مکانیزم پره فرونتال برای کنترل اجرایی بر روی حواس پرتی عاطفی در افسردگی اساسی تغییر داده شده

Abstract A dysfunction in the interaction between executive function and mood regulation has been proposed as the pathophysiology of depression. However, few studies have investigated the alteration in brain systems related to executive control over emotional distraction in depression. To address this issue, 19 patients with major depressive disorder (MDD) and 20 healthy controls were scanned using functional magnetic resonance imaging. Participants performed an emotional oddball task in which infrequently presented circle targets required detection while sad and neutral pictures were irrelevant novel distractors. Hemodynamic responses were compared for targets, sad distractors, and for targets that followed sad or neutral distractors (Target-after-Sad and Target-after-Neutral). Patients with MDD revealed attenuated activation overall to targets in executive brain regions. Behaviorally, MDD patients were slower in response to Target-after-Sad than Target-after-Neutra stimuli. Patients also revealed a reversed activation pattern from controls in response to this contrast in the left anterior cingulate, insula, right inferior frontal gyrus (IFG), and bilateral middle frontal gyrus. Those patients who engaged the right IFG more during Target-after-Neutral stimuli responded faster to targets, confirming a role of this region in coping with emotional distraction. The results provide direct evidence of an alteration in the neural systems that interplay cognition with mood in MDD.

1. Introduction Emotional distraction often interferes with cognitive processing (Johnson et al., 2005 and Dolcos and McCarthy, 2006). One of the cardinal features of major depressive disorder (MDD) is an inability to disengage from negative thoughts, memories and events in order to sustain attention towards on-going cognitive tasks (Lyubomirsky et al., 1998, Wenzlaff and Bates, 1998, Ellenbogen et al., 2002 and Siegle et al., 2002). In turn, susceptibility to emotional distraction adversely affects the capabilities of patient to cope with the demands of daily living (Ottowitz et al., 2002 and Rogers et al., 2004). Despite the established clinical importance of executive dysregulation of emotional processing in MDD, alterations in neural functioning associated specifically with this aspect of the disorder are not yet clear. An influential neurobiological model proposed for mood regulation posits a failure of coordination in dorsal and ventral brain systems subserving executive control and emotional processing, respectively (Mayberg, 1997). Mayberg and colleagues propose that the rostral anterior cingulate (ACC) and related areas in the inferior and medial prefrontal cortex (PFC) may serve critical roles in balancing the relative influence of these brain systems to guide goal-directed behavior and maintain healthy mood. In healthy populations, emotional Stroop and emotional Go/NoGo tasks have been employed to investigate inhibitory cognitive control over emotional distraction by presenting task-irrelevant emotional information simultaneously with task-relevant stimulus features. The ACC, particularly its rostral and ventral aspects, is consistently activated by emotional interference in these tasks (Whalen et al., 1998, Elliott et al., 2000, Bishop et al., 2004, Etkin et al., 2006 and Shafritz et al., 2006). This region has been associated with mediating conflict between competing responses (Carter et al., 1998 and Botvinick et al., 2001), monitoring for the occurrence of response conflict in information processing (Carter et al., 1998, Carter et al., 2000, Botvinick et al., 1999, Barch et al., 2000 and Botvinick et al., 2001), and error monitoring and detection (Rubia et al., 2003). In addition, a number of studies using the Stroop, Go/No Go and other attention-demanding tasks suggest a role of the right inferior frontal gyrus (IFG) in inhibitory processes relevant for successful cognitive performance and executive function (Jonides et al., 1998, Konishi et al., 1998, D'Esposito et al., 1999, Smith and Jonides, 1999, Liddle et al., 2001, Rubia et al., 2003 and Aron et al., 2004). Of particular relevance to the present study, Dolcos and McCarthy (Johnson et al., 2005 and Dolcos and McCarthy, 2006) revealed a role of the IFG in inhibiting emotional distraction in healthy adults. While subjects performed a working memory task, activation in the IFG was enhanced when the subject was distracted by negative emotional pictures relative to distracting neutral or scrambled pictures. Subjects with great activity to emotional distracters in the IFG tended to rate emotional