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) users show no improvement in speech understanding when 
speech is masked by modulated rather than steady maskers. In normal hearing, 
correlated modulation on spectrally distant flanking bands (FBs) of noise can 
lead to reduced thresholds of a tone masked by a modulated narrow-band 
on-frequency masker (OFB), an effect known as comodulation masking release (CMR). 
Comodulation is thus a process that could potentially improve detection in 
electric hearing since electric hearing with current CIs relies on envelope 
information. 

CMR was investigated with CI participants listening to modulated pulse trains 
applied to separate electrodes of their CI. Envelopes of a sinusoidally 
amplitude-modulated (SAM) band of noise (OFB) summed with a tone, and those 
of two SAM FBs of noise were extracted to modulate constant rate pulse trains 
presented in staggered order on three distant electrodes. The envelopes of 
pulse trains on the FB electrodes were modulated either in or out of phase, 
i.e. co- or anti-modulated, with the OFB envelope. Three rates of SAM, 
i.e. 20, 14, and 8 Hz, were used. Tone thresholds were determined for pulse 
trains modulated by the envelope of OFB plus tone applied on the OFB electrode 
alone or together with the co- or anti-modulated pulse trains on FB electrodes. 

When stimulation was presented on OFB and FB electrodes, tone thresholds 
increased compared to those obtained for stimulation on the OFB electrode 
alone. This trend was observed for all modulation rates used. Thresholds were 
lower in the co- than in the anti-modulated condition for one of the subjects, 
in principle indicating CMR. However, since thresholds always remained higher 
than for the OFB presented alone, this threshold difference might be due to 
simultaneous masking from the FBs and not be related to CMR.