Ivan Zea Armenta: Decoding population activity in the early visual system of the mouse

BCCN Berlin / Technische Universität Berlin

Abstract

 

The mice visual pathway has been thoroughly studied and described using electrophysiology. In recent years, high density recording electrodes have allowed to record hundreds of neurons simultaneously from different brain regions. Instead of analyzing neuronal data from single brain regions, the simultaneous recordings allow us to study the visual system as a whole. In the current thesis, we analyzed the activity from such simultaneous recordings. The recordings analyzed were obtained mainly from the early visual system (retinal ganglion cells and superior colliculus) but we also had data from cortical regions (primary visual cortex and latero medial area). In this analysis, we defined and quantified characteristic activity patterns that elucidate visual functions. Although some of these defining features were analyzed with classical methods, we opted to use a decoding model to use as a proxy to the brain’s behaviour. Using an interpretable linear model, we described the similarities and differences between the different brain regions recorded. In particular, we were able to observe the differences that arise from the pathway's hierarchical structure. We observed how the time representation of a visual stimulus diluted with increasing depth in the hierarchy, and we also supported a population level representation of the visual stimuli.

 

Additional Information

Master Thesis Defense

 

Organized by

Prof. Dr. Henning Sprekeler   & Dr. Jens Kremkow   / Lisa Velenosi