Recent studies with magnetoencephalography (MEG) have shown that the human auditory cortex is particularly sensitive to reduction in sound and speech quality. In this paper, we examine whether this sensitivity is also visible in the electroencephalogram (EEG) and whether it is possible to detect subconscious processes which can then be used to improve the behavioral assessment of speech quality. In order to compare the physiological results with behavioral measurements, we degraded a speech stimulus (vowel /a/) in a scalable way and asked for a pair comparison (PC) and a comparison category rating (CCR) of the degraded stimuli. In addition, the brain activity of eleven healthy subjects was measured with EEG, focusing on event-related potentials (ERPs). We found that the threshold as set by the Modulated Noise Reference Unit (MNRU) for the PC and CCR are on a similar signal-to-noise level. We trained classifiers, which were found capable of distinguishing between events which are seemingly similar at the behavioral level (i.e., no button press). Converging evidence suggests that the classifier results could reflect subconscious cortical sensitivity to sound degradations.
Cite as: Antons, J.-N., Schleicher, R., Wolf, I., Porbadnigk, A.K., Blankertz, B., Möller, S., Curio, G. (2010) Neural correlates of speech degradation – Subjective ratings and brain activation in case of signal-correlated noise. Proc. 3rd International Workshop on Perceptual Quality of Systems (PQS 2010), 104-107, doi: 10.21437/PQS.2010-19
@inproceedings{antons10_pqs,
author={Jan-Niklas Antons and Robert Schleicher and Ingo Wolf and Anne K. Porbadnigk and Benjamin Blankertz and Sebastian Möller and Gabriel Curio},
title={{Neural correlates of speech degradation – Subjective ratings and brain activation in case of signal-correlated noise}},
year=2010,
booktitle={Proc. 3rd International Workshop on Perceptual Quality of Systems (PQS 2010)},
pages={104--107},
doi={10.21437/PQS.2010-19}
}