Abstract: PURPOSE: To evaluate the characteristics of emotion-related brain networks and the differences between no-memory load condition and high-memory load condition with fMRI.
METHODS: 26 participants underwent comprehensive psychological assessments, including Emotion Regulation Questionnaire-10(ERQ, based on the emotion regulation, we divided it into two parts: Reappraisal and Suppression), Toronto Alexithymia Scale (TAS-12, including the difficulty identifying feelings questions and the difficulty describing feelings questions), General Health Questionnaire-12(GHQ). Participants completed an emotional Face N-Back (EFNBACK) task . In this task, participants were asked to respond to the prescribed letters when watched a pseudo-random letter sequence. It includes 0-back no-memory load condition (EF‐0‐back) and 2-back high-memory load condition (EF‐2‐back). Faces with expressions of fear, happiness or neutral faces as distractors, or no faces (no-face) randomly appear on both sides of the letters. There are eight stimulus blocks: two memory load conditions (0-back and 2-back), each of which has one of four facial interference states. The task includes three runs of 7 minutes and 4 seconds, with a total of 24 blocks, which are presented in pseudorandom order. Each block contains 12 500ms trails, including two sides of a letter that are either the same picture of the actor's facial expression or no picture.
RESULTS: Across all participants, 3 significant associations were found: TAS-12 score positively correlated with EF-2-back neutral face versus no face rAMY-rOFA (r = 0.612, P = 0.001), EF-0-back fearful face versus no face lAMY-rOFA (r = 0.510, P = 0.009); The negative correlation between EF-0-back happy face versus neutral face rAMY-rOFA and TAS-12 score (r = -0.537, P = 0.006). Significant positive correlations were found between EF-2-back versus EF-0-back fearful face rAMY-lFFA (r = 0.580, P = 0.002), EF-2-back versus EF-0-back happy face rAMY-rOFA (r = 0.512, P = 0.009, Figure 3b) and Reappraisal score. Suppression score negatively correlated with EF-0-back happy face versus no face left amygdala-AC (r = -0.501, P = 0.011) and EF-0-back fearful face versus neutral face rAMY-lpSTS (r = -0.511, P = 0.009).
CONCLUSIONS: Our study found that even under high-demand working memory task conditions, implicit emotion processing can still be activated, and subsequent research is needed to determine whether implicit emotion regulation exists.

Key words: emotion processing, emotion regulation, working memory, implicit emotion