Skip to Content »

 Todd Woodward

Download Seminar Poster PDF

Talk Title:
Constrained Principal Component Analysis for fMRI Data: Relating fMRI Signal to Experimental Conditions.

Thursday, December 9, 2010, 6:00 pm

Assistant Professor, Department of Psychiatry, University of British Columbia

Todd Woodward

Functional magnetic resonance imaging (fMRI) data from one subject
typically involves measurement of blood-oxygen level dependent (BOLD)
changes in the whole brain, measured every two seconds, over a period
of about 12 minutes, at 65,000 brain regions. Over this 12 minute
period, participants are asked to repeatedly carry out some cognitive
task, for example, to remember 8 numbers for a short period of time
(e.g., 4 seconds). Researchers using this technology need to determine
which brain networks are invoked in response to the task (and not some
other aspect of thinking, like “this scanner is uncomfortable!”), but
also (and importantly), which are differentially involved in the
experimental conditions that are built into the task. The method that
our laboratory has developed for this purpose is named constrained
principal component analysis (fMRI-CPCA), and combines multivariate
multiple regression with principal component analysis (i.e., singular
value decomposition). Multivariate multiple regression is used to
separate the BOLD signal variance into what is predictable from the
task and what is not, and singular value decomposition is used to
determine the dominant patterns of intercorrelation between the brain
regions for each source of BOLD signal variance, thereby forming brain
networks. Changes over time in these brain networks can be correlated
with the timing of the task to determine their relative involvement in
the experimental conditions. fMRI-CPCA is available for download free
of charge (

Please note:
Trainees are invited to meet with the VanBUG speaker for open discussion
of both science and career paths. This takes place 4:30-5:30pm in either
the Boardroom or Lunchroom on the ground floor of the BCCRC

Recommended Readings
FIR-CPCA analysis


Introductory Speaker:
Raymond Lim, MSc Candidate, Pavlidis Lab, Centre for High Throughput Biology, University of British Columbia

Wide Scale Comparison of Transcriptome Profiles

Download PresentationPDF