1.

Demographics, recreational substance use and alcohol measures in LTAAS, LTAA, and NSAC.

	LTAAS (n = 36)		LTAA (n = 23)		NSAC (n = 23)
Characteristic	Mean or n	SD or %	Mean or n	SD or %	Mean or n	SD or %	F or χ2
Age, (yrs)	47.85	7.30	47.91	6.76	47.99	6.70	0.003
Education, (yrs) Subject	13.79	2.65	13.70	1.94	15.39	2.08	4.14*,a
Female, n (%)	20	55.56%	8	34.78%	8	34.78%	3.54
Nicotine Dependence
Lifetime, n (%)	14	38.89%	5	27.74%	2	8.70%	6.97*
Current, n (%)	7	19.44%	2	8.70%	0	–	5.60
Marihuana Dependence
Lifetime, n (%)	8	22.22%	0	–	0	–	11.33**
Current, n (%)	0	–	0		0		–
Proportion of 1st degree relatives who are problem drinkers (total number/number of relatives)	0.32	0.31	0.33	0.27	0.14	0.19	3.57*,b
Proportion of 1st degree relatives who are drug users (total number/number of relatives)	0.20	0.27	0.11	0.21	0.04	0.10	3.86*,c
Age started drinking (yrs)	13.67	3.22	15.17	4.12	–	–	–
Average Alcohol Dose (Lifetime; standard number of drinks per month)	198.43	157.38	188.51	163.03	–	–	–
Length of Abstinence from Alcohol (number of days)	2069.00	1749.05	2888.78	2848.14	–	–	–

LTAAS, long-term abstinent from alcohol and stimulants; LTAA, long-term abstinent alcoholics; NSAC, non-substance abusing controls; SD, standard deviation.

^aTukey Post Hoc tests showed that NSAC had significantly higher levels of education when compared to LTAAS (p = 0.04) and LTAA (p = 0.03). Level of education was not significantly different between LTAAS and LTAA (p = 0.89).

^bTukey Post Hoc tests showed that NSAC had a lower proportion of 1st degree relatives who are problem drinkers when compared to LTAAS (p = 0.051) and LTAA (p = 0.061). Proportion of 1st degree relatives who are problem drinkers was not significantly different between LTAAS and LTAA (p = 0.99).

^cTukey Post Hoc tests showed that LTAAS had a significantly higher proportion of 1st degree relatives who are problem drug users when compared to NSAC (p = 0.02). Proportion of 1st degree relatives who are problem drug users was not significantly different between LTAAS and LTAA (p = 0.29).

^*p < 0.05.

^**p < 0.01.

Table 2.

Breakdown of Substance Use in LTAAS Group.

Characteristic	Only meth dependence (n = 8)		Only cocaine dependence (n = 16)		Cocaine and meth dependence (n = 12)
	Mean	SD	Mean	SD	Mean	SD
Average lifetime alcohol dose (standard number of drinks per month)	216.66	140.71	183.45	111.99	162.58	167.40
Alcohol dose during peak use (number of drinks per month)	265.08	171.82	393.85	401.05	268.33	195.84
Length of abstinence from alcohol (number of days)	1594.88	1317.19	2968.44	2489.97	2504.25	1911.62
Average lifetime meth dose (grams per month)	34.25	24.92	–	–	20.91	15.60
Meth dose during peak use (number of days)	38.56	30.56	–	–	23.80	18.28
Length of abstinence from meth (grams per month)	1761.43	1340.25	–	–	2728.91	1888.12
Average Lifetime Cocaine Dose	–	–	39.85	93.61	23.31	27.98
Cocaine Dose During Peak Use	–	–	47.25	91.75	48.52	56.32
Length of Abstinence from Cocaine (number of days)	–	–	3184.28	3219.36	4406.17	2345.63

LTAAS, long-term abstinent from alcohol and stimulants; SD, standard deviation

Table 3.

Current and lifetime psychiatric diagnoses in alcoholic groups.

