BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Computational Neuroscience Journal Club - Mohsen Sadeghi (CBL)
DTSTART:20160705T150000Z
DTEND:20160705T160000Z
UID:TALK66780@talks.cam.ac.uk
CONTACT:Daniel McNamee
DESCRIPTION:Mohsen Sadeghi will cover:\n\n* Credit assignment in movement-
 dependent reinforcement learning\n* S McDougle\, M Boggess\, M Crossley\, 
 D Parvin\, R Ivry\, J Taylor\n* PNAS (2016)\n* http://www.pnas.org/content
 /113/24/6797.long\n\nWhen a person fails to obtain an expected reward from
  an object in the environment\, they face a credit assignment problem: Did
  the absence of reward reflect an extrinsic property of the environment or
  an intrinsic error in motor execution? To explore this problem\, we modif
 ied a popular decision-making task used in studies of reinforcement learni
 ng\, the two-armed bandit task. We compared a version in which choices wer
 e indicated by key presses\, the standard response in such tasks\, to a ve
 rsion in which the choices were indicated by reaching movements\, which af
 fords execution failures. In the key press condition\, participants exhibi
 ted a strong risk aversion bias\; strikingly\, this bias reversed in the r
 eaching condition. This result can be explained by a reinforcement model w
 herein movement errors influence decision-making\, either by gating reward
  prediction errors or by modifying an implicit representation of motor com
 petence. Two further experiments support the gating hypothesis. First\, we
  used a condition in which we provided visual cues indicative of movement 
 errors but informed the participants that trial outcomes were independent 
 of their actual movements. The main result was replicated\, indicating tha
 t the gating process is independent of participants' explicit sense of con
 trol. Second\, individuals with cerebellar degeneration failed to modulate
  their behavior between the key press and reach conditions\, providing con
 verging evidence of an implicit influence of movement error signals on rei
 nforcement learning. These results provide a mechanistically tractable sol
 ution to the credit assignment problem.
LOCATION:Cambridge University Engineering Department\, CBL\, BE-438 (http:
 //learning.eng.cam.ac.uk/Public/Directions)
END:VEVENT
END:VCALENDAR