Modeling dynamics of the lateral geniculate nucleus of the thalamus
The proposed project aims to advance computational neuroscience by enhancing the MouseNet model, a Convolutional Neural Network (CNN) designed to emulate the structure and function of the mouse visual cortex. By incorporating 3D convolutions and realistic temporal dynamics, the project seeks to model the lateral geniculate nucleus (LGN) dynamics more accurately. This enhancement will provide deeper insights into the mechanisms governing visual processing, contributing to our understanding of neural computation.
The expected outcome of the project is a refined MouseNet model that better captures the dynamics of LGN activity, thus improving accuracy within the training of specific visual tasks, particularly those involving motion. Participating institutions stand to benefit from improved computational models that better capture biological processes, facilitating research in various fields such as artificial intelligence, especially computer vision and neuroscience.
Voir la description complète du projetBryan Tripp
Ukrainian Catholic University
Computer science
Artificial Intelligence; Health and Related Sciences & Technology
University of Waterloo
Globalink Research Award