Abstract

Speaker: Roxana Zeraati, Max Planck Institute for Biological Cybernetics, Germany.

Title: Model-free and model-based methods for estimating neural timescales

Abstract: Ongoing neural activity unfolds across a wide range of timescales that are suggested to reflect the diversity of behavioral timescales. However, to gain a holistic understanding of how neural timescales relate to behavior and cognitive function we need to develop analysis methods to precisely estimate neural timescales from limited data recorded during cognitive tasks. In this talk, I will first give an overview of neural timescales and various findings relating them to behavior. Then, I will dive into different model-free and model-based methods for estimating timescales. In particular, I will present our recent model-based Bayesian method that estimates timescales using generative models, allowing for unbiased estimation of timescales from short experimental recordings. Such methods based on generative models can be adopted for estimating timescales across different data modalities, by incorporating various mechanistic assumptions and data features.

Bio: Roxana Zeraati has a background in physics and biomedical engineering. She completed her PhD at the University of Tübingen working with Anna Levina and Tatiana Engel. Currently, she is a postdoctoral researcher in Peter Dayan’s lab at the Max Planck Institute for Biological Cybernetics.

Venue: MRC CBU West Wing Seminar Room and Zoom https://us02web.zoom.us/j/82385113580?pwd=RmxIUmphQW9Ud1JBby9nTDQzR0NRdz09 (Meeting ID: 823 8511 3580; Passcode: 299077)