Combining Different Tools for EEG Analysis to Study the Distributed Character of Language Processing

Abstract

Recent studies on language processing indicate that language cognition is better understood if assumed to be supported by a distributed intelligent processing system enrolling neurons located all over the cortex, in contrast to reductionism that proposes to localize cognitive functions to specific cortical structures. Here, brain activity was recorded using electroencephalogram while volunteers were listening or reading small texts and had to select pictures that translate meaning of these texts. Several techniques for EEG analysis were used to show this distributed character of neuronal enrollment associated with the comprehension of oral and written descriptive texts. Low Resolution Tomography identified the many different sets (si) of neurons activated in several distinct cortical areas by text understanding. Linear correlation was used to calculate the information H(ei) provided by each electrode of the 10/20 system about the identified si. H(ei) Principal Component Analysis (PCA) was used to study the temporal and spatial activation of these sources si. This analysis evidenced 4 different patterns of H(ei) covariation that are generated by neurons located at different cortical locations. These results clearly show that the distributed character of language processing is clearly evidenced by combining available EEG technologies.