Vibrotactile stimulation based on the fundamental frequency can improve melodic contour identification of normal-hearing listeners with a 4-channel cochlear implant simulation

Abstract

Cochlear implant (CI) users’ poor speech recognition in noise and music perception may be both due to their limited access to pitch cues such as the fundamental frequency (F0). Recent studies showed that similar to residual low-frequency acoustic hearing, vibrotactile presentation of the F0 significantly improved speech recognition in noise of CI users. The present study tested whether F0-based vibrotactile stimulation can improve melodic contour identification (MCI) of normal-hearing listeners with acoustically simulated CI processing. Each melodic contour consisted of five musical notes with one of nine contour patterns (rising, falling, or flat in each half of the contour). The F0 of the middle note was 220 or 880 Hz, and the frequency intervals between adjacent notes were 1, 3, or 5 semitones. The F0 of each note was extracted in real time and transposed to a vibration frequency centered around 110 Hz at the right forearm top. MCI was tested in five experimental conditions (with a 4- or 8-channel CI simulation alone, vibrotactile stimulation alone, and 4- or 8-channel CI simulation plus vibrotactile stimulation), each after the same amount of brief training was provided. Results showed that discrimination of vibrotactile stimuli significantly improved from chance to near perfect as the vibration frequency interval increased from 0.25 to 3 semitones. The MCI performance with vibrotactile stimulation alone was similar to that with the 4-channel CI simulation alone, but was significantly worse than that with the 8-channel CI simulation alone. Significant improvement in MCI performance with the addition of vibrotactile stimulation was only found with the 4-channel CI simulation when the middle F0 was 880 Hz and when the frequency intervals were 3 or 5 semitones. The improvement in MCI performance with than without vibrotactile stimulation was significantly correlated with the baseline MCI performance with 4-channel CI simulation alone or with the MCI performance difference between vibrotactile stimulation and 8-channel CI simulation. Therefore, when the simulated or real CI performance is relatively poor, vibrotactile stimulation based on the F0 may improve MCI with acoustic CI simulations and perhaps in real CI users as well.

Publication
Frontiers in Neuroscience, 13, 1145