Psychiatric diagnoses	Current diagnoses count			Lifetime diagnoses count
	LTAA (n = 23) Count (n)	LTAAS (n = 36) Count (n)	LTAA vs LTAAS Odds Ratio	LTAA (n = 23) Count (n)	LTAAS (n = 36) Count (n)	LTAA vs LTAAS Odds Ratio
Mood	3	9	0.45	13	17	1.45
Dysthymia	0	0	N/A	1	0	∞
Manic episode	0	2	∞	2	2	1.62
Hypomanic	0	1	∞	0	1	∞
Major depressive disorder	2	5	0.59	10	13	1.36
Bipolar	1	3	0.50	3	3	1.65
Anxiety	1	6	0.23	7	12	0.88
Agoraphobia	0	0	N/A	2	2	1.62
Obsessive compulsive	0	0	N/A	0	0	N/A
Panic disorder	0	2	∞	1	3	0.50
Social phobia	0	0	N/A	0	0	N/A
Post-traumatic stress disorder	1	5	0.28	6	11	0.80
Attention deficit hyperactivity disorder	2	1	2	4	6	1.05
Antisocial personality disorder	2	6	3.33	3	17	0.17*
Conduct disorder	0	0	N/A	3	18	0.15*

Significant χ² (2-sided):

^*p < 0.05; LTAAS, long-term abstinent from alcohol and stimulants; LTAA, long-term abstinent alcoholics.

2.2. Behavioral task

In a separate session before the scan, participants completed a task that requires cognitive flexibility, an important behavioral construct needed to maintain long-term abstinence. The Intradimensional/extradimensional set shift task (IED; Cambridge Neuropsychological Test Automated Battery) measured the ability to learn associations between a stimulus and response and to switch to new associations after stimulus contingencies are reversed.

2.3. Imaging data acquisition

Resting-state functional connectivity magnetic resonance imaging (fcMRI) data were collected using a twelve-channel head coil on a Siemens Tim Trio 3.0 T scanner (Siemens Medical Solutions, Erlangen, Germany) located at Queen’s Medical Center in Honolulu. Subjects were instructed to lay motionless in the scanner with eyes closed. The imaging sequence was a gradient-echo spiral in/out sequence with TE = 30 ms, TR = 2000 ms, flip angle = 60°, 28 interleaved axial 5 mm thick contiguous slices, FOV = 22 cm, and a 3.44 mm × 3.44 mm in-plane resolution (64 × 64 matrix size; Glover and Law, 2001; Noll et al., 1995). The resting-state scan acquired 123 volumes with a scan time of 4:06.

A high-resolution T1-weighted structural image was acquired (MPRAGE) with parameters of TE = 4.11 ms TR = 2200 ms, flip angle = 12°, 160 sagittal slices, slice thickness = 1 mm, slice gap = 0.5 mm, FOV = 256 mm.

2.4. FcMRI data preprocessing

All imaging data was preprocessed using AFNI (Analysis of Functional NeuroImages) and FSL (FMRIB Software Libraries; Oxford, United Kingdom). Preprocessing consisted of: dropping first 3 TRs to account for magnet field homogenization; slice-time correction; three-dimensional motion correction; skull stripping; temporal despiking; spatial smoothing (6 mm); mean-based intensity normalization; temporal band-pass filtering (0.009–0.1 Hz); and linear and quadratic detrending.

Probabilistic independent component analysis was conducted for each individual to denoise individual data by removing components that represented noise such as head motion, scanner artifacts, and physiological noise base on spatial and temporal characteristics detailed in the MELODIC (FSL) manual (http://fmrib.ox.ac.uk/fslcourse/lectures/melodic.pdf)

All image registrations were conducted with FSL-FLIRT (Jenkinson et al., 2002). First, individual data was registered to individual’s high-resolution T1-weighted structural image which generated a transformation matrix. High-resolution T1-weighted structural image was then registered to a standard Montreal Neurological Institute (MNI-152) brain which generated a second transformation matrix. The two transformation matrices were used to register each individual’s preprocessed and denoised fcMRI data to MNI standard space for group analysis.

