Back

Sleep neural code perpetuates the evolving negativity bias under stress

Um, S. M.; Okubo-Suzuki, R.; Nomoto, M.; Hanafy, M. K.; Murayama, E.; Yamada-Nomoto, K.; Asai, H.; Choko, K.; Suzuki, A.; Inokuchi, K.

2026-03-04 neuroscience
10.64898/2026.03.03.709197 bioRxiv
Show abstract

Negativity bias is the distorted cognitive processing of how self-experiences are negatively perceived through rumination and overgeneralization. Despite proposed theories of higher-order emotional representation and the involvement of default mode networks, the biological mechanisms underlying how negative emotional bias is persistently generated remain unknown. Here, we show that under negative emotional state, sleep orchestrates neural dynamics to shape persistent negativity bias. A negative emotional state, induced by repeated social defeat stress (SoD), generates negativity bias in both memory-recall and novel experiences such as fear conditioning. Longitudinal Ca2+ imaging of the hippocampus across days revealed that in emotional states, negative schema-like representations that encode the semantic aspect of negativity across experiences are generated. These negative schema-like representations are continuously processed along with dynamic co-reactivations of distinct emotional experiences during sleep. Closed-loop optogenetic silencing of stress experience-tagged cellular ensembles during sleep, but not during awake, prevents negativity bias in future memory-recall and novel experiences. Finally, sleep-active, but not sleep-nonactive, SoD cells predict and decode emotional behaviours during learned and novel situations. Together, these results reveal the biological emergence of semantic-like representations of emotional experiences and sleep causally coordinate neuronal dynamics to persistently shape negativity in future conscious states. These findings offer a new perspective on how higher-order emotional processing during sleep may determine our emotional interpretation of past and future experiences.

Matching journals

The top 3 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.