ICS, CNE, HCR Research Seminar on "Learning about neural computation from sparse recordings "
ICS;CNE, HCR Research Seminar on "Learning about neural computation from sparse recordings" by Marcel Nonnenmacher Department of Electrical Engineering at TUM, within the Computational Neuroengineering (CNE) group
Abstract: In this talk, I will present our research on the effects of subsampling of neural population activity on statistical modelling and on the outcomes of subsequent analyses. The last decade brought great progress in experimental techniques for simultaneously recording the activity from large populations of neurons. This led to new analysis methods that aim to better understand these high-dimensional time-varying signals. Yet few models and analyses specifically address the subsampling of activity still inherent to these recordings: Due to technical limitations on the field of view and sampling frequencies, modern techniques in most cases still cannot simultaneously record the large neural populations that are involved in representing and processing incoming stimuli. In a first line of work, we investigated effects of subsampling on a series of recent studies on population coding. Here, we showed that subsampling effects can lead to false positives in a proposed test for 'critical' behaviour of neural activity. In a second line of work, we develop algorithms to fit statistical models of population dynamics to incomplete data recorded from multiple sub-populations one after another. This allows to assemble a full picture of the population dynamics by 'stitching' together multiple subsampled recordings.
Bio: Marcel Nonnenmacher is a Ph.D. candidate at the Department of Electrical Engineering at TUM, within the Computational Neuroengineering (CNE) group. His research is supervised by Prof. Jakob Macke. His research interests include neural population coding and statistical modelling of neural dynamics. He received his Master of Science from the University of Tuebingen, and Bachelor of Science degree from the University of Osnabrueck.
Date: 24. October 2018, 16:30 h Place: ICS Karlstraße 45, 6th floor room 6009