Neural Representation of Temporally Jittered Vowels: A Computational Model of Auditory Aging

Older adults face difficulty in the hearing process due to age-related issues. Research had shown that one of the main contributing factors to this problem is the loss of neural synchrony (phase-locking ability) in the auditory-nerve fibers with the increase in age. This neural asynchrony contributes to the lower speech perception in the listeners. Different behavioral studies had shown that this age-related neural asynchrony could be modeled through an acoustic temporal jitter algorithm. In the current study, this hypothesis is verified through a computational auditory-nerve model. The input to the model is considered as a steady-state jittered single vowel, whereas the output is a peri-stimulus time histogram response at a particular characteristic frequency of the auditory-nerve fiber. A synchronized rate is obtained from these peri-stimulus time histogram response to analyze the neural representation. The synchronized rate for a particular vowel /i/ had shown that along with the original frequency component (2.25 kHz), the fiber also responds to a spurious frequency component (2.35 kHz). Therefore, synchronized rate clearly showed a loss of neural synchrony, which might reduce speech intelligibility among older adults. To restore the lost neural synchrony in the jittered vowels, a simple moving average filter is proposed. This simple technique had shown a decrement of 45.43% in the spurious frequency component of the recovered vowel, which suggests an improvement in the phase-locking ability of the auditory-nerve fiber. A quantitative analysis is implemented on the synchronized rate of vowel /i/ to show a significant decrement in the spurious frequency component after passing through the moving average filter. Additionally, an acoustic time and frequency domain analysis is done on the jittered vowels.