Tuesday, February 12, 2019

17:15h

The talk is open to the public.
Guests are cordially invited!

Hosted by Stefan Rotter

vCal iCal

Bernstein Center Freiburg
Lecture Hall (ground floor)
Hansastr. 9a
79104 Freiburg

Tell a friend! More seminars

Abstract

Extracellular electrophysiology is one of the most important and widely-used approach to investigate how the brain functions. Recent advances in both instrumentation, for example Multi-Electrode Arrays with high spatial density and channel count, and computational modeling, give us powerful tools to improve standard techniques involved in the electrophysiological pipeline. The latter consists of several steps, starting with spike sorting, neural cell-type classification and localization, and finally electrical stimulation.
In my talk, I will show how each of the aforementioned steps can benefit from novel methods that combine computer simulations and data-driven approaches. Realistic biophysical simulations can be used, for example, to construct recordings of known-ground-truth for spike sorting evaluation and validation, to employ supervised learning methods, such as deep learning, for cell-type recognition and 3D localization, and to optimize stimulation paradigms that increase cell selectivity and accuracy. Finally, I will discuss some limitations of the conventional biophysical modeling scheme and possible solutions involving more sophisticated techniques.

Literature

Alessio P. Buccino, Michael Kordovan, Torbjørn V. Ness, Benjamin Merkt, Philipp D. Häfliger, Marianne Fyhn, Gert Cauwenberghs, Stefan Rotter, and Gaute T. Einevoll (2018). Combining biophysical modeling and deep learning for multielectrode array neuron localization and classification. Journal of Neurophysiology (Link)