2.5. Region of interest selection and seed generation

We previously reported significant differences between long-term abstinent alcoholics and non-substance abusing controls in resting state networks generated with seeds in nucleus accumbens (NAcc) and subgenual anterior cingulate (sgACC; Camchong et al., 2013a). NAcc was selected because of its key role in processing the rewarding effects of alcohol or drugs (Everitt and Robbins, 2005; Koob and Le Moal, 1997). Repeated exposure to alcohol or drugs have shown to generate long-lasting synaptic changes in nucleus accumbens, re-organizing its connections within the appetitive drive network (Lee and Dong, 2011). SgACC was selected because of its key role in exerting control on emotion (Kelly et al., 2009), particularly in alcoholics (Salloum et al., 2007). Dysfunctional emotion regulation in alcoholics (e.g., extremes in emotional responsiveness to social situations, negative affect, and mood swings) has been associated with prefrontal dysfunction (Lyvers, 2000). The present study used these same regions, NAcc and sgACC, as seeds (3.5 mm radius) to generate the appetitive drive and executive control networks respectively (Camchong et al., 2013a).

2.6. Resting state individual-level analysis

The average time-series was extracted for each seed (sgACC and NAcc) for each participant. A multiple regression analysis on the denoised data was performed between the extracted average time-series from the seed and all voxels in the brain. This generated a map with a correlation coefficient r for each voxel, for each individual, for each seed. Correlation coefficients (r) were transformed to standardized z values. Resulting standardized z maps showed the degree of correlations with the corresponding seed averaged time-series for each seed for each participant.

2.7. Resting state group-level analysis

To investigate brain functional organization in LTAAS we conducted three sets of analyses. First, to examine whether LTAAS show similar adaptive mechanisms previously found in LTAA, we looked for group differences within specific clusters in which we previously found RSS differences between LTAA and NSAC (Table 4, Figs. 1 and 2). We conducted an analysis of variance (ANOVA) on mean z-scores extracted for each individual within the clusters in Table 4 as dependent variables and group membership as the fixed factor. Analysis included post hoc Tukey tests to look for specific RSS differences between groups within these clusters. Second, we conducted an independent samples t-test using a whole-brain analysis comparing RSS between LTAAS and NSAC for each seed. Third, to look for specific differences related to stimulant dependence, we conducted an independent samples t-test using a whole-brain analysis comparing RSS between LTAAS and LTAA for each seed. A threshold/cluster method derived from Monte Carlo simulations (AlphaSim, AFNI) was applied to control for false positive findings. Monte Carlo simulations (1000 iterations) accounted for the full-width half-maximum Gaussian filter with a connectivity radius of 7.1 mm. On the basis of these simulations, the family-wise ± of 0.025 was preserved with an a priori voxel-wise probability of 0.005 and three-dimensional clusters with a minimum volume of 1536 ± μL (192 voxels).

Table 4.

Regions in which LTAA have previously shown RSS differences when compared to NSAC (Camchong et al., 2013a) on which current group analysis was based.

Anatomy of region	L/R	x	y	Z	LTAAS vs NSAC p-value	LTAAS vs LTAA p-value
(A) NAcc
Dorsolateral prefrontal cortex, BA 10	L	−27	61	26	0.21	0.55
Caudate	L	−12	−1	15	0.27	0.08
	R	11	3	11
Anterior nucleus of thalamus	L	−10	−14	16	0.67	0.02*
	R	10	13	14
Medial dorsal thalamus	L	−6	−25	5	0.41	0.18
	R	8	−21	7
(B) Subgenual ACC
Dorsolateral prefrontal cortex, BA 8 and BA 46	R	26	28	54	0.02*	0.82
	R	52	34	20
Anterior nucleus of the thalamus	L	−12	2	12	0.99	0.01*
	R	6	−2	10

Italics, regions in which LTAA had higher RSS than NSAC in previous paper (Camchong et al., 2013a). LTAA, long-term abstinent alcoholics; LTAAS, long-term abstinent from alcohol and stimulants; RSS, resting state synchrony; NSAC, non-substance abusing controls; NAcc, nucleus accumbens; ACC, anterior cingulate cortex; BA, Brodmann area; L, left; R, right.

