BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Perceiving and Remembering Objects and Faces - Viljami Salmela\, A
 cademy Research Fellow\, University of Helsinki
DTSTART:20161025T130000Z
DTEND:20161025T140000Z
UID:TALK68710@talks.cam.ac.uk
CONTACT:Paul Bays
DESCRIPTION:_Neural representations underlying perception and memory of lo
 w-level\nvisual features\, such as color and orientation\, are quite well\
 nunderstood. Much less is known about how the primary visual features\nare
  integrated into objects and whether similar principles hold for\nobject m
 emory than for memory of primary features. In the first part\nof the talk\
 , I present studies of shape perception and memory. One\nsuggestion in the
  literature is that shapes are represented by the\nmeans of their radial f
 requency (RF) components. Using fMRI and\nrepresentational similarity anal
 ysis (RSA) we showed RF specific\nresponses patterns in the intermediate v
 isual areas\, supporting the\nrole RF analysis in shape perception. In ano
 ther set of studies\, we\nmeasured the effect of RF shape amplitude and co
 mponent phase\ndifference on memory performance. For amplitude modulated s
 hapes\,\nmemory precision gradually declined as the number of shapes\nto-b
 e-remembered was increased. For two-component and phase modulated\nshapes\
 , the memory performance dropped already for two items. These\nresults sug
 gest similar properties for memory of shapes and primary\nfeatures\, but h
 ighlight the role of the behavioral task on memory\nperformance. In the se
 cond part of the talk\, I present our recent\nstudy on face memory. We con
 tinuously varied facial expressions of\nfive basic emotions and measured m
 emory precision as a function of\nmemory load. Happy expressions were reme
 mbered most precisely\, then\ndisgusted\, fearful\, angry and sad expressi
 ons. However\, for each\nemotional expression\, memory performance decline
 d similarly as a\nfunction of memory load. Memory performance correlated m
 ore with\nemotion discrimination than identification. Furthermore\, precis
 ion for\nfaces did not decline when memory was loaded with gratings\, sugg
 esting\nseparate memory systems for faces and features. In the third part 
 of\nthe talk\, I present some ongoing work and future plans. We recently\n
 applied RSA to combine EEG and fMRI data in order to simultaneously\nhave 
 both high spatial and temporal resolution. I briefly describe the\nmethod 
 and how we are going to apply it to face perception and memory._
LOCATION:Nick Mackintosh Seminar Room\, second floor\, Department of Psych
 ology\, Downing Site\, Cambridge
END:VEVENT
END:VCALENDAR