Abstract

Introduction Activation of neuronal networks is typically associated with a change in spectral power in electric signals within the network. The role of different spectral bands in mediating cognitive processes is still not fully understood. In human language networks, a hierarchical model has been suggested in which different types of linguistic information is transferred in different oscillatory bands. Within that model, attention is assumed to correlate with increase in gamma and decrease in alpha band power. Electrocorticography (ECoG) allows the segregation of neural representations independently of stimulus variation. Aim of the study Here, for the first time, we investigate whether selective attention to the different components of a syllable (i.e., consonants or vowels) modulates the spectral power of oscillatory activity (i.e., broadband gamma and alpha) and whether these changes relate to behavioral performance. Methods To address this issue, we measured ECoG signals in 11 patients with refractory epilepsy who were undergoing pre-surgical screening. Neuronal responses were measured with intracranial electrodes (i.e., grids, strips, and depth hippocampal electrodes). The patients engaged in cognitive testing in which auditory consonant–vowel syllable sounds were compared with written letters (i.e., delay match-to-sample task). The behavioral relevance of different components of the syllables was manipulated in four levels; attention was directed either to the complete syllable, only to the vowel, only to the consonant, or the syllable sound was irrelevant. Analysis We calculated: (a) bipolar differentials between each pair of neighboring channels; (b) time–frequency representations of power between 2 and 150 Hz; (c) an index of power change, per trial, relative to power during pre-stimulus baseline period; and (d) average power modulation for two frequency bands (i.e., 5–15 and 70–150 Hz). Results Presentation of stimuli was associated with an increase in broadband gamma (70–150 Hz) and a decrease in alpha (5–15 Hz) over sensory cortex (i.e., auditory cortex for auditory stimuli, and visual cortex for visual stimuli), while alpha power was increased in domain unspecific cortical areas. Gamma band activity could be robustly mapped to stimulus parameters but was only weakly related to attentional state or behavioral performance. Alpha band activity was more related to cognitive state and performance than to stimulus features. Broadband gamma power changes were related to stimulus physical properties. Conclusion These results give new insights into the interplay between alpha and gamma oscillations within the human language network, here with a focus on phonological aspects, and the possible role of these oscillations in the processing of sensory stimuli and behavioral performance during selective attention.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call