Seminar: Learning Algorithms in Neocortex-Cerebellum Circuits

Event Details
  • Date/Time:
    • Tuesday January 19, 2021
      11:00 am - 12:00 pm
  • Location: Via Zoom
  • Phone:
  • URL: Join Zoom Meeting
  • Email:
  • Fee(s):
    N/A
  • Extras:
Contact

Questions? Contact Christy Kelley

Summaries

Summary Sentence: Presented by Mark Wagner, postdoctoral fellow at Stanford University

Full Summary: No summary paragraph submitted.

PRESENTED BY
Mark Wagner, Ph.D.

Postdoctoral Fellow
Stanford University

 

ABSTRACT
In striking contrast to the rest of the brain, evolutionary expansion of the cerebellum has kept pace with that of neocortex, and these two structures contain ~99% of all neurons in humans. The entire neocortex and the cerebellum are also reciprocally connected by some of the densest long-range projections in the brain that are universal across mammals. My work aims to crack the underlying cortex-cerebellum algorithm, by designing novel strategies for multi-site two-photon calcium imaging and optogenetics, viral-genetic single-cell connectivity mapping, and computational circuit modeling. By developing two-photon calcium imaging in cerebellar granule cells during operant behavior, I unexpectedly found reward-related neural activity demonstrating a novel cerebellar cognitive function. More recently, I devised simultaneous two photon calcium imaging of both neocortical output and downstream cerebellar input during weeks-long skill learning, revealing the development of cooperative cortico-cerebellar dynamics. My related studies of the cerebellar climbing fiber system showed that behavioral learning drives the emergence of large-scale neural synchronization transitions during task performance. My ongoing neuroengineering research employs several interconnected approaches: deciphering broad multi-area neural circuit computations; dissecting distributed synaptic plasticity processes during learning; and developing computational models of circuit function informed by novel synaptic connectivity maps. Understanding learning algorithms in biological circuits has the potential for major impact in machine learning, whose neural net architectures are inspired by features of neocortical circuits.

 

BIOGRAPHY
My undergraduate and masters bioengineering work at Harvard University with Maurice Smith used control theory approaches to understand human motor learning. When I came out to Stanford University for my Ph.D., I worked in Mark Schnitzer’s lab in Applied Physics. I adapted the precision motor behaviors I worked with in humans for mice, while learning to incorporate cutting-edge in vivo two photon microscopy tools for large-scale neural activity monitoring and perturbation in learning and behaving animals. I’ve done my postdoctoral studies with Liqun Luo in the Biology department at Stanford, incorporating novel viral-genetic circuit mapping techniques into my broader efforts to decipher neural computation in learning animals.

Additional Information

In Campus Calendar
No
Groups

Wallace H. Coulter Dept. of Biomedical Engineering

Invited Audience
Faculty/Staff, Postdoc, Public, Graduate students, Undergraduate students
Categories
No categories were selected.
Keywords
No keywords were submitted.
Status
  • Created By: Joshua Stewart
  • Workflow Status: Published
  • Created On: Jan 12, 2021 - 9:07am
  • Last Updated: Jan 12, 2021 - 9:16am