CCNCS Seminar Details
Multivariate Pattern Analysis in Neuroimaging: fMRI, EEG and MEG [Joint with CBMI]
|Speaker:||Dr Li Su|
|Date/Time:||Wednesday 16 March 2011, 4.00pm|
|Location:||SW101 (the Brian Spratt Room) in Computing|
Multivariate Pattern Analysis (MVPA) has been successfully applied to fMRI (e.g. Haxby et al., 2001; Kriegeskorte et al., 2006; Haynes and Rees, 2006) and (quasi) time series data (Mourão-Miranda et al., 2007). The particular approach of Representational Similarity Analysis (RSA; Kriegeskorte et al., 2008a) has also demonstrated potential to integrate neuroscientific data from different modalities, experimental designs, or even species. For instance, RSA has been used to relate cell-recording from monkey Inferior Temporal cortex (IT) with the blood-oxygen-level dependent responses from human IT (Kriegeskorte et al., 2008b). Unlike mass-univariate approaches such as SPM (Friston et al., 1995), RSA is based on the pattern-information that is naturally embedded in multi-channel recording of neural activations. Despite some previous endeavours, application of MVPA in time series data is still very limited. Here, we present both well established methods for applying MVPA in particular RSA to fMRI data and our recent developments of applying RSA to MEG/EEG data (Su et al., 2010).