Abstract
Introduction: There is a growing interest in differences in patterns of brain activity across individuals and their relation to cognitive strategies/performance (e.g., Mckay et al., 2003; Vogel et al. 2004). Here, we use multivariate pattern classifiers to quantitatively analyze individual differences in brain activity (electroencephalography, EEG) during a perceptual categorization task (Philiastides et al., 2006). We sought to compare the discriminant information across observers in preparatory, visually evoked and late-post stimulus temporal epochs of the EEG activity. Method: Twenty naive observers perceptually categorized briefly (40 ms) presented images of cars and faces embedded in filtered noise while their EEG activity was recorded from 64 electrodes. Each observer participated in 1000 trials and indicated their decision using a confidence rating. Results: Classifier performance (area under the ROC) identifying an individual from single trial EEG activity during pre-stimulus time intervals (preparatory) was close to chance (7.13%; chance = 5%) but then systematically increased to 65.21% during the time-interval of 200-250 ms post-stimulus presentation (visually evoked epoch). Classifier accuracy then monotonically decreased for late post stimulus intervals ([[gt]]=250 ms). The temporally localized nature of the neural activity differentiating individuals stands in contrast to the neural activity differentiating face vs. car trials which has a similar onset (120–200 ms) but persists until 500 ms post-stimulus. Conclusion: Our findings suggest that the primary source of individual differences in EEG activity during a perceptual categorization task is restricted to visually evoked neural activity.