Avoiding actions with negative consequences is fundamental to adaptive behaviour, yet individuals differ drastically in their tendencies to avoid to punishment. Using a conditioned punishment task in male and female rats, we replicate findings of a bimodal distribution of punishment avoidance, and identify four distinct behavioural phenotypes (Sensitive, Hypersensitive, Insensitive, Generalised) that differed in their punishment avoidance, unpunished reward-seeking, and Pavlovian fear. We examined how these phenotypic differences related to c-Fos (transcriptional activity marker) across prefrontal cortex (PFC) regions. While region-level c-Fos differences were minimal, covariance and graph-theoretic analyses revealed three functional PFC networks predominantly composed of medial PFC, ventrolateral PFC, or sensorimotor regions. Activity in the ventrolateral PFC network predicted punishment avoidance, while activity in the medial PFC network predicted generalised suppression during initial punishment learning. These findings suggest phenotypic differences in punishment sensitivity is underpinned by differential recruitment of distinct cortical networks that drive adaptive versus maladaptive responses to negative outcomes.
The Hidden Power of Normalization: Exponential Capacity Control in Deep Neural Networks
arXiv:2511.00958v1 Announce Type: cross Abstract: Normalization methods are fundamental components of modern deep neural networks (DNNs). Empirically, they are known to stabilize optimization dynamics and