^*p < 0.05.

Fig. 1.

Three-dimensional MNI brain in neurological orientation showing enhanced executive control RSS in LTAAS within regions in which LTAA previously showed significantly higher strength of RSS than NSAC between sgACC and right DLPFC (Camchong et al., 2013a). Bar graph shows mean RSS strength for each group (error bars are one standard error). Red lines in bar graphs represent significant differences between groups (p < 0.05). MNI, Montreal Neurological Institute; LTAA, long-term abstinent alcoholics; LTAAS, long-term abstinent alcoholics with comorbid stimulant dependence; NSAC, non-substance abusing controls; RSS, resting-state synchrony; NAcc, nucleus accumbens; sgACC, subgenual anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex.

Fig. 2.

Three-dimensional MNI brain in neurological orientation showing lack of attenuation of RSS in appetitive drive regions in LTAAS within regions in which LTAA previously showed significantly lower strength of RSS than NSAC with the (A) NAcc and (B) sgACC seeds (Camchong et al., 2013a). Bar graphs show mean RSS strength for each cluster for each group (error bars are one standard error). Red lines in bar graphs represent significant differences between groups (p < 0.05). MNI, Montreal Neurological Institute; LTAA, long-term abstinent alcoholics; LTAAS, long-term abstinent alcoholics with comorbid stimulant dependence; NSAC, non-substance abusing controls; RSS, resting-state synchrony; NAcc, nucleus accumbens; sgACC, subgenual anterior cingulate cortex; MD-Thal, medial dorsal thalamus; AN-Thal, anterior nucleus of the thalamus.

2.8. Correlates of RSS

To examine how RSS differences found in LTAAS are related to behavior, we conducted correlations between RSS strength and behavioral measures in the LTAAS group. To minimize the number of correlations, we used the average RSS for (1) clusters within the executive control network ((Kerns et al., 2004; MacDonald et al., 2000); sgACC and DLPFC), and (2) clusters within the appetitive drive processing network (caudate, thalamus, nucleus accumbens; Everitt and Robbins, 2005). The following correlations were examined. First, based on previous findings of a significant positive correlation between executive control RSS and IED performance in the LTAA group (Camchong et al., 2013a), we investigated this same relationship in the LTAAS group. Second, because we recently found that higher current antisocial symptom counts in short-term abstinent alcoholics was significantly positively correlated with the appetitive drive RSS (Camchong et al., 2013b), we examined this same correlation in the LTAAS group. Third, we examined correlations between RSS strength and length of abstinence.

3. Results

3.1. Behavior results

LTAAS had significantly greater antisocial disposition and current ASPD symptoms than NSAC. LTAA had significantly greater antisocial disposition than NSAC, but had comparable current ASPD symptoms to NSAC. LTAAS had more lifetime ASPD symptoms than LTAA, but LTAAS and LTAA did not differ in current ASPD symptoms, or in measures of antisocial disposition (Table 5). There were no group differences on IED performance (Table 6).

Table 5.

Antisocial personality disorder (ASPD) measures previously reported in Fein and Fein 2012 in STAA and LTAA groups.

	LTAAS	LTAA	NSAC	LTAAS vs NSAC p-value	LTAAS vs LTAA p-value
Lifetime ASPD symptom count	11.67 ± 4.28	7.35 ± 4.53	2.74 ± 3.05	5.11E-9***	0.0004***
Current ASPD symptom count	2.00 ± 1.82	1.09 ± 1.73	0.61 ± 0.94	0.005**	0.088
California Psychological Inventory (socialization scale)	16.43 ± 3.26	17.43 ± 4.08	22.04 ± 3.72	3.94E–7***	0.548
Minnesota multiphasic personality inventory (psychopathic deviance scale)	25.53 ± 5.77	22.83 ± 5.80	16.83 ± 4.48	1.88E–7***	0.158

LTAAS, long-term abstinent from alcohol and stimulants; LTAA, long-term abstinent alcoholics; NSAC, non-substance abusing controls.

