Poster No:
1047
Submission Type:
Abstract Submission
Authors:
Qingqing Meng1, Angela Wong1, Jennifer Clemesha1,2, Jorge Mejia1
Institutions:
1National Acoustic Laboratories, Sydney, NSW, Australia, 2Cochlear Limited, Sydney, NSW, Australia
First Author:
Qingqing Meng
National Acoustic Laboratories
Sydney, NSW, Australia
Co-Author(s):
Angela Wong
National Acoustic Laboratories
Sydney, NSW, Australia
Jennifer Clemesha
National Acoustic Laboratories|Cochlear Limited
Sydney, NSW, Australia|Sydney, NSW, Australia
Jorge Mejia
National Acoustic Laboratories
Sydney, NSW, Australia
Introduction:
Human brain activity measured with magnetoencephalography (MEG) has been shown to track the hierarchical linguistic units embedded in connected speech (Ding et al., 2016). In addition, the responses at hierarchical levels that are dissociated from any acoustic cues to them, such as phrase and sentence level, can be directly modulated by changes in speech intelligibility caused by spectral degradation or immediate prior knowledge on speech signal (Meng et al., 2021, 2022).
Methods:
In the current study, we introduce background noise and manipulate its sound level relative to the target speech to reduce the Signal-to-Noise Ratio (SNR) and speech intelligibility. We hypothesis that the tracking responses can be simultaneously but differently modulated by the variations in SNR and associated changes in speech intelligibility. Continuous Electroencephalography (EEG) responses to isochronous speech sentences presented alone and together with a multi-talker babble noise at three different levels were measured in nineteen normal hearing participants. Subjective ratings of speech understanding were also collected from individual participant after each stream of sentences presentation, across all background noise conditions. Sensor level coherence between the EEG recordings and the temporal regularities of multi-level linguistic units were calculated and mapped back to the cerebral cortex using a whole-brain frequency domain beamforming technique (Gross et al., 2001).
Results:
Results showed that concurrent brain responses to hierarchically nested linguistic structure in connected speech can be reliably measured using EEG, even in the presence of background noise. Driven by the reduction in speech intelligibility, cortical activities coherent to "abstract" linguistic units with no accompanying acoustic cues (phrases and sentences) were reduced relative to the "no noise" condition, and lateralized to single cerebral hemisphere. In contrast, brain response tracking word units that are aligned with syllable/acoustic onsets, were bilateral and exhibited a systematic reduction as the noise level increased (SNR decreased) across conditions.
Conclusions:
This dissociated result suggests that brain processes of encoding linguistic information during speech perception are directly affected by speech intelligibility, which in turn are powerfully shaped by acoustic properties of the speech signal. Subjective rating scores on speech understanding differed significantly across all conditions and correlated with the magnitude of EEG tracking responses at all linguistic levels. These results provide an objective and sensitive neural level index of speech intelligibility and speech perception which has the potential to be further developed into clinical applications for the assessment of speech-in-noise performance among hearing impaired population, such as hearing aid users and cochlear implant recipients.
Language:
Speech Perception 1
Modeling and Analysis Methods:
EEG/MEG Modeling and Analysis 2
Keywords:
Electroencephaolography (EEG)
Hearing
Language
Source Localization
1|2Indicates the priority used for review
Provide references using author date format
Ding, N., Melloni, L., Zhang, H., Tian, X., & Poeppel, D. (2016). Cortical tracking of hierarchical linguistic structures in connected speech. Nature Neuroscience, 19(1), 158.
Gross, J., Kujala, J., Hämäläinen, M., Timmermann, L., Schnitzler, A., & Salmelin, R. (2001). Dynamic imaging of coherent sources: Studying neural interactions in the human brain. Proceedings of the National Academy of Sciences, 98(2), 694–699.
Meng, Q., Hegner, Y. L., Giblin, I., McMahon, C., & Johnson, B. W. (2021). Lateralized cerebral processing of abstract linguistic structure in clear and degraded speech. Cerebral Cortex, 31(1), 591–602.
Meng, Q., Hegner, Y. L., Giblin, I., McMahon, C., & Johnson, B. W. (2022). Prior exposure to speech rapidly modulates cortical processing of high-level linguistic structure (p. 2022.01.25.477669). bioRxiv. https://doi.org/10.1101/2022.01.25.477669