Distinct roles of hippocampus and neocortex in symbolic compositional generalization

The paper investigated which brain regions support symbolic compositional generalization by having participants combine two discrete features (shape and color) to make novel spatial inferences while measuring BOLD responses. It found that the hippocampus encoded elementary visual attributes in a high-dimensional parallel representation enabling flexible individuation, whereas in vmPFC, PPC, and V1 the neural patterns for novel composites could be predicted from familiar elements as linear combinations. It further reported a regional distinction: vmPFC performed this composition in a high-dimensional format, while PPC and V1 used a low-dimensional, spatial, response-consistent frame of reference. The paper’s caveat is that the conclusions are based on fMRI BOLD measures, which are indirect correlates of neural activity. The paper does not explicitly discuss endometriosis or adenomyosis; it was included in the corpus via a keyword match in the upstream search index.

Read from the paper's body, not the abstract. Not a substitute for reading the paper. No clinical advice. How this works