[CTN] CTN seminar: Dr. Masami Tatsuno, Jan 24th, PAS 2464, 3:30

[Matthijs van der Meer]

Dear all,

Please join us for next Tuesday's CTN seminar (Jan 24th) by Dr. Masami
Tatsuno (University of Lethbridge, AB). Title and abstract follow below.

Time and place are the usual, 3.30pm on Tuesday in PAS 2464.

If you would like to meet with Dr. Tatsuno, and/or come to dinner
afterwards, please let me know.

Hope to see you all there!

- Matt


Dr. Masami Tatsuno
Assistant Professor
Department of Neuroscience
Department of Physics and Astronomy
iCORE Scholar in Neural Dynamics and Information Processing
Canadian Centre for Behavioural Neuroscience

Title: Information-geometric measure for estimation of connection
strength under correlated input (Yimin Nie and Masami Tatsuno)

Abstract: The brain processes information in a highly parallel manner.
Determining the relationship between neural spikes and synaptic
connections plays a key role in analyzing electrophysiological data.
Information geometry (IG) has been proposed as a powerful analysis tool
for multiple spike data, providing useful insights into the statistical
interactions within a population of neurons. Previous work has
demonstrated that IG measures can be used to infer the connection
strength between two neurons in a neural network. This property is
useful in neuroscience because it provides a way to estimate
learning-induced changes in synaptic strengths from extracellular
neuronal recording. A previous study has also shown, however, that this
property would hold only when inputs to neurons are not correlated.
Since neurons in the brain often receive common inputs, this would
hinder application of the IG to real data. We investigated two-neuron-IG
measures in higher-order log-linear models to overcome this limitation.
First we mathematically showed that the estimation of symmetrically
connected synaptic weight under correlated inputs can be improved by
taking into account higher-order interactions. Second we numerically
showed that the estimation can be also improved for more general
asymmetrically connected networks. Considering the size of practically
available data in an experiment and the number of connections in the
brain, we showed that the two-neuron-IG measure calculated with 4-5
neuronal interactions would provide estimation of connection strength
within approximately 10% accuracy. These studies suggest that the
two-neuron-IG measure with higher-order interactions is a robust
estimator of connection strength, providing a useful analytical tool for
real multi-neuronal spike data.