Neural encoding of complex signals in the healthy and impaired auditory systems

Saradha Ananthakrishnan, Purdue University

Abstract

Individuals with sensorineural hearing loss (SNHL) typically experience difficulty in understanding speech. Our current knowledge of deficits in speech perception and encoding consequent to SNHL is restricted to psychophysical studies in humans and single-unit experiments in animals. However, the nature of degradation in neural encoding of speech following hearing impairment in humans has not been extensively researched. The objective of this dissertation is to provide a systematic evaluation of neurobiological signature of hearing loss at the subcortical level using an objective electrophysiological non-invasive neural index, the frequency following response (FFR). Subcortical neural encoding of speech signals is explored by quantifying the effects of hearing loss on brainstem processing of acoustic features important for pitch and speech perception, namely the envelope (F0) and temporal fine structure (TFS) (formant structure). In order to capture neural encoding of hearing impaired speech perception in various real-world situations, brainstem representations are studied in response to a variety of stimuli presented in a number of different listening situations. Subcortical neural representations of envelope and TFS in response to stimuli presented in quiet listening conditions are investigated in the first part of the dissertation. Evidence from the brainstem FFR suggests that neural phase locking of both envelope (F0) as well as TFS (formant related harmonics) is reduced in hearing impaired (HI) subjects as compared to normal hearing (NH) subjects, when stimuli are unadjusted for audibility. The question then emerges if these degraded neural representations of envelope and TFS persist when stimuli are presented at equal audibility. Comparisons of the brainstem FFR at equal audibility levels between normal hearing and hearing impaired continue to demonstrate group differences, albeit reduced, suggesting that degradation of the neural representation in hearing loss cannot be attributed wholly to audibility. Rather, these representations appear to reflect a complex interplay of attenuation and distortion effects subsequent to SNHL. Further, envelope and TFS encoding are sensitive to pitch contour and formant structure. The second part of the dissertation addresses subcortical encoding of envelope and TFS cues following SNHL in degraded listening conditions such as reverberation and background noise. Results indicate a definite degradation of subcortical speech encoding with increased background noise and reverberation in both NH and HI subjects, although these effects are dependent on stimulus, level and type of degradation. Thirdly, this dissertation examines sources of variation in brainstem speech encoding. Overall, findings suggest that degree of hearing loss, hearing aid satisfaction and music experience may be strong predictors of the fidelity of neural representation of certain acoustic features as reflected in the FFR in hearing impairment. Finally, the results of this dissertation establish the FFR as a viable technique to measure brainstem speech encoding in hearing impaired listeners to a range of stimuli in a variety of listening conditions. Translation of the brainstem FFR from the lab to the clinic would add great value to the existing audiological test battery, and the potential clinical applications of the FFR are discussed.

Degree

Ph.D.

Advisors

Krishnan, Purdue University.

Subject Area

Audiology

COinS