Biologically inspired CNN for ganglion cell response prediction

Design and analyze a bio-inspired CNN for the retina


We train convolutional neural networks to predict responses of retinal ganglion cells and then further use them to gain insights into the retinal circuitry. However, so far we mostly build generic CNNs that do not reflect the structure and connectivity of a biological retina.


The focus of this project is to develop a CNN that is aligned with our knowledge of retinal architecture. The key aspect will be the integration of an inhibitory layer via weight and connectivity constraints, designed to reflect the inhibitory layer in the retina.


Within the context of this project, several research questions could be explored: - Can we construct a bio-inspired neural network capable of predicting neural responses? - What are the differences between a biologically inspired network and a generic CNN? - What role does inhibition play in the biologically inspired network?


To apply please email Michaela Vystrčilová stating your interest in this research and detailing your relevant skills.

Neural Data Science Group
Institute of Computer Science
University of Goettingen