BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:BMI Seminar // Tim Gentner - Intrinsic Geometry of a Combinatorial
  Sensory Neural Code for Birdsong
DTSTART:20221109T160000
DTEND:20221109T170000
DTSTAMP:20260428T022452Z
UID:10488bca67ba78011fa114bb79bff96fa8a6e2b3d64529af9b69ca75
CATEGORIES:Conferences - Seminars
DESCRIPTION:Tim Gentner\, University of California\, San Diego\, USA\nUnde
 rstanding the nature of neural representation is a central challenge of ne
 uroscience. One common approach to this challenge is to compute receptive 
 fields by correlating neural activity with external variables drawn from s
 ensory signals. But these receptive fields are only meaningful to the expe
 rimenter\, not the organism\, because only the experimenter has access to 
 both the neural activity and knowledge of the external variables. To under
 stand neural representation more directly\, recent methodological advances
  have sought to capture the intrinsic geometry of sensory driven neural re
 sponses without external reference. To date\, this approach has largely be
 en restricted to low-dimensional stimuli as in spatial navigation. In this
  talk\, I will discuss recent work from my lab examining the intrinsic geo
 metry of sensory representations in a model vocal communication system\, s
 ongbirds. From the assumption that sensory systems capture invariant relat
 ionships among stimulus features\, we conceptualized the space of natural 
 birdsongs to lie on the surface of an n-dimensional hypersphere. We comput
 ed composite receptive field models for large populations of simultaneousl
 y recorded single neurons in the auditory forebrain and show that solution
 s to these models define convex regions of response probability in the sph
 erical stimulus space. We then define a combinatorial code over the set of
  receptive fields\, realized in the moment-to-moment spiking and non-spiki
 ng patterns across the population\, and show that this code can be used to
  reconstruct high-fidelity spectrographic representations of natural songs
  from evoked neural responses. Notably\, we find that topological relation
 ships among combinatorial codewords directly mirror acoustic relationships
  among songs in the spherical stimulus space. That is\, the time-varying p
 attern of co-activity across the neural population expresses an intrinsic 
 representational geometry that mirrors the natural\, extrinsic stimulus sp
 ace.  Combinatorial patterns across this intrinsic space directly represe
 nt complex vocal communication signals\, do not require computation of rec
 eptive fields\, and are in a form\, spike time coincidences\, amenable to 
 biophysical mechanisms of neural information propagation.\n 
LOCATION:SV 1717 https://plan.epfl.ch/?room==SV%201717 https://epfl.zoom.u
 s/j/62550191552?pwd=OFRDVzBHb0cyVG1wYkxUSFFEMUNPdz09
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR