Impact of noise and noise reduction on processing effort: A pupillometry study

Speech understanding in adverse listening conditions induces cognitive demands. Processing effort typically increases for acoustically degraded speech. Moreover, processing effort depends on listener-specific factors, such as cognitive abilities and hearing loss. While noise reduction schemes are able to attenuate interfering noise, the impact on processing effort has not yet been clarified. The present study examined the benefit of noise reduction on processing effort allocated for speech recognition. The influence of noise reduction on processing effort was evaluated by measuring the Peak Pupil Dilation (PPD) of 24 hearing-impaired listeners while performing a speech recognition task. Speech recognition performances and PPDs were measured at two different signal-to-noise ratios (SNRs) corresponding to either the individual 50% correct (L50) or the 95% correct (L95) performance level with and without noise reduction. Processing effort increased at the L50 compared to the L95. Moreover, the benefit of the noise reduction was demonstrated for both the L50 and the L95, i.e., where speech recognition was at ceiling performance. The results emphasized the relevance of the PPD measure as an indicator of processing effort in situations where the traditional speech reception measures fail due to ceiling effects.