distracters as less distracting, suggesting that activity in the IFG indexed successful inhibition of emotional distraction. It is unknown, however, whether recruitment of this region during emotional distraction is altered in clinical populations, such as MDD, and whether dysregulation of IFG activity has behavioral consequences on task performance. Functional neuroimaging of MDD patients during emotional tasks has implicated dysfunction in frontolimbic regions, providing initial support for Mayberg's neuroanatomical model of mood regulation. For instance, a sustained emotional response in the amygdala during a personal relevance rating task and decreased dorsolateral PFC activity on a digit-sorting task has been reported in MDD relative to controls (Siegle et al., 2002 and Siegle et al., 2006). Siegle and colleagues also reported a decreased correlation of amygdala and dorsolateral PFC activity in the MDD group (Siegle et al., 2006); however, this study did not address how emotional responses affected subsequent brain activity associated with executive control. George et al. (1997) reported decreased activity in the ACC in MDD patients who performed an emotional Stroop task while undergoing positron emission tomography (PET) scanning. Elliott et al. (2002) reported attenuated neural responses to emotional relative to neutral targets in the ventral ACC and the posterior orbitofrontal cortex during an emotional Go/No Go task in MDD. However, because the latter attentional studies used blocked designs, it is not possible to disambiguate responses to different stimulus events and time epochs during the task. In the present event-related functional magnetic resonance imaging (fMRI) study, we extended these initial neuroimaging findings in MDD to investigate alterations in executive and emotional processing systems during an attentionally demanding visual oddball task with intermittent emotional distraction by presentation of sad pictures (Wang et al., 2005 and Wang et al., 2006). This task more closely models the disruption of on-going task-relevant cognitive processes by sporadic mood-congruent thoughts in MDD than Stroop or Go/No Go tasks in which the emotional stimuli are themselves task-irrelevant. Furthermore, because the emotional distractors are temporally separated from presentation of the attentional targets, we can evaluate whether emotional dysregulation in MDD leads to performance decrements and differential brain activation to task-relevant attentional targets that follow emotional distractors close in time. Our previous studies using this paradigm in healthy adults have consistently shown that the attentional targets activate dorsal frontoparietal structures and the sad distractors activate ventral frontolimbic structures, including the amygdala (Wang et al., 2005 and Wang et al., 2006). In the present study, we first compared brain activation patterns in MDD and controls to attentional targets and sad distractors separately in order to address the main effect of depression on executive and emotional processing, respectively. Next, to probe the lingering impact of the sad distractors on executive function, we isolated activity to attentional target events that were preceded by the sad distractors (relative to those preceded by neutral distractors). We hypothesized that brain regions such as the ACC and the right IFG might be activated by this contrast and play critical roles in executive control requiring reallocation of attentional resources to task-relevant processing from task-irrelevant emotional distraction. We predicted that relative to the control group, the MDD group would have decreased activation to targets following sad distractors in these regions.

3. Results 3.1. Behavioral data: RT and emotional ratings MDD patients had slower RTs to attentional targets during the scan, t(37) = 2.69, P < 0.05, regardless of the valence of the previous distractor. The mean (S.D.) RT was 663.4 (78.8) ms for the MDD group and 590.0 (94.9) ms for the control group. A repeated measures ANOVA on target subtype (Targets-after-Sad, Target-after-Neutral) using group (MDD, control) as the between-subjects factor revealed a significant main effect of group, F(1,38) = 5.01, P < 0.031, with MDD patients showing slower responses. The main effect of target subtype was also significant, F(1,37) = 5.67, P = 0.023, with sad distractors showing slower responses. Although there was no significant interaction effect, the RT slowing was exaggerated in MDD patients when the targets were preceded by sad distractors ( Fig. 2a). The MDD group was significantly slower in response to the Target-after-Sad event than the control