From sounds to scenes: Comprehension of environmental sounds by cochlear implant users

Abstract

Performance outcomes for cochlear implant (CI) users traditionally focus on measures of speech perception. However, existing research indicates that environmental sound identification tasks also remain challenging for adult CI users compared to normal-hearing (NH) or hearing-impaired (HI) peers. In contrast, anecdotal reports indicate that environmental sound perception improves post-implantation. Methodological choices may contribute to this discrepancy; CI users may be benefiting from more integrative higher-level processes that are not adequately measured by source identification tasks. This study employs two alternative tasks that are designed to assess perception of the semantic properties of environmental sounds. The first is a comprehension tasks, which requires listeners to make inferences about naturalistic sound scene recordings (e.g. which other activities might you expect to take place, or at which time of day does this scene most likely occur). The second task presents a triplet of isolated environmental sound recordings and requires participants to select the sound that does not belong. Preliminary data indicate that CI users are able to perform the tasks with varying levels of proficiency relative to NH and HI listeners. Comprehension tasks focused on context-dependent semantic processing may, thus, complement findings from more traditional single-sound identification tasks.