Neural representation of spatial and non-spatial auditory attention in EEG signals

Abstract

Neural representation, capturing the content and format of encoded information, provides insight into the internal states of neural units. Studies of neural representation contrast with studies of neural processes, which focus on how one neural unit influences another. Representational similarity analysis (RSA), a multivariate analysis approach, has been used in previous studies to explore the neural representation of object categories in various neuroimaging modalities. In this study, we employed RSA to examine the neural representation of executive function. We designed an experiment involving a rich set of conditions where participants engaged in an auditory task requiring either spatial or non-spatial attention. We extracted representational features from their electroencephalography (EEG) scalp voltage and alpha power and compared these features with ideal conceptual models representing perfect categorization of different attentional states. The results demonstrate the feasibility of investigating internal cognitive states using RSA. Specifically, we identified time intervals during which attentional state contrasts, such as differences between attention types or locations, manifested in the measured neural responses from scalp voltage and alpha power distributions.