^**p < 0.01.

^***p < 0.001.

Table 6.

Analysis of variance results showed no overall group differences in Intradimensional/Extradimensional Set Shift task performance.

Behavioral measure	NSAC (n = 23)		LTAAS (n = 36)		LTAA (n = 23)		Overall ANOVA
	Mean	SD	Mean	SD	Mean	SD	F-value
Total number of blocks completed successfully	8.48	0.90	8.52	0.84	8.17	1.27	0.96
Total number of trials completed on all attempted blocks	92.39	21.62	90.22	20.15	91.43	18.16	0.09
Adjusted total number of errors	29.70	19.92	27.94	19.11	34.96	28.08	0.72

LTAAS, long-term abstinent from alcohol and stimulants; LTAA, long-term abstinent alcoholics; NSAC, non-substance abusing controls; SD, standard deviation.

3.2. Analysis within clusters previously found to be different between LTAA and NSAC

Enhanced executive control RSS in LTAAS

LTAAS showed significantly higher RSS between sgACC and right dorsolateral prefrontal cortex (DLPFC; Fig. 1) than NSAC. There were no differences in executive control RSS between LTAAS and LTAA.

No significant attenuation in appetitive drive network RSS in LTAAS

In contrast to our finding in LTAA (Camchong et al., 2013a), LTAAS did not show significant attenuation of RSS between appetitive drive regions in Table 4 (Fig. 2). LTAAS showed significantly higher RSS between anterior nucleus of the thalamus (AN-Thal) and both seeds than LTAA.

3.3. LTAAS vs LTAA

Independent sample t-test results revealed higher RSS between the NAcc seed and: right mid-posterior insula (Fig. 3A), right thalamus, and right caudate (Table 7A) in LTAAS than LTAA. LTAAS also showed higher RSS between the sgACC seed and: left anterior insula, bilateral caudate, right anterior nucleus of the thalamus, right putamen, and left DLPFC (Table 7B).

Fig. 3.

Three-dimensional MNI brain in neurological orientation showing regions in which LTAAS showed significantly higher RSS strength between insula and NAcc than (A) LTAA and (B) NSAC. Bar graphs show mean RSS strength for each group (error bars are one standard error). Red lines in bar graphs represent significant differences between groups (p < 0.05). MNI, Montreal Neurological Institute; LTAA, long-term abstinent alcoholics; LTAAS, long-term abstinent alcoholics with comorbid stimulant dependence; NSAC, non-substance abusing controls; RSS, resting-state synchrony; NAcc, nucleus accumbens.

Table 7.

RSS differences specific to LTAAS when compared to (A) NSAC and (B) LTAA.

(A) LTAAS > LTAA
RSS with NAcc					LTAAS	LTAAS
	L/R	x	y	z	vs NSAC	vs LTAA
					p-value	p-value
Mid-posterior insula	R	42	−4	8	0.036*	0.001**
Thalamus	R	14	−17	11	0.642	0.006**
Caudate	R	16	3	13	0.518	0.001**
RSS with sgACC					LTAAS	LTAAS
	L/R	x	y	z	vs NSAC	vs LTAA
					p-value	p-value
Dorsolateral prefrontal cortex	L	−44	16	27	0.021*	0.001**
Caudate	L	−10	7	12 1	0.069	0.00004***
	R	15	19		0.334	0.0004***
Putamen	R	18	3	11	0.077	0.0001***
Anterior and mid-insula	L	−33	−5	18	0.151	0.0002***
(B) LTAAS > NSAC
RSS with NAcc					LTAAS	LTAAS
	L/R	x	y	z	vs NSAC	vs LTAA
					p-value	p-value
Cingulate	Middle	0	13	26	0.003**	0.099
Mid-posterior insula	L	−48	10	6	0.005**	0.310
	R	46	−1	12	0.004**	0.022*
RSS with sgACC					LTAAS	LTAAS
	L/R	x	y	z	vs NSAC	vs LTAA
					p-value	p-value
Dorsolateral prefrontal cortex	R	49	32	12	0.0002***	0.006**
Putamen	L	−21	18	−6	0.0004***	0.046
	R	25	−6	11	0.0054**	0.002**
Rostral ACC	L	−4	38	23	0.0019**	0.004**

