Rumination and Over-Recruitment of Cognitive Control Circuits in Depression

Abstract

Background:

Rumination is associated with greater cognitive dysfunction and treatment resistance in major depressive disorder (MDD), yet its underlying neural mechanisms are not well understood. Since rumination is characterized by difficulty in controlling negative thoughts, the present study investigated whether rumination is associated with aberrant cognitive control in the absence of negative emotional information.

Methods:

Individuals with MDD (n=176) and healthy volunteers (n=52) completed the Stop Signal Task with varied stop signal difficulty during functional magnetic resonance imaging. In the task, a longer stop signal asynchrony made stopping difficult (Hard-stop) while a shorter stop signal asynchrony allowed more time for stopping (Easy-stop).

Results:

In MDD participants, higher rumination intensity was associated with greater neural activity in response to difficult inhibitory control in the frontoparietal regions. Greater activation for difficult inhibitory control associated with rumination was also positively related to state fear. The imaging results provide compelling evidence for the neural basis of inhibitory control difficulties in MDD individuals with high rumination.

Conclusions:

The association between higher rumination intensity and greater neural activity in regions involved in difficult inhibitory control may provide treatment targets for interventions aimed at improving inhibitory control and reducing rumination in this population.

1. Introduction

Major depressive disorder (MDD) is a common mental health disorder, with 8.4% of adults in the United States reporting at least one major depressive episode in 2020 (National Institute of Health, https://www.nimh.nih.gov/health/statistics/major-depression). Although MDD is considered a heterogeneous disorder without a clearly established mechanism, with a variable course and response to treatment (1), dysfunctional affective processing such as sustained negative affect is characteristic of this disorder (2,3). In addition to negative affect, rumination marked by the tendency of repetitive negative thinking is a crucial risk factor for depression. Repetitive negative thinking is associated with a series of adverse outcomes in depression including its onset, severity, and treatment outcome (4,5,6,7,8). Since rumination is intertwined with the difficulty of regulating negative thoughts with negative affect on depression, the association between rumination and cognitive control has been studied (9,10,11).

Cognitive control utilizes several processes including inhibiting task-irrelevant processing, shifting attention and resources to task-relevant processing, and updating information in light of the current need (12). Extant research indicates that impairment of cognitive control in depression could be related to the dysregulation of negative thoughts and affect underlying rumination (13,14). However, the association between rumination and cognitive control is still unsettled. Some studies suggest an association between rumination and deficits in cognitive control (15,16,17), while other studies reported no such significant associations (18,19,20). Available meta-analyses do not indicate impairments in overall cognitive control linked to rumination such that some found negative associations between rumination and inhibition and set-shifting but no association between rumination and working memory whereas the other yielded a negative association between rumination and discarding irrelevant information but no other relationship between rumination and cognitive control functions including inhibition (17,20). Another meta-analysis study examining the relationship between response inhibition and psychopathology reported only a small or negligible relationship in MDD compared to other disorders (21).

Inherent in depression is a deficit in the regulation of negative information. Several studies suggest cognitive control deficits associated with rumination as inhibiting and updating negative information (13,22,23,24). In fact, rumination is closely related to individuals’ responses to difficult and negative events. The difficulty in suppressing negative information has been linked to reactivity to stress (9,25,26). Difficulties inhibiting and updating negative emotional information predicted the intensity of rumination upon stress, and in turn, rumination seems to be associated with heightened reactivity to stress in MDD (27,28). These findings imply an interplay between the cognitive process and the affective process with rumination.

Negative thoughts and negative affect tend to divert individuals from making situationally appropriate responses. In this regard, engagement in negative information processing may consume cognitive resources that could have been used for task processing (29,30). Consequently, negative thoughts and affect may result in summoning more effort to achieve the same level of performance due to inefficient allocation of cognitive resources to task processing (31,32,33). Similarly, rumination may impose an interference load as task-irrelevant processing (34,35,36). Behaviorally, both processing inefficiency and interference could be negatively related to accuracy and response time (RT), although the relationship between accuracy and RT is more complex beyond simple speed-accuracy tradeoffs (37,38) and different models yield different predictions and mixed findings about the role of rumination on accuracy and RT (39,40,41).

