مقررات رفتاری در معتادان متامفتامین: مطالعه fMRI

Abstract The goal of this study was to extend our previous findings of abnormal prefrontal function in methamphetamine (MA) abusers and controls and to link the imaging data to behavioral, demographic and drug use variables. We used a fast event-related functional magnetic resonance imaging (fMRI) design to examine trial-to-trial reaction time (RT) adjustments in 30 MA abusers and 30 controls. A variant of the Stroop task was employed to measure influence of response conflict on RT, including the level of trial-to-trial RT adjustments seen after conflict trials. Compared to control subjects, MA abusers exhibited reduced RT adjustments and reduced activation in the prefrontal cortex (PFC) after conflict trials. RT adjustment correlated negatively with PFC brain activity in the MA group, while a trend for a positive correlation was observed in controls. No correlations were observed between task performance or brain activity and age, education or drug use variables. These data support our previous findings that the ability to adapt a behavioral response based on prior experience is compromised in MA abusers. Interestingly, these impairments do not appear to be linked to drug use patterns or to educational levels.

Introduction Worldwide use of methamphetamine (MA) is now estimated to be at 51 million users (Degenhardt et al., 2008, Roehr, 2005, United, 2008 and United, 2009), with global abuse of amphetamine/methamphetamine now surpassing that of cocaine and opiates combined (United Nations, 2009). Approximately 5% of the adult population in the United States has used MA on at least one occasion and Emergency Department admissions related to MA use have doubled during the period of 1994 to 2002 (SAMSHA, 2004). Numerous imaging studies suggest that fronto-cingulate regions of the brain are affected by MA abuse (Ernst et al., 2000, London et al., 2004, Nordahl et al., 2005, Salo et al., 2007 and Volkow et al., 2001). Consistent with abnormalities in brain structure and function, cognitive impairments have also been observed in MA abusers on tasks that require the suppression of task irrelevant information (Monterosso et al., 2005 and Salo et al., 2007), decision-making (Kalechstein et al., 2003 and Paulus et al., 2003), working memory (McKetin and Mattick, 1998) and cognitive control (Nestor et al., 2011 and Salo et al., 2009a). Neurobiological models of addiction propose that ventral brain regions (i.e., orbital frontal cortex and nucleus accumbens) contribute to the impulse to seek drugs, whereas the recruitment of fronto-cingulate regions may be critical to control those prepotent impulses (i.e., cognitive control) (Jentsch and Taylor, 1999). In the context of addiction, cognitive control can be interpreted as the inhibition of a prepotent response (e.g., habitual drug use) in order to carry out behaviors associated with long-term rewards and positive outcomes (e.g., abstaining from drug use). 1.1. Study rationale The goal of the current study was to replicate and extend our preliminary findings that examined cortical mechanisms of cognitive control in a small group of MA abusers (Salo et al., 2009b). In the current study we increased our sample significantly in order to examine correlations between brain activity and behavior. In order to examine the neural substrates of cognitive control relevant to addiction, we conducted a fast event-related functional magnetic resonance imaging (fMRI) study in which we examined trial-to-trial reaction time (RT) adjustments using a variant of the single-trial Stroop task in 30 chronic MA abusers and 30 controls. This version of the Stroop task (Kerns et al., 2004) creates conditions in which performance (RT and accuracy) reflects the ability to recognize and resolve conflict at the time of response selection (i.e., within a trial), as well as the ability to flexibly adapt behavior, such as using exposure to conflict situations to change subsequent behavior (i.e., trial-to-trial adjustments after conflict trials). Many lines of research have shown that the information-processing system can adjust its strategy on a trial-to-trial basis and that the utilization of such a plan of action may be coupled with the subjective view that a particular strategy will produce a positive outcome (Gratton et al., 1992). Given that one of the hallmark behaviors associated with addiction is the failure to choose adaptive strategies to achieve future positive outcomes, this paradigm is well suited to examine the role of cognitive control in MA dependence. We hypothesized that in this increased sample of MA abusers we would observe correlations between abnormal patterns of trial-to-trial RT adjustments (i.e., conflict adaptation) and reduced PFC activity. Given the recent claim that many cognitive studies in MA abuse are confounded by group differences in age and education, we also wanted to test these relationships directly.