LTAA, long-term abstinent alcoholics; LTAAS, long-term abstinent from alcohol and stimulants; RSS, resting state synchrony; NSAC, non-substance abusing controls; NAcc, nucleus accumbens; ACC, anterior cingulate cortex; L, left; R, right.

^*p < 0.05.

^**p < 0.01.

^***p < 0.001.

3.4. LTAAS vs NSAC

Independent sample t-test results revealed higher RSS between the NAcc seed and (1) bilateral mid-posterior insula (Fig. 3B) and (2) cingulate (Table 7B) in LTAAS than NSAC. LTAAS showed higher RSS between the sgACC seed and (1) bilateral putamen, (2) right DLPFC and (3) perigenual/rostral ACC (Table 7B) than NSAC.

3.5. RSS behavioral correlates in the LTAAS group

3.5.1. Is RSS correlated with task performance?

Contrary to our hypothesis, LTAAS did not show a significant correlation between executive control RSS and IED performance as previously found in LTAA (Camchong et al., 2013a; Bonferroni correction required p-value <0.0125 = 0.05/4; one RSS measure plus three IED measures Table 6). Because insula has also been implicated in executive control (Schmaal et al., 2013), we conducted an exploratory analysis to explore the correlation between insula RSS and IED performance. We found a positive correlation between the total number of trials needed to complete an IED block (more trials less cognitive flexibility) and strength of insular RSS in the LTAAS group (r = 0.39, p = 0.020; not corrected for multiple comparisons). Linear regression analysis revealed that insula RSS predicted IED performance in the LTAAS group (Beta = 0.386, significance 0.020). Insula RSS, however, did not predict IED performance in the LTAA or the NSAC group. An analysis of covariance with group membership as fixed factor, IED performance as dependence variable, and insular RSS as covariate, revealed that there was a significant group by RSS interaction (F = 3.95, p = 0.023), in which insular RSS significantly predicts IED performance in the LTAAS group only, but not the LTAA or NSAC groups. Moreover, a test looking for differences in correlation coefficients, based on Fisher’s Z-transform (Paul, 1989), revealed significant differences among correlation coefficients between groups (χ² = 7.53886, p = 0.023; Fig. 4A).

Fig. 4.

Scatter plots showing positive correlations between strength of insula RSS and (A) lack of cognitive flexibility (i.e. more trials needed to complete a block) and (B) current antisocial personality disorder symptom counts. Each dot represents an individual LTAAS (solid black), LTAA (solid gray), or NSAC (green outline) subject. LTAAS, long-term abstinent alcoholics with comorbid stimulant dependence; LTAA, long-term abstinent alcoholics; NSAC, non-substance abusing controls; RSS, resting-state synchrony.

3.5.2. Is RSS strength correlated with ASPD?

Contrary to our hypothesis, LTAAS did not show significant correlations between ASPD current symptom count and appetitive drive RSS. Exploratory analysis of insular RSS correlates, however, showed a significant positive correlation between the current ASPD symptom count and insular RSS (r = 0.43, p = 0.010, not corrected for multiple comparisons; Fig. 4B).

3.5.3. Higher insular RSS correlated with length of abstinence

Bonferroni correction required a p-value < 0.0056, when correlating average RSS (executive, appetitive and insular) with length of abstinence (alcohol, methamphetamine or cocaine) (p-value = 0.05/9). LTAAS with history of cocaine dependence (n = 28) showed a significant negative correlation between cocaine abstinence duration and insular RSS (r = −0.568, p = 0.002).