Regardless of behavioral differences, rumination was often associated with greater neural activity in the frontal and parietal regions where goal-directed control and attention are often localized (42,43), implying aberrant neural activity for cognitive control. Rumination was also associated with default mode network functioning (see [44] for review). Frontal regions, including the left inferior frontal gyrus, have been implicated in the neural circuit regulating cognitive processes (45,46). In particular, response inhibition related to rumination involves activity in the frontoparietal network including the dorsolateral prefrontal cortex, sensory motor cortex, temporoparietal junction, and precuneus (see [43] for review).

Besides, individuals with MDD and intense rumination showed aberrant fear processing including extended fear response during extinction (46), as hypothesized in a preliminary communication (47). These findings suggest that fear may represent a negative valence of emotion contributing to reduced cognitive control in highly ruminative MDD individuals. Regarding uncontrollable stress response which tends to be linked to rumination, patients with severe depression show impaired fear extinction (48,49). These findings imply that negative affect such as fear may overload or interfere with cognitive processing in highly ruminative MDD patients, even in the absence of explicit emotional material.

The Stop Signal Task (SST [50]) is a well-established experimental paradigm for examining cognitive control, specifically inhibition of prepotent responses. SST typically includes variable ranges of stop signal delay by trial, calibrated according to each individual’s average response time. The stop signal presented earlier in the trial allows more time to abort the prepotent go response (Easy-stop), whereas the stop signal presented later in the trial leaves less time to stop the imminent go response if the response has not been already committed (Hard-stop). In this paradigm, Hard-stop tends to put more demands on the task for response inhibition, which may lead to being perceived as a difficult event compared to Easy-stop.

In the present study, we tested the hypothesis that rumination is associated with aberrant inhibitory control in MDD when the demand for inhibitory control is high, due to the cognitive effects of rumination while responding to difficult situations, instead of the deficiency in overall inhibitory control. In light of previous studies, we expected greater activity for inhibitory control upon high demand for control. We also hypothesized that aberrant control processing associated with rumination would be related to negative affect (e.g., fear). That is, high fear would be associated with aberrant inhibitory control that is related to rumination. We thus predicted less efficient neural processing and more negative affect (i.e., more fear) for difficult inhibitory control with higher rumination as a major interface linking environmental demands and rumination in MDD.

2. Methods

2.1. Participants

The present study participants were drawn from the first 500 individuals who completed the baseline assessments in the Tulsa-1000 study (T-1000). Inclusion and exclusion criteria for the T-1000 study are described in Supplementary Information (see [51] for the full screening procedure). The Mini International Neuropsychiatric Inventory (MINI, [52]) was used for MDD diagnosis, followed by consultation with board-certified psychiatrists. Only those who were diagnosed with MDD (n=184) were included in the present study along with healthy volunteers (n=53) from the initial 500 (total=237) to examine rumination in depression. Nine participants (8 MDD, 1 HC) were excluded from the analysis due to excessive head motion and imaging quality, leaving a total of 228, those with MDD (n=176, 122 with anxiety & 54 without anxiety) and HCs (n=52). Participants completed the Ruminative Responses Scale (RRS, [53]) as the rumination measure and the Positive Affect Negative Affect Schedule-Fear (PANAS-Fear, [54]) for reporting their current state fear, as well as the Patient Health Questionnaire (PHQ-9, [55]) and the Overall Anxiety Severity and Impairment Scale (OASIS, [56]) for the symptom severity of depression and anxiety respectively. The study was approved by the Western Institutional Review Board. Participants provided written informed consent and received remuneration for study participation. The demographic and clinical information of participants is displayed in Table 1.

Table 1.

Demographic and Clinical Characteristics of Participants

	MDD (n=176)	HC (n=52)	p-value
Age	34.43 ± 11.15	31.78 ± 10.96	0.13
Male, n (%)	40 (22.7)	25 (48.1)	0.001
Ethnicity, n (%)			0.74
Asian	2 (1.1)	2 (3.8)
Black	13 (7.4)	2 (3.8)
Hispanic	8 (4.6)	3 (5.8)
Native American	28 (16.0)	7 (13.5)
White	118 (67.4)	36 (69.2)
Other	6 (3.4)	2 (3.8)
Education, n (%)			0.59
No High School	6 (3.4)	0 (0.0)
High School	24 (13.8)	7 (13.5)
Some College	70 (40.2)	21 (40.4)
College or Higher	74 (42.5)	24 (46.2)
BMI (SD)	28.78 (5.35)	27.81 (5.61)	0.26
Employed, n (%)	114 (66.7)	43 (84.0)	0.03
Income (SD)	$61,225 (90,978)	$58,417 (49,753)	0.84
Psychotropic Medication, n (%)	115 (65.3)	7 (13.5)	<0.001
OASIS (SD)	9.53 (3.55)	1.35 (1.98)	<0.001
PHQ-9 (SD)	12.89 (4.97)	0.92 (1.43)	<0.001
PANAS-Fear (SD)	13.12 (4.99)	7.63 (2.04)	<0.001
RRS (SD)	56.03 (11.64)	30.25 (8.16)	<0.001

