The effects of auditory stimulus level and speech recognition performance on fNIRS measured cortical activation in adults with normal hearing and adults with cochlear implants
Sheffield, Sterling Wilkinson
An objective measure, not requiring a behavioral response, of speech recognition performance in individuals with cochlear implants (CIs) would be beneficial in directing clinical recommendations. Neuroimaging can be such a measure but is difficult in the CI population. Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique that is viable for CI recipients. This dissertation began to examine the potential of fNIRS in this population by determining the effects of speech recognition performance and stimulus intensity level on fNIRS recorded cortical activation. We hypothesized that fNIRS responses would be correlated with both intensity level and speech recognition performance. Thirteen adults with bilateral CIs and 16 adults with normal hearing were included in the study. Experiment one used signal-correlated noise to determine the effect of intensity level in the range of soft to loud speech (45-75 dB SPL). Experiment two used both signal-correlated noise and sentences presented in background babble at three different signal-to-noise ratios to determine the effect of speech recognition performance on cortical activity. fNIRS data were recorded in the left hemisphere. The experiment one showed a positive correlation of auditory cortical activation with stimulus intensity level in both groups. In contrast, experiment two revealed a difference between the groups. The adults with normal hearing had stronger activation in the auditory cortex, inferior frontal gyrus, and the inferior parietal lobe with higher speech recognition. In contrast, the CI group had the opposite effect with stronger activation in those same areas with lower speech recognition. The stimulus level results are consistent with the literature and out hypothesis. The results of the second experiment are consistent with our hypothesis and the literature in only the normal hearing group. This is the first study to use neuroimaging for speech recognition in noise in individuals with CIs. Research is needed to further examine cortical activation patterns for speech recognition in noise in these two groups. These results support the potential of fNIRS as a measure of speech recognition performance at a group level. The individual variability was large, however, and might limit the potential of fNIRS at an individual level.