English

Multimodal neural networks better explain multivoxel patterns in the hippocampus

Neurons and Cognition 2022-01-28 v1 Machine Learning Neural and Evolutionary Computing

Abstract

The human hippocampus possesses "concept cells", neurons that fire when presented with stimuli belonging to a specific concept, regardless of the modality. Recently, similar concept cells were discovered in a multimodal network called CLIP (Radford et at., 2021). Here, we ask whether CLIP can explain the fMRI activity of the human hippocampus better than a purely visual (or linguistic) model. We extend our analysis to a range of publicly available uni- and multi-modal models. We demonstrate that "multimodality" stands out as a key component when assessing the ability of a network to explain the multivoxel activity in the hippocampus.

Keywords

Cite

@article{arxiv.2201.11517,
  title  = {Multimodal neural networks better explain multivoxel patterns in the hippocampus},
  author = {Bhavin Choksi and Milad Mozafari and Rufin VanRullen and Leila Reddy},
  journal= {arXiv preprint arXiv:2201.11517},
  year   = {2022}
}

Comments

Oral at SVRHM Workshop (NeurIPS 2021)

R2 v1 2026-06-24T09:05:27.842Z