Test–Retest Reliability of TRF-Derived Measures of Cortical Tracking of the Speech Envelope

Cortical tracking of the speech envelope is an emerging, noninvasive measure of neurophysiological processing of speech that is being widely adopted. It demonstrates good ecological validity, as it allows researchers to study human processing of continuous, naturalistic speech containing dynamic spectrotemporal variations and rich linguistic content. While measures of cortical tracking have strong clinical and research applications, there is a lack of research documenting the reliability of these measures, including how they are affected by the stimulus and how the stimulus is represented, as well as electroencephalography (EEG) acquisition and analysis parameters. In this study, we measured the test–retest reliability of cortical tracking of the speech envelope across different stimuli (an audiobook vs a podcast), stimulus features (broadband envelope and its derivative, multiband envelope and its derivative), reference electrodes (average mastoid vs common average reference), and EEG frequency bands (delta, theta, full) in 26 neurotypical adults (20 female) to assess the potential for cortical tracking to serve as a reliable measure of speech processing. We also examined the amount of data required to achieve stable reliability estimates. We observed moderate to good test–retest reliability for most parameters with as little as 390 s of data, supporting the utility of cortical tracking of the speech envelope as a reliable tool for assessing speech processing. The findings of this study will allow researchers to more effectively design and implement studies on cortical tracking in neurotypical adults and adults with language disorders. Significance Statement This study highlights the importance of cortical tracking of the speech envelope as a reliable, noninvasive measure of how the brain processes speech. By demonstrating moderate to good test–retest reliability across various stimuli, features, reference electrodes, and electroencephalography frequency bands with as little as 390 s of data, the research supports the use of cortical tracking in both clinical and research settings. These findings will help researchers design better studies on speech processing in both neurotypical adults and those with language disorders, ultimately advancing our understanding of neural processing of speech and language.