|
Abstract of Hannah's Poster Presentation
Speech Based Optimization of Hearing Devices
Authors: Hannah W. Siburt, BS, Alice Holmes, PhD, Rahul Shirvastav, PhD and Lee Krause, BS
Background:
Cochlear implants (CI) are devices which provide auditory information for patients with severe-to-profound hearing loss. Modern CIs are sophisticated devices capable of advanced signal processing. However, no standard procedures exist to identify optimal programming settings for individual patients. An optimization programming routine was designed to improve speech recognition using real-time Distinctive Feature (DF) analysis (Jakobson, et al., 1963) of phoneme errors. The CI device is adjusted based on phoneme recognition abilities, not subjective quality. This approach may provide a standardized CI fitting method, individualizing programming in an efficient manner.
Methods:
Twenty experienced CI users were recruited from the UF Speech & Hearing Clinic and participated in three sessions, two weeks apart, using a repeated measures design Subjects' speech processors were reprogrammed using the optimization procedure with over 25 programs tested in 30 minutes. Outcome measures were completed after two weeks experience with the optimized programs and compared to baseline. The study was conducted under UF IRB-01 approval. Participants received monetary compensation.
Results:
Analysis of Variance with Repeated Measures (ANOVA-R) was completed for each outcome measure. Percent scores were transformed into Arcsine units (RAU Transform) to stabilize the error (Studebaker, 1985). The ANOVA-R indicated a significant difference among the three conditions using both Sphericity Assumed analysis (p < 0.003) and Greenhouse-Geisser analysis (p < 0.004). Significant improvements were found with the optimized procedure compared to standard clinical methods.
Conclusions:
This procedure shows promise to improve patient performance and increase satisfaction in a shorter test time. The optimization method resulted in improved subject performance in all measures of word and phoneme identification with optimized maps, as well as in noise. Future applications for this method include expanding the number of optimized parameters in CIs and applying the procedure to other digital hearing devices.
|
