[2603.03343] Neuro-Symbolic Decoding of Neural Activity

Quantitative Biology > Neurons and Cognition arXiv:2603.03343 (q-bio) [Submitted on 22 Feb 2026] Title:Neuro-Symbolic Decoding of Neural Activity Authors:Yanchen Wang, Joy Hsu, Ehsan Adeli, Jiajun Wu View a PDF of the paper titled Neuro-Symbolic Decoding of Neural Activity, by Yanchen Wang and 3 other authors View PDF HTML (experimental) Abstract:We propose NEURONA, a neuro-symbolic framework for fMRI decoding and concept grounding in neural activity. Leveraging image- and video-based fMRI question-answering datasets, NEURONA learns to decode interacting concepts from visual stimuli based on patterns of fMRI responses, integrating symbolic reasoning and compositional execution with fMRI grounding across brain regions. We demonstrate that incorporating structural priors (e.g., compositional predicate-argument dependencies between concepts) into the decoding process significantly improves both decoding accuracy over precise queries, and notably, generalization to unseen queries at test time. With NEURONA, we highlight neuro-symbolic frameworks as promising tools for understanding neural activity. Comments: Subjects: Neurons and Cognition (q-bio.NC); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) Cite as: arXiv:2603.03343 [q-bio.NC]   (or arXiv:2603.03343v1 [q-bio.NC] for this version)   https://doi.org/10.48550/arXiv.2603.03343 Focus to learn more arXiv-issued DOI via DataCite (pending registration) Submission history From: Joy Hsu [view email] [v1] Sun, 22 Feb 20...