Abstract
Pattern classification techniques have been widely used to differentiate neural activity associated with different perceptual, attentional, or cognitive states, especially using fMRI. Here, we endeavored to predict the scope of attention, local or global, from EEG data. A group of 15 participants were presented with Navon-type hierarchical stimuli and were asked to indicate which of two target letters (H or S) was presented, irrespective of whether the target appeared at the local or global level. On each trial, the target letter appeared equiprobably at each level, paired with an irrelevant distracter (E or A) at the other level. EEG data were recorded from 64 channels at a sampling rate of 1024 Hz, and were subsequently high-pass filtered at 0.1 Hz. We applied linear support vector machine (SVM) classifiers with a cross-validation procedure at each time-point (~1 ms). EEG data were reliably differentiated for local vs. global attention on a trial-by-trial basis, emerging as a specific spatiotemporal activation pattern over posterior electrode sites during the 300-700 ms interval after stimulus onset. Whereas the classifiers reliably differentiated the scope of attention, a more traditional event-related potential (ERP) analysis of the same data did not. In sum, multivariate pattern analysis of EEG, which reveals unique spatiotemporal patterns of electrophysiological activity distinguishing between behavioral states, is a highly sensitive tool for characterizing the neural correlates of attention.
Meeting abstract presented at VSS 2013