Suppressing unwanted emotional memories disrupts their neural representation and amplifies subsequent negative emotional reactivity : evidence from EEG-based decoding

Abstract

Unwanted emotional memories can be voluntarily forgotten by suppressing their recall. Yet, it remains unclear whether and how such retrieval suppression impacts the spontaneous/implicit and deliberative/explicit emotional reactivity towards suppressed emotional memories upon their re-exposure. In the present dissertation, we examined the memory and emotional after-effects of retrieval suppression. Thirty-seven participants first underwent an emotional Think/No-Think (eTNT) task with electroencephalography being recorded. To gauge neural and behavioural indices of participants’ spontaneous/implicit emotional reactivity, we combined EEG recording with an adapted affect misattribution procedure; we refer to this task as the Implicit Affect Task. For deliberative/explicit emotional reactivity, we measured participants’ self-reported affect ratings towards cues of emotional memories and towards emotional memories per se.

We found behavioural evidence that suppressing the retrieval of emotional memories induced forgetting in the post-eTNT cued recall task. In terms of emotional rating, we found that retrieval suppression enhanced deliberative/explicit negative emotional reactivity towards previously suppressed negative memories when they were directly re-encountered. Memory suppression did not alter the implicit and explicit emotional ratings toward memory cues. EEG analyses showed that during the eTNT task, suppression enhanced neural signals associated with inhibitory control (frontocentral N450) while reducing electrophysiological activity associated with episodic recall (parietal P300), working memory maintenance (prefrontal NSW), and emotional reactivity (parietal LPP) relative to voluntary retrieval. Multivariate EEG-based decoding during the eTNT revealed that neural activity pattern was distinct upon viewing cues associated with emotionally negative vs. neutral memories during suppressed trials, only during 210-340 ms post cue onset. During later time window (400-3000 ms post cue), the neural activity pattern for cues associated with negative vs. neutral memories was indistinguishable during suppressed trials. In voluntarily retrieved trials, there was distinct neural activity pattern for cues paired with negative vs. neutral memories at multiple clusters, specifically during 190-260 ms and 1960-2050 ms post cue onset. The decoding evidence suggests that suppressing emotional memories deteriorated their valence-dependent neural representation.

Neural marker sensitive to emotional reactivity (parietal LPP) recorded during the Implicit Affect Task was also attenuated for No-Think cues than for Think-cues, suggesting that the emotional after-effects of memory suppression persisted after eTNT. Further corroborating the emotional after-effects of memory suppression, multivariate decoding analyses revealed that neural activity pattern was indistinguishable only between suppressed negative and neutral memory. Taken together, our results suggest that suppressing unwanted memories can induce their forgetting as well as disrupt emotion-dependent neural representation in response to cues. However, suppression also amplified negative emotional reactivity towards previously suppressed negative memories when they are re-confronted. Our findings offer fine-grained insight into the dynamic influence of memory suppression on the implicit and explicit affect reactivity and bear theoretical implications for future research on the link between active forgetting and affect regulation.