Yannik Bauer

Thesis topic: Robust visual processing of feedforward and feedback signals in the mouse dLGN

Research:
I work as part of the DFG-funded Collaborative Research Center project 'Robust Vision' (SFB1233) with my joint supervisors Prof. Laura Busse (LMU), Prof. Philipp Berens (U Tübingen) and Prof. Thomas Euler (U Tübingen), to investigate the neural basis of robust vision, particularly the role of feedback information in making neural signal processing robust against the extreme variability of naturalistic visual inputs. For instance, what enables biological systems, like us, to recognize visual objects in all sorts of contexts without apparent effort where our most advanced machine vision algorithms, including deep networks, fail?

One interesting question of this project concerns what visual signal transformations are carried out in the dorsolateral geniculate nucleus (dLGN) of the thalamus, the next stage after the retina in the visual processing chain. Far from being a simple relay station, the dLGN may make essential contributions to the reliability, efficiency, and robustness of visual information coding, mediated by effects of feedforward information, local inhibition and cortico-thalamic feedback. The point of our research is not only about biological vision, but also about general principles of neural and machine computation, including the roles of feedback, inhibition, robustness and network stability.

Background:
2014 - 2016: MSc Neural & Behavioural Science, University of Tübingen
2011 - 2014: BA Psychology & Philosophy, University of Oxford

Supervisor: Prof. Dr. Laura Busse