Abbreviations: MDD, major depressive disorder; HC, Healthy Control; BMI, Body Mass Index, OASIS, Overall Anxiety Severity and Impairment Scale; PHQ-9, Patient Health Questionnaire; RRS, Ruminative Response Scale.

2.2. Imaging task: Stop Signal Task (SST)

The task requires participants to control a motor response by making a response (‘go’ ‘ trial) or stopping the response (‘stop’ trial) according to a trial cue (50). The task consisted of 6 blocks separated by 12-s breaks, with each block comprised of 48 trials, 36 go trials (75%) and 12 stop trials (25%), yielding a total of 288 trials. A trial was 1300-ms long followed by a 200-ms blank interval. A trial cue (‘X’ or ‘O’) in white was presented on a black background for each trial. On the ‘go’ trials, participants were instructed to make a response following the trial cue, pressing the left button for ‘X’ and the right button for ‘O’ as quickly as they could. On the ‘stop’ trials, an auditory tone signaled participants to stop the button-press response, with the color change of the trial cue from white to red. The delivery time of the tone (i.e., the stop signal delay) was individually calibrated for each participant, by their mean response times (RTs) collected from practice. The stop signal delay ranged 500-ms, 400-ms, 300-ms, 200-ms, 100-ms, or 0-ms less than one’s mean RT. Thus, the stop trials presented stop signals early in trials (500, 400, 300 less than mean RTs) were relatively easier to stop responses (Easy-stop) whereas the stop trials with stop signals presented late (200, 100, 0 less than mean RTs) were more difficult to stop responses (Hard-stop).

2.3. Imaging data acquisition and preprocessing

Two identical GE MR750 3T scanners equipped with 8 RF channel phased array coils were used to acquire both structural and functional scans in the same site. T₁-weighted 3D high-resolution anatomical images were acquired using the magnetization-prepared rapid gradient echo (MP-RAGE) pulse sequence with FOV 240 × 192 mm², flip angle 8°, voxel size of 0.938 × 0.938 × 0.9 mm³, TR/TE = 5/2.012 ms, and 186 axial slices. T₂*-weighted echo-planar images (EPIs) were acquired in FOV 240 × 240 mm², flip angle 78°, voxel size of 1.875 × 1.875 × 2.9 mm³, TR/TE = 2000/27 ms, and axial plane with the sensitivity encoding (SENSE) factor of 2. Each volume was comprised of 39 slices in an interleaved sequence and 256 EPI volumes were collected. Preprocessing and statistical analyses of MRI data were performed using the Analysis of Functional NeuroImages software suite (AFNI, [57], http://afni.nimh.nih.gov). The first three EPI volumes were discarded to allow signal equilibrium, and the remaining volumes were despiked to remove transient signal spikes. For preprocessing, slice-time correction to the first slice was done with co-registration to a T1-weighted anatomical image. Motion correction via affine registration was performed, with censoring time points with large head motion using the threshold of an average Euclidean norm (ENORM) of the derivatives of the six motion parameters greater than 0.3. The resultant data were normalized to the MNI space with resampling of 2 mm isotropic voxels and smoothed with an isotropic 4 mm full width at half maximum (FWHM) Gaussian kernel.

