The effect of intensity on an EEG-based objective measure of speech intelligibility

Recently an objective measure of speech intelligibility, based on neural responses, has been developed in normal hearing listeners. However, the population for whom this method is being developed are patients with a hearing loss, requiring overall higher speech intensities to reach a good speech understanding level. As these higher intensities could influence the outcome, we investigated the influence of stimulus intensity on this objective measure. Similar to literature investigating the effect of intensity on cortical responses to non-speech stimuli, we hypothesized that increasing stimulus intensity would (1) increase the amplitudes and (2) decrease the latencies of the peaks in the neural response.

We recorded the electroencephalogram (EEG) in 20 normal-hearing participants while they listened to a narrated story. The story was presented at intensities varying from 10 to 80 dB A. To investigate the brain responses, we analyzed neural tracking of the speech envelope because the speech envelope is known to be essential for speech understanding. Envelope tracking can be measured by reconstructing the envelope from EEG using a linear decoder and by correlating the reconstructed envelope with the actual envelope. We investigated the delta (0.5-4 Hz) and the theta (4-8 Hz) band at each intensity. We also investigated the latencies and spatial components of the responses in more detail using temporal response functions.

Preliminary results show that when presenting the stimuli at higher intensities, response latencies of the peaks between 0 and 250ms shorten, similar to literature using non-speech stimuli. The amplitudes of these peaks, on the other hand, behave opposite to literature, decreasing with increasing stimulus intensity. Despite these latency and amplitude changes as a function of intensity, we can still objectively measure speech intelligibility. These results indicate that the objective measure of speech intelligibility can reliably be used in patients requiring higher stimulus intensities to enhance their speech understanding.

Acknowledgements: Research of Eline Verschueren (1S86118N) and Jonas Vanthornhout (1S10416N) is funded by a PhD grant of the Research Foundation Flanders (FWO). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement No 637424 to Tom Francart). Further support came from KU Leuven Special Research Fund under grant OT/14/119.