Auditory and visual cortical activity and relationship to functional speech outcomes in post-lingual cochlear implant users: a functional near-infrared spectroscopy (fNIRS) study

Cortical reorganisation as a result of deafness has been proposed as a potential contributor to variability in cochlear implant (CI) outcomes. Specifically, crossmodal activity in auditory and visual cortices has been shown to correlate with auditory-only speech understanding in post-lingually deaf CI users. However, it remains unclear whether these associations are adaptive’ or ‘maladaptive’ for functional outcome, with varied associations reported. Importantly, there has been variability across studies in the types of visual and auditory stimuli used, e.g., ‘basic’, such as gratings and noise, vs. ‘speech-based’ visual and auditory stimuli, which are likely to engage different cortical networks. This study therefore aims to 1) examine activations to both basic and speech-based visual and auditory stimuli in the same cohort of post-lingually deaf CI users, and a group of age-matched, normally-hearing controls, and 2) investigate whether a relationship exists between ‘crossmodal’ activation and speech recognition measures in auditory, visual-only and audiovisual conditions, and whether this changes depending on the paradigm used (basic vs. speech). fNIRS was employed as a non-invasive technique that is not susceptible to electrical or magnetic artifacts common to other neuroimaging methods such as EEG and fMRI. Two test paradigms were administered in a block design, using 1) basic auditory and visual stimuli - modulated speech-shaped noise and morphing concentric gratings, and 2) speech stimuli - concatenated IEEE sentences presented in an auditory-only and a visual-only condition. Responses were measured via a 52-channel optode array positioned over temporal and occipital regions. Speech recognition performance on IEEE sentences were assessed in auditory-only, visualonly and audiovisual conditions. Group-level analysis using a general linear mixed effects model indicates significant activations across both groups, paradigms and conditions. Activations are observed in temporal areas to auditory stimuli and in occipital areas to visual stimuli in both groups, for both basic and speech-based stimuli. Initial results indicate significant cross-modal activity in temporal regions at the group level in CI users, but not in normally-hearing controls. In summary, evidence of differential activations depending on the nature of auditory and visual stimuli used is observed between CI users and normallyhearing controls. The study presents a correlational analysis, examining reported associations between cross-modal activity to both types of visual and auditory stimuli and speech recognition in auditory-only, visualonly and audiovisual conditions.