Previous researchers in speech perception looked for invariant cues that uniquely specified the phonetic identity of a speech segment. Instead, researchers discovered multiple, context-dependent cues used by listeners in making phonetic judgements. The goal of this research is to investigate the mapping from these perceptually relevant, acoustic cues to linguistic units in stop consonant perception by comparing the performance of statistical and neural models to results from human psychophysical studies, which will be undertaken. Through this comparison, the research will create a functional model of stop consonant perception, determine how much of perception is dictated by these cues, provide insight into whether more complex representations are needed, and provide a basis for searching for other acoustic cues. The first stage will extend an initial list of known perceptually relevant acoustic cues by testing possible acoustic cues on human subjects to see if a given cue affects phonetic judgment. The second stage will compare human and model performance on synthetic speech stimuli, using the acoustic cues that were determined in the first stage, to determine if a more complex representation is needed, and which representation most closely emulates human perception. Finally, the performance of both humans and the previously-trained models will be evaluated on novel real speech tokens to determine how much of natural stop consonant perception can be accounted for based solely on these cues. In addition, the discrepancies will be used as an impetus for discovering more acoustic cues. This research will not only provide insight into speech perception and provide a model, but it could also be used to improve hearing aids by accentuating the relevant attributes.

Agency
National Institute of Health (NIH)
Institute
National Institute on Deafness and Other Communication Disorders (NIDCD)
Type
Research Project (R01)
Project #
7R01DC000194-13
Application #
2125054
Study Section
Hearing Research Study Section (HAR)
Project Start
1983-01-01
Project End
1997-03-31
Budget Start
1995-04-01
Budget End
1996-03-31
Support Year
13
Fiscal Year
1995
Total Cost
Indirect Cost
Name
Massachusetts Eye and Ear Infirmary
Department
Type
DUNS #
073825945
City
Boston
State
MA
Country
United States
Zip Code
02114
Rosowski, John J; Bowers, Peter; Nakajima, Hideko H (2018) Limits on normal cochlear 'third' windows provided by previous investigations of additional sound paths into and out of the cat inner ear. Hear Res 360:3-13
Chhan, David; McKinnon, Melissa L; Rosowski, John J (2017) Identification of induced and naturally occurring conductive hearing loss in mice using bone conduction. Hear Res 346:45-54
Ravicz, Michael E; Rosowski, John J (2017) Chinchilla middle ear transmission matrix model and middle-ear flexibility. J Acoust Soc Am 141:3274
Chhan, David; Bowers, Peter; McKinnon, Melissa L et al. (2016) Middle-ear and inner-ear contribution to bone conduction in chinchilla: The development of Carhart's notch. Hear Res 340:144-152
Ravicz, Michael E; Rosowski, John J (2013) Inner-ear sound pressures near the base of the cochlea in chinchilla: further investigation. J Acoust Soc Am 133:2208-23
Chhan, David; Röösli, Christof; McKinnon, Melissa L et al. (2013) Evidence of inner ear contribution in bone conduction in chinchilla. Hear Res 301:66-71
Ravicz, Michael E; Rosowski, John J (2013) Middle-ear velocity transfer function, cochlear input immittance, and middle-ear efficiency in chinchilla. J Acoust Soc Am 134:2852-65
Röösli, Christof; Chhan, David; Halpin, Christopher et al. (2012) Comparison of umbo velocity in air- and bone-conduction. Hear Res 290:83-90
Puria, Sunil; Rosowski, John J (2012) Bekesy's contributions to our present understanding of sound conduction to the inner ear. Hear Res 293:21-30
Ravicz, Michael E; Rosowski, John J (2012) Chinchilla middle-ear admittance and sound power: high-frequency estimates and effects of inner-ear modifications. J Acoust Soc Am 132:2437-54

Showing the most recent 10 out of 40 publications