Event: Neurobiology Laboratory, German Primate Center (DPZ)

Author: Alireza Fathian

Summary: The interaction among brain areas has been a topic of increasing interest. With technological advances in recent years, we are able to record from large numbers of neurons across multiple brain areas; therefore, proper data analysis methods are crucial to gain a more extensive understanding of the higher-level mechanisms of the brain. In this presentation, I presented a recent multi-area dimensionality reduction method called Delayed Latents Across Groups (DLAG) in which the neural activity in two different areas (for example, area A and area B) is separated into three sets of population activity patterns: (1) The within area activity patterns, (2) Across area activity patterns from area A to area B (3) Across area activity patterns from area B to area A. In the end, with the ability to extract smooth and bidirectional activity patterns, DLAG offers unique advantages for characterizing the temporal structure of activity within and across areas, and it is also a powerful tool for exploratory data analysis.