group, t(37) = 3.05, P = 0.004, and within the MDD group, RTs to Target-after-Sad events were significantly slower than RT to Target after-Neutral events, t(37) = 2.31, P = 0.027. a) The influence of emotional distraction on reaction time (RT) to the ... Fig. 2. a) The influence of emotional distraction on reaction time (RT) to the presentation of subsequent targets. The MDD group revealed significantly slower RT when a target was preceded by a sad distractor (Target-after-Sad) than when a target was preceded by a neutral distractor (Target-after-Neutral), ⁎P < 0.05. b) Comparison of the percentage of pictures out of the set of 100 that were rated as ‘mildly happy’, ‘neutral’, ‘mildly sad’, ‘sad’ and ‘very sad’ by controls and MDD subjects. Results indicate a bimodal distribution. The MDD group rated significantly more pictures as ‘very sad’ (⁎P < 0.05), indicating a negative emotional processing bias. Figure options Subjective ratings for the picture distractors were analyzed using Student's t tests. The MDD group rated more pictures as ‘very sad’ than controls, t(37) = 2.51, P = 0.018, suggesting a negative emotional processing bias ( Fig. 2b). 3.2. fMRI results Random-effect/within-group analysis revealed that the control group showed activation in response to attentional targets in the dorsal executive system, including bilateral middle frontal gyrus (MFG), IFG (BA45 and 47), dorsal ACC, posterior cingulate, supramarginal gyrus, superior parietal lobule, precuneus, thalamus and striatum (Fig. 3a). The control group showed activation in response to the sad distractors (relative to neutral distractors) in ventral posterior and frontolimbic regions, including bilateral extrastriate cortex, fusiform gyrus, amygdala, anterior insula, IFG (BA47), and middle and superior temporal sulci. These results are consistent with our previous findings in healthy adults using the same paradigm (Wang et al., 2005). The depressed group activated similar regions as the controls (Fig. 3a). For clarity, we only report below the results that show significant differences between the two groups rather than separate main effects within the two groups. a) Voxel-based random effect analysis for the MDD and control groups on the ... Fig. 3. a) Voxel-based random effect analysis for the MDD and control groups on the activation to attentional targets at the peak time point. Both of the groups activated a dorsal executive system, although the MDD group had a lesser extent of activation (false discovery rate-corrected P < 0.05, spatial extent of five contiguous voxels). b) Significantly decreased activation in the MDD group compared with the control group from a two-sample t test on activation to attentional targets at the peak time point (P < 0.001, uncorrected, spatial extent of five contiguous voxels). ACC = anterior cingulate cortex; BG = basal ganglia; IFG = inferior frontal gyrus; MFG = middle frontal gyrus; L = left hemisphere; R = right hemisphere. Figure options 3.2.1. Alterations in executive and emotional processing systems in MDD A direct statistical comparison between the MDD and control groups revealed activity reductions in the MDD group to attentional targets in the executive system (Fig. 3b, Table 2) including superior frontal gyrus, middle frontal gyrus, supramarginal gyrus, insula, basal ganglia, and bilateral IFG (left BA47 and right BA45). Activation patterns to the sad distractors relative to neutral ones were comparable between the groups. However, MDD patients showed greater deactivation than controls to sad stimuli relative to baseline in dorsal frontoparietal regions, including the middle frontal gyrus, supramarginal gyrus, precuneus, postcentral gyrus and temporal parietal conjunction area ( Table 3, Fig. 4). Note that most of these regions showed activation in response to attentional targets ( Fig. 3 and Fig. 4). The task-induced deactivation usually occurs during an active task relative to a “resting” or “passive” baseline in a default-mode network. It is postulated to result from a reallocation of processing resources ( McKiernan et al., 2003). The magnitude of deactivation increases with task difficulty ( McKiernan et al., 2003). The enhanced deactivation for emotional distractors in the regions associated with cognitive function may reflect reallocation of attentional resources to sad distractors ( Posner and Dehaene, 1994 and Drevets and Raichle, 1998). Thus the increased deactivation might suggest a strong emotional distraction effect in MDD. Table 2. Brain regions showing significant group differences (threshold P < 0.001, five contiguous voxels) in activation to attentional targets (MDD < control) Region BA area Hemisphere