Auditory Continuity Illusion and Perceptual Restoration of Speech in Noise: Relationship with Speech Intelligibility in Noise in Individuals with Normal Hearing, and Cochlear

Abstract

Understanding speech in adverse acoustic conditions is a subject of interest in those with normal hearing sensitivity, and more so in those with hearing loss. Top-down processing plays a critical role when the bottom-up encoding becomes unreliable due to signal distortion. The present study addressed two related measures of top-down processing: auditory continuity illusion and phonemic/perceptual restoration of speech in noise. The study aimed to investigate the relationship between behavioral auditory continuity perception, mismatch negativity for auditory continuity, phonemic restoration of speech in noise and speech perception in noise performance in those with normal hearing and mild to moderate sensorineural hearing loss. Behavioral continuity perception was assessed using natural vowel /a/, pure tones: 500 Hz and 4000 Hz, and 1000 Hz amplitude modulated tone. Phonemic restoration of speech in noise was measured as the difference in speech identification scores between sentences interrupted by silences and those interrupted by silences filled with speech noise. The mismatch negativity was measured for the vowel /a/ at 0 and -10 dB SNRs. The relationship between parameters of these measures and how their correlation with speech perception in noise performance was probed. The behavioural continuity percept was the strongest for the amplitude modulated tone, followed by the pure tones, and the weakest for the vowel /a/ in both the groups, and the difference between groups was not significant for any stimulus. The phonemic restoration magnitude was appreciable in the normal hearing group, but was drastically affected in the SNHL group, and was negative in many. The mismatch negativity elicited in the normal hearing group was significantly different between the 0 and -10 dB SNRs while no such difference was observed in the SNHL group. Significant correlation was found between behavioral measures in the normal hearing group but in the SNHL group. The correlation, however, was significant between behavioral continuity ratings and MMN amplitude in both the groups. Continuity perception measures (both behavioral and electrophysiological) were not correlated with speech in noise perception in either group. We conclude that continuity perception is subtly affected, and phonemic restoration is severely affected in the SNHL group. Absence of differences in continuity perception at the behavioral level, but evidence of deficits at the pre-attentive physiological level suggests compensatory mechanisms in those with SNHL. Absence of correlation between measures of restoration in in the SNHL group is suggestive of disrupted top-down processing secondary to deficient bottom-up input.