Related Experiment Video
Updated: Jan 16, 2026
06:50
Analyzing Neural Activity and Connectivity Using Intracranial EEG Data with SPM Software
Published on: October 30, 2018
9.9K
Dynamical Modeling of Behaviorally Relevant Spatiotemporal Patterns in Neural Imaging Data.
Mohammad Hosseini1, Maryam M Shanechi1,2,3
1Electrical and Computer Engineering, Viterbi School of Engineering, University of Southern California (USC), Los Angeles, CA, USA.
Arxiv
|October 3, 2025
Summary
We developed SBIND, a deep learning framework for analyzing brain imaging data. SBIND models neural dynamics and behavior, outperforming existing methods in predicting neural-behavioral relationships.
Area of Science:
- Neuroscience
- Computational Neuroscience
- Machine Learning
Background:
- High-dimensional neural imaging (e.g., widefield calcium, functional ultrasound) offers insights into brain-behavior relationships.
- Modeling complex neural dynamics is challenging due to high dimensionality, spatiotemporal dependencies, and irrelevant activity.
- Current models often reduce dimensionality, potentially losing crucial behavior-related information and spatiotemporal structure.
Purpose of the Study:
- To introduce SBIND, a novel deep learning framework for modeling spatiotemporal dependencies in neural images.
- To disentangle behaviorally relevant neural dynamics from other neural activity.
- To validate SBIND on widefield imaging and explore its application to functional ultrasound imaging.
Main Methods:
- SBIND employs a data-driven deep learning approach to model neural image data.
- The framework captures local and long-range spatial dependencies across the brain.
- SBIND is designed to differentiate behaviorally relevant neural dynamics.
Main Results:
- SBIND effectively models spatiotemporal dependencies in neural imaging data.
- The model successfully dissociates behaviorally relevant neural dynamics.
- SBIND demonstrates superior performance in neural-behavioral prediction compared to existing models.
- The framework shows applicability to both widefield calcium and functional ultrasound imaging.
Conclusions:
- SBIND provides a powerful tool for analyzing neural imaging data and understanding brain-behavior links.
- The framework advances the modeling of neural dynamics in high-dimensional imaging modalities.
- SBIND facilitates the investigation of neural mechanisms underlying behavior.