MNI coordinates T value Cluster size x y z Middle frontal gyrus BA10 L − 42 53 0 3.33 28 R 46 42 4 3.23 26 Inferior frontal gyrus BA47 L − 49 25 − 11 4.21 42 BA45 R 32 14 25 4.03 25 Precentral gyrus L − 28 4 39 4.63 132 Insula L − 42 11 − 11 3.31 36 Superior temporal gyrus BA38 R 49 11 − 11 3.80 29 Precuneus BA7 L − 7 − 67 60 4.09 18 Basal ganglia L − 28 4 − 4 3.43 53 R 25 − 11 18 3.89 69 Middle occipital gyrus BA19 L − 32 − 77 21 3.89 26 Cerebellum L − 25 − 81 − 35 3.92 91 Note: BA=Brodmann area, MNI=Montreal Neurological Institute coordinate system. Table options Table 3. Brain regions showing significant group differences (threshold P < 0.001, five contiguous voxels) in deactivation to sad relative to neutral distractors (MDD > control) Region BA area Hemisphere MNI coordinates T value Cluster size x y z Middle frontal gyrus BA9 L − 28 28 35 3.41 9 Supramarginal gyrus BA40 R 56 − 32 25 3.44 26 Precuneus BA7 L − 7 − 63 46 4.50 55 Postcentral gyrus BA43 L − 49 − 18 21 4.00 68 Superior temporal gyrus BA41 L − 35 − 32 11 4.06 24 Note: BA=Brodmann area. MNI=Montreal Neurological Institute coordinate system. Table options Significantly increased deactivation in response to sad vs. neutral contrast in ... Fig. 4. Significantly increased deactivation in response to sad vs. neutral contrast in the MDD group relative to the control group. Brain maps, voxel-based two-sample t test analysis at the peak time point (P < 0.001, uncorrected, spatial extent of five contiguous voxels); Waveforms, ROI analysis showing the hemodynamic response with time course. The MDD group revealed increased deactivation in response to sad relative to neutral distractors and decreased activation in response to targets compared with the control group. MFG = middle frontal gyrus, SMG = supramarginal gyrus. Figure options Unlike other studies that reported significantly stronger activation to negative emotional stimuli in the emotional system in MDD (Sheline et al., 2001, Siegle et al., 2002, Siegle et al., 2006, Fu et al., 2004 and Canli et al., 2005), the MDD group only showed increased activation relative to the control group using a less stringent statistical threshold (P < 0.05 uncorrected, spatial extent of five voxels) in the emotional system including the left IFG (BA47), hypothalamus, and inferior temporal gyrus, as well as the right uncus, ventral basal ganglia, and ventral extrastriate cortex. We further investigated whether antidepressant medication weakened the activation to emotional stimuli in the MDD group. Relative to the medicated group (n = 11), the medication-free group (n = 8) revealed stronger activation in the left amygdala in response to sad distractors when compared with neutral distractors (P < 0.001 uncorrected, spatial extent of five voxels). 3.2.2. Influence of emotional distraction on the processing of subsequent attentional targets To investigate the influence of emotional distraction on the processing of subsequent attentional targets and the responses of the executive control system to emotional distraction, we first conducted an ANOVA to compare activation to the Target-after-Sad and Target-after-Neutral events in MDD patients relative to controls. The group × stimulus type interaction (Table 4 and Fig. 5) revealed significant effects in the left rostral ACC (BA32) and anterior insula (BA31), right IFG (BA44), and bilateral middle frontal gyrus (BA10). Follow-up ANOVA and Bonferroni post hoc tests were conducted on functional ROIs extracted from these regions and were correlated with the behavioral RT effects. Findings from each of these regions are discussed, in turn, below. Table 4. Brain regions that show an interaction effect of group × target subtype Contrast Region BA area Hemisphere Cluster size Peak voxel ROI F value ROI P value MNI coordinates x y z Control > MDD MFG BA10 L 6 − 39 42 14 6.348 0.016 Target-after-Neutral > Target-after-Sad BA10 R 40 35 42 28 9.580 0.004 BA9 R 7 39 18 35 IFG BA44 R 13 49 14 18 6.966 0.012 Control > MDD Target-after-Sad > Target-after-Neutral Insula BA13 L 25 − 28 18 14 28.801 0.000 ACC BA32 L 46 − 11 35 11 7.233 0.011 Note: the coordinates represents the location of the voxel which showed peak activation in the region using voxel-based ANOVA analysis (threshold P < 0.001, five contiguous voxels); the F and P value were from ROI-based ANOVA analysis. Table options Brain regions which revealed significant interaction effect of group (control, ... Fig. 5. Brain regions which revealed significant interaction effect of group (control, MDD) × target subtype (Target-after-Sad, Target-after-Neutral) in the voxel-based analysis (top), and the mean signal percentage change in these regions