BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Population Inference for Functional Brain Connectivity - Genevera 
 I. Allen (Rice University)
DTSTART:20150304T110000Z
DTEND:20150304T120000Z
UID:TALK58206@talks.cam.ac.uk
CONTACT:Dr R.E. Turner
DESCRIPTION:Functional brain connections are the set of statistical relati
 onships between neural activity in different parts of the brain\; these ar
 e typically estimated from neuroimaging data such as functional MRI. In mu
 lti-subject studies\, many are interested in identifying the individual fu
 nctional brain connections or patterns of connections that are different b
 etween two groups of subjects or across a clinical population.  Popular ap
 proaches to this problem include estimating a network for each subject\, a
 nd then assuming the subject networks are fixed\, conducting inference ove
 r network metrics.  These approaches\, however\, fail to account for the v
 ariability and network estimation error associated with estimating each su
 bject’s brain network\, thus resulting in incorrect inferences.  \n\nIn 
 this talk\, we study this problem using Markov Networks as the model for b
 rain connectivity.  Statistically\, our problem can be described as conduc
 ting large scale inference over network edges or groups of edges post grap
 hical model selection\, part of a novel statistical paradigm we call Popul
 ation Post Selection Inference (popPSI).  We show that for this popPSI pro
 blem\, current approaches in neuroimaging have both low statistical power 
 and highly inflated false positive rates.  We then develop a new procedure
  which we term R^3\, standing for resampling\, random effects\, and random
  penalization.  Our approach uses the correct two-level random effects mod
 el to account for network variability within a subject (due to network est
 imation) as well as between subjects.  Through simulation studies we show 
 that our method solves many of the problems associated with existing techn
 iques\, yielding substantial improvements in terms of both error control a
 nd statistical power. We conclude our talk by applying our methods in two 
 case studies - a color-sequence synesthesia study and a neurofibromatosis 
 one study.\n\nJoint work with Manjari Narayan and Steffie Tomson.
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR