BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Computational Neuroscience Journal Club - Greg Sotiropoulos (CBL)
DTSTART:20161108T160000Z
DTEND:20161108T170000Z
UID:TALK69031@talks.cam.ac.uk
CONTACT:Daniel McNamee
DESCRIPTION:Greg Sotiropoulos will cover:\n\n* Inferring the brain's inter
 nal model from sensory responses in a probabilistic inference framework\n*
  Richard D Lange\, Ralf M Haefner\n* bioXriv (October 2016)\n* http://bior
 xiv.org/content/early/2016/10/19/081661.full.pdf\n\nABSTRACT:\nDuring perc
 eption\, the brain combines information received from its senses with prio
 r information about the outside world. The mathematical concept of probabi
 listic inference has previously been suggested as a framework for understa
 nding both perception and cognition. Whether this framework can explain no
 t only behavior but also the underlying neural computations has been an op
 en question. We propose that sensory neurons' activity represents a centra
 l quantity of Bayesian computations: posterior beliefs about the outside w
 orld. As a result\, sensory responses\, just like the beliefs themselves\,
  should depend both on sensory inputs and on prior information represented
  in other parts of the brain. We show that this dependence on internal var
 iables induces variability in sensory responses that -- in the context of 
 a psychophysical task -- is related both to the structure of that task and
  to the neurons' stimulus tuning. We derive analytical predictions for the
  correlation between different neurons' responses\, and for their correlat
 ion with behavior. Furthermore\, we show that key neurophysiological obser
 vations from much studied perceptual discrimination and detection experime
 nts agree with those predictions. Our work thereby provides a normative ex
 planation for those observations\, requiring a reinterpretation of the rol
 e of correlated variability for sensory coding. Finally\, the fact that se
 nsory responses (which we observe) are a product both of external inputs (
 which we control) and of internal beliefs\, allows us to reverse-engineer 
 information about the subject's internal beliefs by observing sensory neur
 ons' responses alone. Population recordings of sensory neurons in animals 
 performing a task can therefore be used to track changes in the internal b
 eliefs with learning and attention.
LOCATION:Cambridge University Engineering Department\, CBL\, BE-438 (http:
 //learning.eng.cam.ac.uk/Public/Directions)
END:VEVENT
END:VCALENDAR