2.4. Imaging data analysis

General linear models were used to examine the association between the intensity of rumination and neural activity (BOLD signal) for response control on SST by the difficulty of control. First, three regressors were constructed on a subject: ‘Go’ (go trials), ‘Easy-stop’ (stop trials with longer times: 500-ms, 400-ms, & 300-ms less than one’s mean RT), and ‘Hard-stop’ (stop trials with shorter times: 200-ms, 100-ms, & 0-ms less than one’s mean RT). Only correct trials were entered into the analysis to prevent the influence of different error rates by the difficulty of control.¹ Incorrect responses were modeled as nuisance covariates with six motion parameters. The hemodynamic response for a trial was convolved with a block function (1-s) from the onset of the task cue. We examined the relationship between rumination (RRS total score) and Hard control (Hard-stop > Easy-stop) with covariates of age and sex in individuals with MDD using a multiple regression model (3dLME) in AFNI (57). A voxel-wise threshold of p < .0005 was adopted with false discovery rate correction (FDR < .05). Significant cluster effects over 100 voxels were queried with post-hoc analyses on beta coefficients extracted from the suprathreshold clusters. Bonferroni correction was used for controlling a type-I error with multiple comparisons in post-hoc analyses. Current state fear (PANAS-Fear) was examined with the beta coefficients to probe the nature of neural activity in the relationship between rumination and cognitive control by difficulty in seeing the relationship between cognitive control and fear in rumination (p < .01).

3. Results

3.1. Participant characteristics

Participants did not differ across age, ethnicity, education, BMI, and income, but individuals with MDD comprised more females, more unemployed, and greater use of psychotropic medications, as well as higher scores in depression, anxiety, and rumination compared to HCs (Table 1). Rumination (RRS) was positively correlated with state fear (PANAS-Fear, r[226] = .54, p < .001) and depression symptom severity (PHQ, r[225] = .76, p < .001), and fear and depression symptom severity were also positively related (r[225] = .57, p < .001).

3.2. Behavioral results

Behavioral accuracy did not differ between MDD and HC (Table 2). Accurate Go responses decreased as a function of rumination intensity, F[1,222] = 7.44, p = .007, without the main group effect. The accuracy of stopping, both Easy- and Hard-stops, was not related to RRS scores. An accuracy analysis incorporating Go and Stop responses (e.g., [Correct Go + Correct Stop] / [Correct Go + Correct Stop + Error Go + Error Stop]) did not yield group differences or an association with RRS.

Table 2.

Mean accuracy (probability) and reaction time (s) by trial type (SD in parenthesis)

Trial	MDD (n=176)	HC (n=52)	p-value
GO
Correct response  Go RT	.90 (.10) .79 s (.14)	.93 (.07) .79 s (.14)	0.06 0.93
Omission error	.08 (.10)	.06 (.06)	0.13
Choice error	.02 (.02) .50 s (.22)	.01 (.02) .54 s (.19)	0.57 0.37
STOP
Easy – Stop  Error (Go) RT	.82 (.13) .50 s (.09)	.80 (.16) .51 s (.09)	0.58 0.70
Hard – Stop  Error (Go) RT	.44 (.18) .66 s (.12)	.39 (.14) .65 s (.12)	0.05 0.69

3.3. Imaging results

The main task effect of overall inhibitory (Stops > Go) was evident in bilateral frontal cortices including middle frontal gyri, extensive temporal and parietal cortices, and visual areas (Supplementary Figure 1). Task effects by each condition are shown in Supplementary Figure 2. Rumination intensity was not related to the overall Stop control (all Stops > Go) above the threshold. Those who did not make sufficient numbers of correct stop responses (< 5) were excluded (n=6) from further analyses (n=222, 170 MDDs & 52 HCs). As shown in Table 3, higher rumination in depression was related to the activity of Hard control (Hard-stop > Easy-stop) in bilateral supramarginal gyri, right middle frontal gyrus, and left postcentral gyrus. These regions exhibited a positive relationship between rumination intensity and neural responses to Hard control in MDD, such that greater BOLD signals for Hard control were found in those with higher RRS scores (Figure 1). No relationship between Hard control and RRS was found in HCs.

Table 3.

Regions showing positive association between RRS scores and cognitive control difficulty

Peak coordinates (x,y,z)			Number of voxels	Region	β	95% CI		F-statistics
−57	−43	41	371	L supramarginal gyrus	+0.045	+0.021	+0.068	15.55 (p = .0001)
5	17	53	238	R middle frontal gyrus	+0.038	+0.015	+0.060	12.58 (p = .0005)
−11	−43	75	206	L postcentral gyrus	+0.044	+0.020	+0.068	13.63 (p = .0003)
59	−23	39	136	R supramarginal gyrus	+0.037	+0.016	+0.058	13.41 (p = .0003)

Note: β denotes the estimated response magnitude after standardization (M=0, SD=1); positive values for increased BOLD signals and negative values for decreased BOLD signals

Figure 1.

