Active Research

Principal Investigator: Michelle R. Molis, Ph.D.

Objective: To compare patterns of vowel identification in noise between and among hearing-impaired and normal hearing listeners and to relate vowel identification performance of individuals to estimates of their spectral resolution abilities.

Background:  Hearing impairment frequently results in reduced frequency selectivity in the auditory periphery. Additionally, there is a loss of perceptual sensitivity that requires inputs be presented at increased levels to ensure audibility - a step that may further decrease frequency selectivity. The result is a smoothed internal representation lacking unambiguous spectral prominences corresponding to formant frequencies. The introduction of a competing background noise further exacerbates this situation. Optimal vowel exemplars are rarely produced by speakers in real world situations. Instead, there is extensive overlap between category tokens. This stimulus ambiguity makes understanding speech in everyday listening situations an even greater challenge for individuals with hearing loss.

Findings: An evaluation of vowel identification with ambiguous stimuli in the presence of competing noise will provide a more realistic evaluation of listeners’ capabilities than is commonly provided in a laboratory setting.