Search bioRxiv⌕ Search

Biology subjects

Mirkovic, B.

Publications and source records attributed to Mirkovic, B..

2 recordsLinked to original sources

Early visual cortex tracks speech envelope in the absence of visual input

Neural entrainment to continuous speech is typically observed within the language network and can be modulated by both low-level acoustic features and high-level meaningful linguistic units (e.g., phonemes, phrases, and sentences). Recent evidence showed that visual cortex may entrain to speech envelope, however its putative role in the hierarchy of speech processing remains unknown. We tested blindfolded participants who listened to semantically meaningful or meaningless stories, either in quiet or embedded in multi-talker babble noise. Entrainment to speech was assessed with forward linear modeling of participants EEG activity. We investigated (1) low-level acoustic effects by contrasting neural tracking of speech presented in quiet or noise and (2) high-level linguistic effects by contrasting neural tracking to meaningful or meaningless stories. Results showed that envelope tracking was enhanced and delayed for speech embedded in noise compared to quiet. When semantic information was missing, entrainment to speech envelope was fastened and reduced. Source modeling revealed that envelope tracking engaged wide neural networks beyond the auditory cortex, including early visual cortex. Surprisingly, while no clear influence of semantic content was found, the magnitude of visual cortex entrainment was affected by low-level features. The decrease of sound SNR-level dampened visual cortex tracking, suggesting an active suppressing mechanism in challenging listening conditions. Altogether, these findings provide further evidence of a functional role of early visual cortex in the entrainment to continuous speech.

neuroscience↗

Impact of Different Acoustic Components on EEG-based Auditory Attention Decoding in Noisy and Reverberant Conditions

Recently, a least-squares-based method has been proposed to decode auditory attention from single-trial EEG recordings for an acoustic scenario with two competing speakers. This method aims at reconstructing the attended speech envelope from the EEG recordings using a trained spatio-temporal filter. While the performance of this method has been mainly studied for noiseless and anechoic acoustic conditions, it is important to fully understand its performance in realistic noisy and reverberant acoustic conditions. In this paper, we investigate auditory attention decoding (AAD) using EEG recordings for different acoustic conditions (anechoic, reverberant, noisy, and reverberant-noisy). In particular, we investigate the impact of different acoustic conditions for AAD filter training and for decoding. In addition, we investigate the influence on the decoding performance of the different acoustic components (i.e. reverberation, background noise and interfering speaker) in the reference signals used for decoding and the training signals used for computing the filters. First, we found that for all considered acoustic conditions it is possible to decode auditory attention with a decoding performance larger than 90%, even when the acoustic conditions for AAD filter training and for decoding are different. Second, when using reference signals affected by reverberation and/or background noise, a comparable decoding performance as when using clean reference signals can be obtained. In contrast, when using reference signals affected by the interfering speaker, the decoding performance significantly decreases. Third, the experimental results indicate that it is even feasible to use training signals affected by reverberation, background noise and/or the interfering speaker for computing the filters.

neuroscience↗