Phase-synchrony metrics allow inspection of inter-subject similarity at high temporal resolution

Recent studies have demonstrated that it is possible to use feature films (see Hasson et al. 2010) and music (Alluri et al. 2012) as highly dynamic naturalistic stimuli in functional magnetic resonance imaging studies. This constitutes a highly significant step forward as it enables one to study cognitive functions that would otherwise be difficult to engage under the neuroimaging laboratory conditions such as emotions, social perception and cognition, and perception of higher-order musical features. Analysis of the resulting highly multidimensional neuroimaging data is not trivial and while functional brain activity under naturalistic viewing conditions has been successfully analyzed by calculating inter-subject correlations of hemodynamic data, inspection of temporal dynamics of inter-subject similarity using inter-subject correlation has been challenging as the correlations have to be calculated over sliding time windows of ~10-20 seconds.

Glerean et al. 2012 show that it is possible to increase temporal resolution by using instantaneous phase synchronization rather than inter-subject correlation as the measure of dynamic (time-varying) functional connectivity. In their study, Glerean et al. applied inter-subject phase-synchrony on a functional magnetic resonance imaging dataset obtained while 12 healthy volunteers watched a feature film. In addition, they compared across-subject similarities of phase-synchrony that take place between brain areas (similarly to the widely used seed-voxel correlation method that also suffers from compromised temporal accuracy), denoting this as seed-based inter-subject phase-synchrony.

The findings of Glerean et al. suggest that the tested phase-synchrony metrics yield results that are consistent with both seed-based correlation and inter-subject correlation methods when inspected over the whole duration of the movie, but provide superior (an order of magnitude better) temporal resolution for estimates of how similarly brains of individual subjects are processing the various features and events of the movie. These results thus provide a significant methodological step forward in making it possible to use highly naturalistic stimuli in neuroimaging studies and remarkably broaden the possibilities of cognitive neuroimaging. The matlab algorithms for calculating the phase-synchrony metrics of functional magnetic resonance imaging data are freely downloadable from http://becs.aalto.fi/bml/software.html

References: 

Alluri V, Toiviainen P, Jääskeläinen IP, Glerean E, Sams M, Brattico E.Large-scale brain networks emerge from dynamic processing of musical timbre, key and rhythm. Neuroimage (2012) 59: 3677-3689. http://dx.doi.org/10.1016/j.neuroimage.2011.11.019

Glerean E, Salmi J, Lahnakoski JM, Jääskeläinen IP, Sams M. FMRI phase synchronization as a measure of dynamic functional connectivity. Brain Connectivity (2012) (epublication ahead of print May 4). http://dx.doi.org/10.1089/brain.2011.0068

Hasson U, Malach R, Heeger DJ. Reliability of cortical activity during natural stimulation. Trends in Cognitive Sciences (2010) 14: 40-48. http://dx.doi.org/10.1016/j.tics.2009.10.011