Towards testing comodulation masking release in electric hearing

Pierzycki, Robert H. and Seeber, Bernhard U.
MRC Institute of Hearing Research, Nottingham NG7 2RD, UK,

Cochlear implant (CI) patients experience difficulty understanding speech in the presence of modulated 
noise, and this can also be observed in simulated CI processing (Qin & Oxenham, 2003). In normal 
hearing, detection thresholds of a tone masked by an amplitude-modulated on-frequency noise masker 
can be reduced when presented together with coherently amplitude modulated flanking bands of noise. 
This effect has been termed comodulation masking release (CMR) and appears to rely on across-channel 
comparisons of temporal envelopes (Hall et al., 1984). However, the temporal fine structure (TFS) 
information is known to also contribute to CMR. Furthermore, most CMR experiments test detection of a 
tone signal and hence listeners may benefit from the perception of its pitch - a cue barely available in CI 
listening. The present studies investigate if CMR is available in CI listening without explicitly encoded TFS 
information and thus the degree to which CMR relies on comodulation of envelopes across multiple 
channels. 
CMR was investigated in normally hearing people in preparation for testing with CI patients. A flanking 
band paradigm was used in which CMR was observed as a decrease of detection threshold of a tone in 
the presence of co-modulated compared to anti-modulated flanking bands of noise. The flanking bands 
mimic stimulation on separate, distant electrodes in electric hearing. The paradigm was further adapted 
to CI listening by using noise bands that covered roughly equal distances on the basilar membrane 
similar to electric stimulation. A vocoder was used to replace the stimuli TFS by extracting stimuli 
envelopes in separate frequency channels and modulating carrier signals (band-pass noises or tones) 
with the extracted envelopes. This was done to simulate the lack of signal-specific encoding of the TFS 
information in electric hearing. 
CMR was observed with and without vocoding, indicating that it is a robust envelope-based process. 
However, replacing the TFS with the vocoder slightly reduced CMR, suggesting that TFS contributes 
somewhat to CMR. Nonetheless, as strong CMR can be observed despite vocoding, it is hoped that 
comodulation of envelopes could improve signal detection in electric hearing. 

Acknowledgements
----------------
Work supported by the intramural programme of the Medical Research Council.

References
----------
Hall J.W., Haggard M.P. & Fernandes M.A. 1984. Detection in noise by spectro-temporal pattern analysis. 
J Acoust Soc Am, 76, 50-56.

Qin M.K. & Oxenham A.J. 2003. Effects of simulated cochlear-implant processing on speech reception in 
fluctuating maskers. J Acoust Soc Am, 114, 446-454.