Neural correlates of suppressing and imagining future threat

Abstract

Suppressing upsetting thoughts can cause psychological distress but might also enhance mental health when used flexibly to suppress the imagination of future threat during challenging times. To investigate the neural correlates of suppressing and imagining future threat, a cohort of 65 participants underwent a previously examined "Imagine/No-Imagine" paradigm while examining brain activation using magnetic resonance imaging. We observed activity of the inferior frontal gyrus, middle frontal gyrus (MFG), superior parietal lobule, and superior occipital sulcus during thought suppression, whereas imagining future threat elicited activation in the bilateral posterior cingulate cortex (PCC) and ventromedial prefrontal cortex (vmPFC). Subjective levels of anxiety, stress, and depression as covariates did not alter these results. To further examine the group and individual-level network dynamics, we conducted dynamic causal modeling (DCM) and group iterative multiple model estimations (GIMME). The DCM model showed that during suppression, the MFG positively influenced the vmPFC and right PCC. In contrast, the vmPFC and the left and right PCC showed positive connections to the MFG during imagining. This suggests that the neural correlates of self-regulation involve an information flow between the PCC and the PFC. In addition, GIMME identified group-level connections between the right and left PCC and between the left PCC and vmPFC, reflecting the information flow during suppression and imagination of future threat, respectively. Considerable interindividual heterogeneity in the connectivity patterns became apparent, pointing to the existence of different biotypes.

Keywords: Brain network connectivity; DCM; GIMME; Network analysis; Neural correlates; Thought suppression.

Fig. 1

The imagine/no-imagine task (INI) paradigm. The experimental protocol spanned two consecutive days. On the first day, participants generated and rated 18 personal fear-provoking scenarios. For example, the participant might describe sitting at the kitchen table in the evening when their partner announces they want a divorce. They could elaborate on how emotions like anger and sadness emerge, and how tears start to well up in their eyes. For each scenario, they provided a reminder word, i.e., ‘divorce’ as an obvious cue and a code word representing a typical aspect of their imagination, i.e., ‘glass’. They rated these scenarios on a five -point scale. On the second day, MRI measurements were conducted. Before scanning, participants underwent a preparatory training session outside the scanner to familiarize themselves with the task. Participants repeatedly imagined events associated with green reminder words while trying to suppress any thoughts related to episodes with red reminder words. During the MRI scan, the protocol began with a localizer and AAH scout scan, followed by T1-weighted (T1w) structural imaging. The fMRI task included four blocks of 30 words each, displayed for 5 s with a variable inter-stimulus interval (1.5 to 7.5 s, averaging 2.5 s). Diagnostic questions were asked between runs to ensure task adherence. After the task runs, a field map was acquired, and data collection concluded with a recall phase where participants identified the previously generated scenarios.

Fig. 2

Each region that survived the FPR threshold for the condition suppress > imagine; (a) IFG = inferior frontal gyrus, (b) MFG = middle frontal gyrus, (c) SOS = superior occipital sulcus, (d) SPL = superior parietal lobule.

Fig. 3

Each region that survived the FPR threshold for the condition imagine > suppress; (a) right PCC, (b) left PCC, (c) vmPFC.

Fig. 4

The winning model of the DCM analysis, including the strength of the connection between the regions. For between-region parameters (measured in Hz), positive values indicate excitation, whereas negative values indicate inhibition. Positive values for self-connections represent increased self-inhibition, whereas negative values indicate disinhibition. In the suppress condition (shown in red), the MFG showed positive connections to the PCCr and the vmPFC, with all regions exhibiting disinhibitory self-regulation. In the imagining condition (shown in green), both PCCs and the vmPFC showed positive connections to the MFG with all regions exhibiting negative self-regulation.

Fig. 5

The GIMME network of the included ROIs. In this graph, solid lines represent contemporaneous relations (lag 0), and dashed lines reflect lagged relations (lag 1). The width of paths corresponds to the estimated path weight (corresponds to the count of connections). Black represents group-level paths (for all participants), and gray represents individual-level paths. Condition (Suppress vs. Imagine) is the exogenous variable and is set to the time points where the participants were asked to suppress. Connections indicate an influence of the task condition on the respective ROI. A: shows the results for the group. B and C: Individual results for two exemplary participants (red paths represent positive weights and blue paths represent negative weights).