Speech processing cues in children with cochlear implants

Despite the proven effectiveness of cochlear implants for audiological rehabilitation for children with bilateral profound sensorineural hearing loss, variability in performance is large in terms of audibility, speech perception and language processing. This research project aimed at providing a better understanding of the way speech-related acoustic cues are utilized in language processing by children with cochlear implants. The knowledge provided by the current work aims at ultimately improving outcomes of speech perception and language processing of children with cochlear implants through tailored speech processing algorithms, rehabilitative tools and guidelines.

We first conducted a narrative literature review on the difficulties in perception and utilization of the speech-related spectro-temporal cues by cochlear implant users. The literature review targets speech-language pathologists, teachers, and other personnel/authorities to help them better understand the difficulties experienced by cochlear implant users. This would then allow them to better tailor the rehabilitation tools, and provide suitable adjustments. Additionally, the literature review points out gaps in research and encourages future work within them. The current greatest challenge for a cochlear implant is the delivery of fine spectro-temporal cues, which results in difficulties with segregating speech from background noise and impaired perception of speech prosodic aspects of intonation, stress, and pitch.

We then conducted two experimental studies focusing on studying the effect of “sonority” in terms of audibility vs. the knowledge of language-universal rules on the novel-word learning abilities of children with cochlear implants. In addition, the effect of optimal oral-language exposure in terms of early implantation and binaural hearing on the novel-word learning ability has been explored. Sonority is a scalar property of the relative loudness/ perceptual prominence of a particular sound segment compared to other sounds of the same length, stress, and pitch. The first experimental study conducted in Greek showed that children with cochlear implants exposed to preimplant periods of auditory deprivation tend to learn words with the static template of perceptually prominent component more accurately, adopting the early-stage word-learning strategy known as template-driven word learning. The second experimental study conducted in a group of Flemish children provided evidence that children implanted below the age of two, who have no other disabilities, who have optimal exposure through bilateral hearing, perform even more in line with normal-hearing listeners in novel-word-learning tasks than children implanted at a later, albeit young, age. However, children with cochlear implants continued to show lower receptive vocabulary scores than their normal-hearing peers, despite the equivalent novel-word learning abilities. This casts a shadow on the role of the additional processes, other than fast mapping, which are required for the development of a lexicon of words.