Characterizing the role of hearing loss in comodulation masking release

The detection of pure tones embedded in noise can be facilitated if amplitude modulations are imposed on the noise masker by multiplying it by low-frequency noise (i.e., modulated noise). This phenomenon is known as comodulation masking release (CMR). It can be quantified by measuring the reduction in the masked threshold observed when the masker is modulated, compared to a masker with the same spectrum and level but with an unmodulated envelope (i.e., unmodulated noise).

CMR in normal-hearing (NH) listeners has been observed through several experimental paradigms. Nevertheless, the perceptual mechanisms responsible for it are still unclear. In the studies investigating CMR in listeners with sensori-neural hearing loss (SNHL), an overall reduction in the amount of masking release was observed. Loss of sensitivity and reduced frequency selectivity were found to be linked to the reduction in CMR. However, the deficits in auditory processing directly responsible for the reduced CMR have not been precisely identified.

The aim of this study is to extend the investigation of CMR in relation to SNHL including other aspects of hearing. CMR was measured in a group of listeners with sloping, mild-to-severe SNHL. Large differences in the CMR across participants were observed, some showing normal CMR, others exhibiting reduced or absent CMR. The amount of CMR was related to several aspects of auditory processing, assessed by means of behavioral experiments. These included measurements of absolute thresholds, estimates of auditory filter width, cochlear compression and sensitivity to temporal fine structure. A statistical analysis of the results indicated a significant effect of frequency selectivity and sensitivity on the amount of CMR in HI listeners, in line with previous studies. No significant effect of the estimated cochlear compression ratio or sensitivity to temporal fine structure was observed. Additionally, simulations have been conducted with a computational model of auditory perception that has been shown to account for several aspects of hearing impairment. The model failed to predict the experimental results, even when the individual elevation of absolute thresholds, increased auditory filter bandwidth and loss of cochlear compression were included in the model front end. This suggests that other deficits in the auditory system, such as the processing of the temporal envelope of the stimulus, might be related to the CMR in listeners with SNHL.