Institute of Cognitive Science,
49090 Osnabrück, Germany
Cortical spike synchrony
as a measure of contour uniformity
My background includes both neuroscience and machine learning engineering. I am interested in combining these two, developing models of brain dynamics which can be biologically plausible and computationally efficient at the same time.
One example of such exciting combination is neuromorphic computing. Neuromorphic chips are designed to work efficiently with spiking neural networks, which are likened to real brain neuronal populations. I am especially interested in modeling mechanisms of brain plasticity on such neuromorphic hardware.
I am currently involved in two interconnected projects. The goal of the first project is to develop a spiking model of V1 brain area, which can demonstrate synchronous neuronal activity in response to visual stimuli with specific geometrical properties. These properties are: visual continuity (e.g. a continuous contour) and orientation similarity (e.g. an image consisting of lines of the same angle). In the brain such synchrony happens due to horizontal connections between neurons. The crucial component of the model is also the connection matrix between model units, which we define before running the model.
In the second project the model will be transferred to neuromorphic hardware, and the connection matrix won’t be predefined, but rather learned via the plastic mechanisms between model units.
Zemliak V and MacInnes WJ (2022) The Spatial Leaky Competing Accumulator Model. Front. Comput. Sci. 4:866029. doi: 10.3389/fcomp.2022.866029
Paper (not RTG related)