Brain regions showing positive association between RRS scores and Hard control

Post-hoc analysis showed that all of these clusters revealed greater activity during Hard-stop with higher RRS scores in MDD after Bonferroni correction (Figure 2): L supramarginal gyrus, F[1,166] = 12.34, p = .002; R middle frontal gyrus, F[1,166] = 13.19, p = .001; L postcentral gyrus, F[1,166] = 19.39, p = .00008; R supramarginal gyrus, F[1,166] = 11.75, p = .002. However, an association between RRS scores and neural activity during Easy-stop was not found, except L supramarginal gyrus F[1,166] = 10.55, p = .006. Hard control activity associated with rumination did not differ between whether participants took psychotropic medications: L supramarginal gyrus region F[1,127] = 2.93, p = .09; R middle frontal region F[1,127] = 1.07, p = .30; L postcentral region F[1,127] = 1.87, p = .17; L supramarginal region F[1,127] = 1.07, p = .17.

Figure 2.

Association between RRS and Hard control in task conditions

Figure 3 displays the relationship between state fear and neural activity for Hard control. That is, greater neural activity for Hard control in MDD is associated with not only higher RRS scores but also higher state fear: L supramarginal gyrus, F[1,168] = 11.05, p = .004; R middle frontal gyrus, F[1,168] = 11.85, p = .003; L postcentral gyrus, F[1,168] = 12.84, p = .002; R supramarginal gyrus, F[1,168] = 12.54, p = .002. No correlation between Hard control and fear was found in HCs. A supplementary analysis with a matching proportion of females in both MDD and HC to control different proportions of females in the two groups also yielded identical patterns of results of the positive association between heightened neural activity and higher rumination and the association between heightened neural activity and higher state fear in the frontoparietal regions in MDD, with increased effect sizes overall (Supplementary Figure 3). The relationship between depression symptom severity assessed with PHQ9 and neural activity during Hard control was not significant after correcting multiple comparisons.

Figure 3.

Positive association between neural activity for Hard control and fear

4. Discussion

This study examined whether rumination in MDD is related to neural processing of inhibitory control with the level of control difficulty using the Stop Signal task during fMRI. There were three main findings. First, more intense rumination was associated with greater neural activation during difficult control in the frontoparietal regions including supramarginal gyri, postcentral gyrus, and middle frontal cortex. Second, neural activity for difficult control linked to rumination was also positively correlated with self-reported state fear but not with behavioral accuracy. Third, there was no association between rumination and overall inhibitory control. Taken together, these findings are most consistent with the notion that MDD participants with high rumination show altered inhibitory processing only when the demand for inhibitory cognitive control is high.

The primary finding of this study is the association between rumination and neural activity for Hard control in bilateral supramarginal gyri, L postcentral gyrus, and R middle frontal gyrus. These regions could participate in the frontoparietal network (intercorrelations among 4 clusters (r): 0.92–0.94), known for guiding goal-directed behavior, namely cognitive control. However, these regions are also parts of other networks, including the default mode network (DMN). Altered functional signals in the frontoparietal network have been reported in psychopathological populations including individuals with MDD and substance use (58,59). At the same time, abnormal DMN function is also associated with psychopathological disorders such as MDD and schizophrenia (60), particularly with rumination (61).

The present finding of greater activity associated with higher rumination for Hard control can be interpreted in different ways albeit non-exclusive. First, greater activity in these regions related to rumination intensity may imply more effortful control processing in these circumstances, indicating that individuals with higher rumination may have to recruit more neural resources for the same level of inhibitory control, particularly when exerting control was difficult. That is, over-recruitment of neural activity is summoned to compensate for lower mental resources (62,63). In addition, greater neural activity probably reflects more engagement in a task due to more effortful but inefficient task processing. Greater neural activity for Hard control with more intense rumination may indicate a higher effort-to-success ratio in highly ruminative individuals, who thus need more effort to attain success (64,65). Further, based on the presence of repetitive negative thinking and implicit negative emotion associated with it, it is plausible that individuals with intense rumination engage in ‘dual tasks’, SST and negative self-referential processing. A dual-task component is known to demand greater neural activity due to increased cognitive load for task-switching and interference between competing tasks (66,67).

