BME PhD Preliminary Exam Announcement for Shulan Xiao (K. Jayant, advisor)

Everyone is invited to attend the public presentation beginning at 11:00 am.

TITLE: THE BIOPHYSICAL BASIS OF SENSORY FEATURE BINDING IN CORTICAL PYRAMIDAL NEURONS

Date: May 19^th, 2023

Time: 11:00 am

Location: MJIS 2001 and Zoom https://purdue-edu.zoom.us/j/9146700686

Advisory committee: Krishna Jayant, Chair; Alexander A. Chubykin; Edward L. Bartlett; Fang Huang; Scott R. Pluta

Abstract:
Feature binding is critical for sensory perception and pattern classification. In cortical layer 2/3 and layer 5 pyramidal neurons (L2/3 and L5 PNs) feedforward inputs synapse onto basal dendritic arbors in distinct spatio-temporal patterns. How these distributed inputs integrate, multiplex, and dictate unique output codes, however, remains poorly mapped. Furthermore, it remains unclear how specific input sequences learn over time to aid pattern classification and whether this plasticity and integration is shaped by top-down feedback. To address these outstanding challenges, I have designed and implemented a new microscope capable of stimulating synapses via 3D transmitter uncaging while simultaneously recording the electrical input-output transformation. Using this platform across in vitro acute slice preparations, I propose to unravel the biophysics of multibranch dendritic integration and feature classification under quiescent and in vivo like states in L5 pyramidal neurons – the primary output layer of the cortex. Here, my aims are focused on dissecting mechanisms that underlie multiplexing across basal dendrites, plasticity mechanisms that assist efficient feature classification, and the role of top-down modulation in shaping basal dendritic integration. I propose innovative methods including two-photon holographic uncaging, dynamic clamp, computational modeling, and in vivo two-photon calcium imaging to test and validate my hypotheses. Successful completion of my aims will lead to a new understanding of how dendritic nonlinearities facilitate efficient feature classification amidst noisy backgrounds in individual L5 PNs, including different subtypes. The outcome of this finding is not only poised to improve our understanding of neural signal processing but could bolster the design of future biologically inspired artificial intelligence-based neuromorphic systems.