Exploring the factors predictive of informational masking in a speech recognition task

The effects of informational masking (IM) can be recast as a question of which cues to sound source identity (auditory object formation) are most useful for overcoming IM. We hypothesize that individual differences are related to specific interactions of stimulus and listener-specific variables that determine the effectiveness of the auditory object formation process. Results from our laboratory generally support the well-established relationship between performance and stimulus variables such as spectrotemporal cues (in this case, voice differences) and spatial cues (talker locations). In addition, the listener-specific variables of age and hearing loss were found to interact with the stimulus variables and to be correlated with potential mediating variables such as interaural time sensitivity and minimum levels at which speech identification was possible. Future work will involve developing predictive models that focus on identifying the mediating variables responsible for increased susceptibility to IM and efficient tests to reveal these relationships in individual listeners. The clinical relevance of the ability to identify factors predictive of IM susceptibility will be discussed, including the potential for improved fitting of hearing aids and cochlear implants.