Speech sound processing by children with cochlear implants : what impact on linguistic development ? Analysis of acoustic production profiles and their relation to phonological and morphosyntactic components.

The doctoral thesis investigates the phonetic and linguistic skills of groups of children with cochlear implants (CIs) and children with typical hearing to explore the links between the perceptual limitations of CIs and language skills. At the phonetic level, various phonetic segments are compared, whose acoustic correlates reflect contrasting difficulties related to sound coding by CIs: nasal and oral vowels (related to low-frequency information), fricative consonants (related to high-frequency information), and voiced and voiceless stop consonants (where the distinction relies on temporal cues that are expected to be better coded by CIs). The results highlight a specific impact of the perceptual limitations of CIs on the productive skills involving vocalic nasality and fricatives, as well as specificities in the phonetic implementation of the voiced/voiceless stop consonant distinction. However, compensation strategies are observed among groups of chil-dren who are exposed early and intensively to Cued Speech, demonstrating the value of multimodal speech activation in their management. Positive effects of early implantation are also found. The study of lexical and grammatical skills shows specific difficulties in processing grammatical and lexical morphemes whose morphophonological alternations are carried by nasal and oral vowels, as well as narrative production skills characterized by lower Mean Length of Utter-ances and fewer occurrences of function words and complex verbal forms. Lexical diversity is also lower in the CI group. A specific link between phonological and grammatical skills is observed in the CI group, revealing a strong interde-pendence that is discussed in connection with phonological theories of morphosyntactic difficulties. Finally, phonological, lexical, and grammatical skills are linked to acoustic production profiles through factor analyses and hierarchical as-cending analyses, which identify different profiles among the children: some with very low performances and others with performances approaching those of their peers with typical hearing of the same auditory age. A one-year follow-up conducted on a subgroup of children, however, reveals a stagnation in performance among the majority of retested children, suggesting slower developmental profiles probably due to the limitations related to perceptual processing.