P04Session 2 (Friday 10 January 2025, 09:30-11:30)Assessment of listening effort for cochlear implant users with pupillometry and subjective ratings: A comparison of three speech-in-noise tests
Background: Cochlear implant (CI) users experience listening as effortful, and especially in background noise. Listening effort can be measured with pupillometry, as based on the observation that the pupil dilates more when listening is more challenging. To assess listening effort in different noisy environments, a speech-in-noise test is needed where the magnitude of pupil dilation is dependent on the SNR. In contrast to typical hearing listeners (TH), however, PPDs have proven to be quite insensitive to SNR for CI users.
Objective: We have compared three speech corpuses to assess which one is most sensitive to measure the pupil response as a function of SNR for CI users. These were digits in noise (DIN), Matrix sentences, and LIST sentences.
Methods: Peak pupil dilation (PPD) was recorded for 18 CI users when listening to DIN triplets, Matrix sentences and LIST sentences at three different SNRs, namely at the speech recognition threshold where SI was 50% (0 dB re SRT), at + 6 dB re SRT, and in quiet. The DIN material consists of triplets drawn from a closed set of 10 digits; the Matrix sentences consist of 5 words with a fixed syntactic structure that provide no context; the LIST sentences are an open set of sentences that were developed especially for CI users.
Results: PPDs were significantly dependent on SNR for the DIN triplets and LIST sentences, but not for the Matrix sentences. For the DIN triplets, the PPD differed significantly between 0 dB re SRT and quiet, and between +6 dB re SRT and quiet. By contrast, for the LIST sentences the PPD was significantly different between 0 dB re SRT and quiet and between 0 dB re SRT and +6 dB re SRT.
Conclusions: The LIST sentences are useful for measuring listening effort with pupillometry in CI users around SRT, whereas DIN triplets are useful at more favorable SNRs. The Matrix sentences yielded PPDs that were unresponsive to changes in SNR and it proved to be a challenging listening test even in quiet.
Funding: This work was funded in part by the Dutch Research Council (NWO) within the Crossover Research program under the ‘INTENSE’ project (grant #17619) and co-funded by Advanced Bionics (Hannover, Germany).
Acknowledgements: We thank the study participants and Nicolas Furnon (Advanced Bionics) for technical support.