3.5.4. Exploratory analysis examining family density associated with RSS

To investigate whether lack of attenuation of appetitive drive RSS is a function of stimulant use or not, we correlated appetitive drive RSS and the proportion of first degree relatives with drug problems in LTAAS. LTAAS showed a significant positive correlation between these two variables (r = 0.35, p = 0.044).

4. Discussion

We previously observed enhanced resting state synchrony (RSS) within the executive control network and attenuated RSS within the appetitive drive network in long-term abstinent alcoholics (LTAA) without comorbid drug dependence vs normal controls (NSAC; Camchong et al., 2013a). We found similar effects, although to a lesser degree, in alcoholics with short-term (~10 weeks abstinent) abstinence (Camchong et al., 2013b). We believe these differences are adaptive mechanisms that support abstinence both because of the graded effect seen in short-term vs long-term abstinence and because these networks play an important role in behavioral aspects needed to overcome addiction such as cognitive inhibitory control and regulation of emotion and appetitive behavior (Hare et al., 2009; Medalla and Barbas, 2009; Naqvi and Bechara, 2010). The present study sought to investigate whether these differences (vs NSAC) were also present in long-term abstinent alcoholics with comorbid stimulant dependence (LTAAS). Additionally, due to its critical role in addiction (Naqvi and Bechara, 2010; Schmaal et al., 2013), we investigated insula RSS differences. Results showed commonalities in LTAA and LTAAS RSS (similar enhanced executive control RSS and left insula RSS) as well as differences (no attenuation of appetitive drive RSS in LTAAS and no enhancement of RSS in right insula in LTAA). These findings suggest common as well as specific targets for treatment in chronic alcoholics with vs without comorbid stimulant dependence.

4.1. Enhancement of RSS in executive control network in LTAAS and LTAA

LTAAS and LTAA show similar higher RSS compared to NSAC in the executive control network, especially between sgACC and right DLPFC (Fig. 1). SgACC activity has been associated with emotional responsiveness and regulation of reward-related behavior (Kelly et al., 2009; Phan et al., 2005), while right DLPFC activity has been associated with executive control evaluation and integration of behavior (Delgado et al., 2008; Hare et al., 2009; McClure et al., 2004). Researchers have proposed that the level of synchronization between ACC and DLPFC increases or decreases in relation to cognitive demand (Hare et al., 2009; Medalla and Barbas, 2009). Since both alcoholic groups showed enhanced RSS between these regions when compared to controls, this difference may reflect the alcoholics’ need to control emotion and behavior to achieve long-term abstinence.

4.2. Lack of attenuation of RSS in appetitive drive network in LTAAS

Chronic stimulant use, which has been directly associated with the reorganization of synaptic connectivity patterns in appetitive drive regions (Jedynak et al., 2007), may have impeded the adaptive attenuation of RSS between regions known to process reward. It should be noted that lack of attenuation of RSS between appetitive drive regions does not lead to relapse. Both, the LTAA and LTAAS groups had multi-year abstinence from all drugs of abuse at the time of assessment, irrespective of the strength of RSS in the appetitive drive network. Moreover, strength of appetitive drive RSS was not related to length of abstinence in LTAA or LTAAS.

Our finding of a positive association between appetitive drive RSS and family history of drug problems in LTAAS suggests that this may be a potential biological marker of vulnerability to stimulant addiction. Elevated appetitive drive network RSS may exist before active stimulant addiction, reflecting the genetic vulnerability to stimulant dependence (Stewart et al., 2012). This finding warrants further research in adolescents with high family density of addiction to disentangle the effects of environment vs genetics.

4.3. Laterality effects of RSS in insula

Enhanced insula RSS in LTAAS (Fig. 3) provides evidence of the important role of the insula in alcohol and stimulant addiction (Goldstein et al., 2009; Naqvi and Bechara, 2010). Research has proposed a wide range of behavioral aspects of addiction in which insula plays a role, such as interoceptive awareness, decision-making, and emotion regulation. RSS correlates revealed that LTAAS with higher insula RSS had difficulties in adapting to changing contingencies during the IED task. These findings imply two potential interpretations for LTAAS: (1) enhanced insula RSS may disrupt the ability to adapt to new contingencies or (2) LTAAS with greater difficulty adapting to new contingencies compensate by enhancing insula RSS. LTAAS performing at levels comparable to LTAA and NSAC supports the latter interpretation. LTAAS with higher demands to internally modulate decision-making and emotion may need to keep the insula network engaged even during rest to be prepared to respond properly to the environment.

