Nonlinearities in bipolar cells and their role for encoding visual signals

von Helene Marianne Schreyer
Dissertation
Datum der mündl. Prüfung:2018-05-08
Erschienen:2019-05-02
Betreuer:Prof. Dr. Tim Gollisch
Gutachter:Prof. Dr. Jochen Staiger
Gutachter:Dr. Jeong Seop Rhee
Gutachter:Dr. Marion Silies
Gutachter:Prof. Dr. Tobias Moser
Gutachter:Prof. Dr. André Fiala
Zum Verlinken/Zitieren: http://dx.doi.org/10.53846/goediss-7423

 

 

Dateien

Name:HeleneSchreyer_PhDThesis_web.pdf
Size:8.52Mb
Format:PDF

Lizenzbestimmungen:

Zusammenfassung

Englisch

Vision begins in the retina, where ganglion cells separate the visual input into ~30 parallel output channels with different response characteristics to visual stimuli. How retinal ganglion cells obtain such a diversity of functional properties is unclear. The diversity appears to evolve along the signal processing stream from photoreceptors to ganglion cells. Along this pathway, bipolar cells represent pivotal elements by connecting the photoreceptors, the horizontal cells and the amacrine cells to the retinal ganglion cells. Despite their crucial position, our knowledge about bipolar cells is limited. Furthermore, simplifying assumptions about their light responses are made. For example, it is broadly assumed that bipolar cells respond to light linearly. In this thesis, we investigated the assumption of linear signal processing in bipolar cells. To do so, we worked on four main goals: Goal 1: Establishing a general characterization of bipolar cells; Goal 2: Assessing nonlinearities in bipolar cells; Goal 3: Predicting bipolar cells’ responses with the linear-nonlinear model and Goal 4: Simultaneous recordings from bipolar and ganglion cells. We investigated the goals in the salamander retina by recording the voltage signals of bipolar cells with single electrodes. We observed a diversity of bipolar cell responses to simple and complex light stimuli (goal 1). We observed nonlinear responses of bipolar cells in their contrast representation and in their input integration (goal 2). Further, mathematical models like the linear-nonlinear model failed to predict responses of some bipolar cells to complex artificial and natural light stimuli (goal 3). Finally, the established method of simultaneous recordings from bipolar and ganglion cells was used to study the connection between bipolar cells and mathematically retrieved subunits in ganglion cells (goal 4). Taken together, our work suggests that nonlinear signal transformation starts at the level of the input integration in bipolar cells and that the bipolar cell nonlinearities have to be taken into consideration for mathematical encoding models in the retina.
Keywords: retina; bipolar cells; output nonlinearity; spatial nonlinearity; encoding models; natural stimuli
 
Statistik