Resting-state synchrony during early alcohol abstinence can predict subsequent relapse

Abstract

Short-term abstinent alcoholics have shown increased engagement of reward regions and reduced engagement of executive control regions. There is no report yet on whether these differences can predict relapse. This is the first study that investigates whether differences in resting-state networks can predict later relapse. Resting-state functional magnetic resonance imaging data were collected from 69 short-term abstinent alcoholics. Participants performed the affective go/no-go task outside of the scanner. At 6-month follow-up, participants were grouped as abstainers (N = 40; age: M = 46.70, standard deviation [SD] = 6.83) and relapsers (N = 29; age: M = 46.91, SD = 7.25). We examined baseline resting-state synchrony (RSS) using seed-based measures. Compared with abstainers, relapsers showed significantly decreased RSS within both the reward and executive control networks as well as within the visual network (P < 0.05). Lower RSS in relapsers could predict relapse (P < 0.05) and was significantly correlated with poor inhibitory control of emotional-laden stimuli (P < 0.017) and with alcohol use (P < 0.05). Results suggest that lower RSS during short-term abstinence may predict subsequent relapse. The association of lower RSS with poorer inhibitory control suggests that low RSS may constitute a faulty foundation for future responses to external cues, which can be manifested as the inability to inhibit behavior.

Keywords: abstinence; alcohol; fMRI; relapse; resting-state networks.

Figures

Figure 1.

t-Maps showing within group t-test results in abstainers (top row) and relapsers (bottom row) in the NAcc (reward), sgACC (executive control), middle occipital gyrus (MOG) (visual), and insular networks.

Figure 2.

(A) Left: NAcc seeds in green. Right: Resulting t-map showing regions in which abstainers had significantly higher strength of RSS than relapsers within the nucleus accumbens network. t-Map is overlaid on a MNI brain in radiological orientation (right is left). (B) Bar graphs (with ±1 standard error bars) showing strength of RSS between subgenual ACC and regions with significant differences between abstainers (blue bars) and relapsers (green bars). ANCOVA analysis showed that differences in all regions showed were still significant after controlling for noise (see F and Sig. values in Table 6A). RSS, resting-state synchrony; NAcc, nucleus accumbens; LN, lentiform nucleus; ACC, anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex.

Figure 3.

(A) Left: sgACC seed in green. Right: Resulting t-map showing regions in which abstainers had significantly higher strength of RSS than relapsers within the sgACC network. t-Map is overlaid on a MNI brain in radiological orientation (right is left). (B) Bar graphs (with ±1 standard error bars) showing strength of RSS between sgACC and regions with significant differences between abstainers (blue bars) and relapsers (green bars). ANCOVA analysis showed that differences in all regions showed were still significant after controlling for noise (see F and Sig. values in Table 6B). RSS, resting-state synchrony; sgACC, subgenual anterior cingulate cortex; DLPFC, dorsolateral prefrontal cortex; LN, lentiform nucleus; MOG, middle occipital gyrus.

Figure 4.

(A) Left: MOG seed in green. Right: Resulting t-map showing regions in which abstainers had significantly higher strength of RSS than relapsers within the MOG network. t-Map is overlaid on a MNI brain in radiological orientation (right is left). (B) Bar graphs (with ±1 standard error bars) showing strength of RSS between MOG and regions with significant differences between abstainers (blue bars) and relapsers (green bars). ANCOVA analysis showed that differences in all regions showed were still significant after controlling for noise (see F and Sig. values in Table 7A). RSS, resting-state synchrony; MOG, middle occipital gyrus; STG, superior temporal gyrus; IPL, inferior parietal lobule.

Figure 5.

(A) Left: Insula seed in green. Right: Resulting t-map showing regions in which abstainers had significantly higher strength of RSS than relapsers within the insular network. t-Map is overlaid on a MNI brain in radiological orientation (right is left). (B) Bar graphs (with ±1 standard error bars) showing strength of RSS between insula and regions with significant differences between abstainers (blue bars) and relapsers (green bars). ANCOVA analysis showed that differences did not reach significance after controlling for noise (see F and Sig. values in Table 7B). RSS, resting-state synchrony; PHG, parahippocampal gyrus.

Figure 6.

Scatterplot showing correlations between generalized RSS strength and commission errors (response to a distractor stimulus) during positive (green), negative (red), and neutral (gray) target blocks in relapsers. Each dot is an individual subject. RSS, resting-state synchrony; AGN, affective go/no-go task.

Figure 7.

Bar graphs showing mean RSS for NSAC (brown), abstainers (blue), and relapsers (green) in specific clusters within the (A) NAcc, (B) sgACC, (C) visual, and (D) insular networks shown in Tables 6 and 7. ANOVA results showed that relapsers were significantly different from both abstainers and NSAC. Abstainers were only significantly different from relapsers. NSAC, non-substance abuse controls.