These proposed mechanisms implicate inefficiency in neural processing associated with rumination. Greater activity for Hard control was found in the frontoparietal regions where cognitive control over executive functions is often proposed to occur, along with attention to handling task demand and cognitive load. Patients with higher rumination showed greater neural activity related to Hard-stops than Easy-stops. Since both stop conditions required inhibitory control, the difference between these conditions could represent inefficient neural processing with higher rumination, particularly in difficult situations. The specificity of neural engagement in the frontoparietal regions suggests that inefficient neural processing associated with rumination is likely to be linked specifically to the difficulty in cognitive control. Thus, we propose that the current findings of greater neural activity in the frontoparietal regions may imply inefficient cognitive control in those with high rumination when subjected to difficult control situations.

Further, greater neural activity for Hard control in the frontoparietal region was associated with state fear, suggesting a specific role of negative self-referential mentation in cognitive control. A positive association between fear and frontoparietal activity for inhibitory control suggests that greater neural activity with rumination in these regions reflects inefficient processing of control demands associated with heightened negative affect, even when those individuals with rumination were able to abort responses (i.e., stopping). Difficult tasks tend to require more mental resources, both cognitive and affective. In fact, difficulty in regulating one’s own negative emotions may exacerbate impairment in cognitive control under difficult situations, possibly because mental resources may have been allocated to task-irrelevant negative emotion processing when cognitive control is even more crucial. Fear correlated to greater neural activity for Hard control may reveal the difficulty of responding to taxing situational demands in ruminative depression. As such, the present results indicate that rumination owes its maladaptive and adverse prognostic features at least in part to deficits in reactivity to environmental changes. In this regard, the present finding is in parallel with a structural finding of aberrant disposition of white matter fibers reaching the postcentral gyrus (68) and the view that the sensory-motor region might have a critical role in MDD and anxiety symptom generation with abnormal environmental demands (69,70,71).

While previous studies used negatively-valenced materials to investigate the dysregulation of negative emotion associated with rumination, the present study employed a cognitive experimental paradigm with neutral materials. Even in this circumstance, the finding of greater neural activity with higher rumination and the relationship between neural activity and fear on response control evidenced aberrant inhibitory control with heightened rumination. The present findings indicate that rumination is associated not only with inefficient control processing of relevant information where an individual is situated but also with heightened negative affect in a challenging circumstance. We should also mention that the potential collinearity among rumination, negative affect, and depression due to their correlations mandates caution while interpreting our results. In this respect, the null finding between Hard control and PHQ9 scores should be tempered, given that the limitation imposed by the depression scale may have not been present if a depression scale with wider score ranges had been used. Future studies examining a potential intervention to ameliorate rumination and its cognitive consequences (e.g., repetitive transcranial magnetic stimulation) targeting the frontoparietal area may be beneficial to examine the causative role of rumination in cognitive control.

This study did not find the association between rumination and overall inhibitory control. While this finding may support the deficits of inhibitory control in rumination interacting with situational difficulty, the null finding warrants a systematic investigation of rumination and inhibitory control with difficulty. Further, the current findings should be interpreted with caution due to limitations. First, the current analysis did not include stop signal reaction times, thus we were unable to examine theoretical models based on reaction time differences with the inhibition index. Second, individuals with MDD were not propensity-matched. Confounding of depression severity and rumination limits interpretations of the findings. Third, the number of healthy volunteers was smaller than the number of MDDs in this study, resulting in relatively smaller effect sizes in HC. Future studies may provide information conjugating the current results with propensity-matched samples and equal numbers of participants in groups. In addition, we used relatively stringent voxel-wise and cluster-level thresholds in an effort to avoid adding false positives to the literature in a field where this is a major concern. However, we acknowledge that this choice could also be seen as a limitation of the current study, as it increased the risk of reporting false negative results. Nonetheless, the present study clearly demonstrates the potential influence of rumination in handling challenging situations with inefficient cognitive processing.

Conclusion

This study provides evidence that rumination as the negative tendency in MDD is intricately linked with inefficient cognitive processing upon demanding environmental challenges. The present findings reveal that rumination is associated with heightened neural activity in the brain regions responsible for cognitive control, particularly in challenging situations that necessitate enhanced cognitive control. These results shed light on the dysfunctional control processing that characterizes rumination in contexts demanding more cognitive control. Considering the critical role of adaptability in navigating ever-changing environments, these findings exhibit the risk of rumination in cognitive control, thereby uncovering a potential target for therapeutic interventions aimed at addressing rumination in depression.