in the ROI analysis (bottom) at the peak time point. Figure options The ACC region showed stronger activation to the Target-after-Sad events relative to the Target-after-Neutral events in controls, but not in MDD patients. Post hoc tests confirmed a significant decline in engagement of this region during the Target-after-Sad events in MDD patients relative to controls, t(37) = 2.29, P = 0.028. Stronger activation to the Target-after-Sad relative to the Target-after-Neutral stimuli in controls is consistent with the rostral ACC's putative role in mediating conflict between prepotent distractors and task-relevant stimulus processing. The reduced activation to Target-after-Sad stimuli in MDD patients suggests possible attenuated function of the ACC in conflict control or monitoring. Similar to the left ACC, the left insula also showed stronger activation to the Target-after-Sad events relative to the Target-after-Neutral events in controls, but the pattern was reversed in MDD patients. In contrast to the rostral ACC and insula, the right IFG (BA44) revealed stronger activation in controls for the Target-after-Neutral events relative to the Target-after-Sad events (Fig. 5). Relative to controls, the MDD group had significantly attenuated activation for the Target-after-Neutral vs. Target-after-Sad contrast, t(37) = 2.23, P = 0.034. The MDD group had relatively stronger activation to Target-after-Sad relative to Target-after-Neutral. Similar to the right IFG, bilateral MFG revealed significantly stronger activation in MDD patients for the Target-after-Sad events relative to the Target-after-Neutral events. Consistent with Elliott et al. (2002), we found that controls had the opposite Pattern—relatively stronger activation in this region to the Target-after-Neutral events ( Fig. 5). Post hoc t tests showed that the interaction effect in the right MFG was driven by the increased activation to the Target-after-Sad events in the MDD group relative to controls (t(37) = 2.30, P = 0.027). To identify brain regions which were directly associated with slower RT to Target-after-Sad relative to Target-after-Neutral events in MDD, linear regression analyses were conducted within the MDD group to correlate the RT of Target-after-Sad vs. Target-after-Neutral with the activations to Target-after-Sad vs. Target-after-Neutral events in these frontal regions. Among the four regions which showed a significant interaction effect, only the activation of the right IFG in MDD revealed a significant correlation with the RT difference (left ACC, r(37) = − 0.38, P = 0.11, left insula, r(37) = − 0.23, P = 0.35, right IFG, r(37) = − 0.60, P = 0.007, left MFG, r(37) =− 0.10, P = 0.69, and right MFG r(37) = − 0.10, P = 0.68). As shown in Fig. 6, MDD patients who had stronger activation to the Target-after-Sad vs. Target-after-Neutral contrast in the right IFG showed faster RT when comparing Target-after-Sad vs. Target-after-Neutral events. Thus, the relatively high activation to Target-after-Sad vs. Target-after-Neutral events in these MDD patients might reflect an effortful inhibition processing as a compensatory effect due to the dysfunction of the ACC. Of note stronger activation of control subjects in the right IFG in response to targets (collapsed across the Target-after-Sad and Target-after-Neutral subtypes), but not to the subtypes of targets, was correlated with faster RT in target detection (rRT, target = − 0.65, P = 0.002). Given that the control group had a reverse pattern from the MDD patients, with stronger activation to Target-after-Neutral vs. Target-after-Sad events, the role of the right IFG may differ between the two groups. Nevertheless, both enhanced activation to targets in controls and enhanced activation to Target-after-Sad vs. Target-after-Neutral events were directly correlated with the behavioral outcome, the speed of target detection, indicating a role of this region in successful reallocation of attention from task-irrelevant stimuli to targets. Overall, the MDD subjects as a group had decreased activation in this region and slower RT. Combined, these data implicate altered function of the right IFG in MDD in executive control over emotional distraction and reallocation of attention on task-relevant stimuli. Emotional distraction activity in the right inferior frontal gyrus (IFG-R) ... Fig. 6. Emotional distraction activity in the right inferior frontal gyrus (IFG-R) correlates with reaction time (RT) to detect subsequent targets in MDD. The percent signal change difference in the IFG-R (Target-after-Sad minus Target-after-Neutral) was correlated with the RT difference to the same stimuli in the MDD patients but not in controls.