As shown in the results section, LTAAS had increased RSS in bilateral insula when compared to NSAC (Fig. 3B, Table 7B), while this effect was only present in right insula when compared to LTAA (Fig. 3A, Table 7A) (There were no differences in left insula RSS between LTAAS and LTAA). Although our previous study using whole-brain analysis (Camchong et al., 2013a) did not find differences in insula RSS between LTAA vs NSAC, current findings of increased RSS in LTAAS vs NSAC prompted us to further explore whether increased RSS in insula is a compensatory mechanism specific to individuals with comorbid alcohol and stimulant dependence or whether this effect can be found to a lesser extent in individuals with alcohol dependence only (LTAA). We conducted a focused analysis honing into insula’s RSS applying a bilateral mask of the insula (Lancaster et al., 2000) to the analysis to increase detection power and decrease the number of multiple comparisons. Results revealed that both, LTAA as well as LTAAS have significantly increased RSS between NACC and left insula when compared to NSAC (LTAA vs NSAC: t(44) = 2.12, p = 0.039; Partial Eta² = 0.093; LTAAS vs NSAC: t(57) = 4.13, p = 0.00016; Partial Eta² = 0.240). LTAAS and LTAA did not differ in left insula RSS (t(57) = 1.99, p = 0.163, Partial Eta² = 0.034). The strength of activity in left insula has been previously associated with regulation of emotion and appetitive behavior (Kelly et al., 2012; Paulus et al., 2012).

The strength of activity in right insula, a region in which LTAAS showed increased RSS vs both NSAC and LTAA (Fig. 3, Table 7), may support arousal and response to aversive emotion such as disgust and anxiety (Goldstein et al., 2009; Kelly et al., 2012), and has shown positive associations with the ability to remain abstinent in individuals with stimulant dependence only (Clark et al., 2012; Paulus et al., 2005). There is evidence for suggesting that individuals that perform well under extremely stressful conditions show increased right insula activity. Navy SEALS, for example, have shown right insula hyperactivity (vs controls) when shifting affect during stress induction, presumably to modulate emotional and interoceptive awareness and effectively respond to the environment (Simmons et al., 2012). Increased activity in right insula has also been reported in abstinent cocaine addicts when exposed to personalized stress scripts (Potenza et al., 2012). We propose that, like Navy SEALS, LTAAS have stress-coping mechanisms that allow them to deal with the stress in their lives without relapsing (Ando et al., 2012; Kiluk et al., 2010). These interpretations may be further explored in studies that evaluate insula network function and organization before and after stress induction.

Present findings suggest a laterality effect in insula RSS with: (1) increased RSS in left insula present in both alcoholic groups, with a larger effect in LTAAS, a compensatory mechanism that is presumably associated with the ongoing need to enhance emotion regulation, and (2) increased RSS in right insula only in LTAAS, possibly due to an ongoing demand to regulate stress to remain abstinent.

4.4. Summary

Current results provide support for three premises. First, enhanced synchrony of executive control regions may be a common adaptive mechanism that supports abstinence among both alcoholic groups. Second, attenuation of RSS within the appetitive drive network is not essential to achieve long-term abstinence in LTAAS, although it may potentially be a factor reflecting stimulant dependence vulnerability. Third, enhanced RSS in left insula may constitute an adaptive difference related to the need to modulate emotion in both LTAAS and LTAA, while enhanced RSS in right insula may facilitate stress-coping mechanisms in LTAAS only, adaptations that presumably facilitate abstinence maintenance. These results suggest common and specific potential treatment targets (to help facilitate achievement and maintenance of long-term abstinence) in the brain’s resting functional organization in chronic alcoholics with and without comorbid stimulant dependence.