3. Results 3.1. Behavioral data 3.1.1. Reaction time analyses Analyses revealed main effects of Stroop word type [F(1,58)=157.49, p<0.0001] and trial-to-trial adjustments [F(1,58)=14.58, p<0.0001]. Planned analyses also revealed that the trial-to-trial adjustment RT effect (cI–iI) differed significantly between the MA abusers and controls [F(1,58)=4.37; p<0.05]. While the controls showed an RT advantage (26-ms benefit) to conflict trials that were preceded by conflict trials (iI), the MA abusers showed no advantage and were actually slower (4-ms cost). These group differences in Stroop performance endured with age, education, NART scores as covariates. Correlational analyses revealed no relationship between education levels and trial-to-trial adjustments in the MA abusers (r=.11; p=0.57) or the controls (r=.15; p=0.44). No correlations were observed between measures of age and premorbid IQ (i.e., NART scores) and trial-to-trial adjustments in either group. Similar to our previously published findings, no group differences were observed on within-trial Stroop conflict effects (F<1) ( Table 2). Table 2. Behavioral results from 30 methamphetamine (MA) abusers and 30 control subjects. Methamphetamine abusers (n=30) Control subjects (n=30) Within-trial Stroop effects, median (S.D.) (ms) Incongruent 767.6 (140.7) 752.5 (156.9) Congruent 629.4 (80.1) 626.7 (118.8) Stroop conflict effect 138.2 125.8 Conflict errors 0.05 (.04) 0.07 (.07) Non-conflict errors 0.02 (.02) 0.03 (.05) Trial-to-trial RT adjustments median (S.D.) (ms) Congruent-incongruent (cI) 765.5 (140.2) 765.4 (167.2) Incongruent-incongruent (iI) 769.7 (147.1) 739.7 (150.2) Trial-to-trial adjustment (cI–iI) −4.2a 25.8 a Significantly different from control group. Table options 3.1.2. Error analyses Analyses revealed a main effect of word type [F(1,58)=36.13, p=0.0001] with both groups making significantly more errors in the incongruent condition (6%) than in the congruent condition (3%). There were no group differences in incongruent [F(1,58)=2.28; p=0.13] or congruent errors [F(1,58)=1.5; p=0.22]. There was no evidence of a speed-accuracy trade-off for both groups (MA abusers; r=0.293; p=0.10, controls: r=0.096; p=0.61]. 3.2. Imaging results We first examined whether activity within the ACC and PFC was associated with within-trial conflict monitoring (I–C) and trial-to-trial adjustments (iI–cI). Using ANOVA procedures with subject as a random variable, significant activation to conflict contrasts (I–C) was observed in the ACC in both groups with no differences observed between groups. In contrast, when we examined the pattern of activation for the trial-to-trial RT adjustments, we found that controls exhibited increased PFC activity on iI sequences compared to the cI sequences. In contrast the MA abusers showed little or no activation within the PFC region (see Fig. 1). We then ran a second-level random-effects analysis to explore group differences across the contrasts of interest. Between-group differences emerged in activation in the prefrontal regions centered in right and left Brodmann areas 6 and 8 (p=0.001, corrected) ( Friston et al., 1996). (See Fig. 2.). These regions showed lower levels of activation in the MA abusers after iI trial sequences relative to cI trial sequences. Table 3 summarizes the data for frontal regions with significantly different activation between the MA abusers and controls. Regions of increased activation associated with trial-to-trial adjustments ... Fig. 1. Regions of increased activation associated with trial-to-trial adjustments (iI–cI) in methamphetamine abusers and controls. Figure options Dorsolateral prefrontal cortex (DLPFC) region showing increased activity related ... Fig. 2. Dorsolateral prefrontal cortex (DLPFC) region showing increased activity related to trial-to-trial adjustments (iI–cI) in control subjects relative to methamphetamine (MA) abusers. (2a) Graph of the average coefficients for the iI–cI statistical contrast in the two groups, averaged within group for the voxels of the DLPFC activation shown in panel 2a. Figure options Table 3. Brain regions with significant group differences associated with trial-to-trial adjustments in 30 methamphetamine (MA) abusers and 30 control subjects. Region Brodmann’s area No. voxels MNIcoordinates x y z Right frontal gyrus 8 28 32 16 46 Right superior frontal gyrus 6 21 26 12 52 Right mid-frontal gyrus 6 40 18 12 76 Right cingulate gyrus 32 14 8 16 44 Right cingulate gyrus 24 28 14 −24 46 Left frontal gyrus 8 69 −32 16 46 Left superior frontal gyrus 6 28 −14 −6 72 Left mid-frontal gyrus – 26 −36 44 −14 Left cingulate gyrus – 11 −14 −42 48 Table options 3.3. Correlations In this expanded dataset significant negative correlations were observed between the PFC beta coefficients and the corresponding behavioral data for the trial-to-trial adjustments in the MA abusers (r=−0.375, p<0.05). A trend significant positive correlation was also observed in the controls (r=−0.33, p=0.07). A direct statistical analysis of correlation coefficients between groups reached significance (p =0.003). These results suggest that while the MA group failed to activate the PFC in order to sustain adaptive trial-to-trial RT adjustments, the controls exhibited increased PFC activation in response to trial-to-trial RT adjustments. Give the physiological interaction between medial and lateral frontal cortices and striatum in mediating cognitive control, we wanted to examine the specificity of brain-behaviorrelationships within the PFC ( Frank et al., 2005 and Pasupathy and Miller, 2005). In order to do this, we conducted correlations within a striatal cluster that showed activation during trial-to-trial adjustments within both MA abusers and controls, but did not show differences between groups. In this correlational analysis, no significant correlations emerged in the MA abusers (r=0.15; p=0.44) or the controls (r=0.12; p=0.54).