Optophysiological analysis of pattern classification strategies in the zebrafish olfactory bulb.
PhD Thesis, University of Basel,
Faculty of Science.
Official URL: http://edoc.unibas.ch/diss/DissB_9779
Classification of overlapping activity patterns is a common problem for sensory systems. For a robust representation of sensory stimuli, neuronal circuits should generalize over input patterns reflecting variations of the same stimulus but separate patterns representing different stimuli, even if they are highly overlapping. In the olfactory bulb (OB), input patterns evoked by related, yet distinct, odors are often highly overlapping but perceived as a different stimulus. Consistent with this observation, overlapping input patterns evoked by different odors become decorrelated at the level of OB output neurons, the mitral cells. Input patterns evoked by different concentrations of the same odor, on the other hand, can be quite different but are usually perceived as the same odor quality. To obtain insights into the principles underlying pattern classification in the OB, I measured activity patterns evoked by different stimulus concentrations across mitral cells using temporally deconvolved 2-photon calcium imaging in zebrafish. In addition, I morphed one stimulus into another by continuously changing their ratio in a binary mixture to examine the transition between representations of similar and non-similar stimuli. The intensity, time course and distribution of mitral cell response patterns changed substantially when odor concentration was varied. Nevertheless, response patterns across the population of mitral cells maintained a high correlation over a substantial concentration range. OB output activity patterns therefore retain concentration-invariant components and generalize over different odor concentrations. Morphing of similar and dissimilar stimuli, in contrast, resulted in abrupt transitions between two or three discrete representations, respectively. These transitions were not caused by shifts in the global network state but mediated by coordinated response changes among small neuronal ensembles. These results indicate that the olfactory bulb classifies input patterns into distinct output patterns in an attractor-like fashion.
|Committee Members:||Lüthi, Andreas|
|Faculties and Departments:||09 Associated Institutions > Friedrich Miescher Institut FMI|
|Bibsysno:||Link to catalogue|
|Number of Pages:||69 S.|
|Last Modified:||30 Jun 2016 10:48|
|Deposited On:||27 Mar 2012 14:03|
Repository Staff Only: item control page