Information Theoretic Tools for Studying Cortical Circuit Function
New optical tools for probing neural circuit functionality require the development of new signal processing and data analysis algorithms. On the signal processing end, we have developed advanced image segmentation algorithms capable of automatically pulling out regions of interest from calcium imaging movies, and a calcium transient detection framework based on Finite Rate of Innovation theory, which performs extremely well. To analyse the resulting data, we continue to work on information theoretic algorithms for studying multivariate time series, and in particular tools for dissecting out distinct statistical components of the neural code, as well as dimensionality reduction approaches necessary to deal with large-scale neural recording technology.
We have worked extensively on neural coding of sensory information, in systems ranging from the visual and somatosensory cortices, to the lateral geniculate nucleus of the thalamus, to the cerebellum. Current work focuses on quantifying encoding and retrieval of spatial information in the hippocampus., and on information processing by ON and OFF pathways in the mouse dorsal LGN.