Cochlear Implant Simulations
ARL main page
Publications & presentations
Demos & Simulations
Office (312) 942 - 3298
Lab (312) 942 - 3316
1653 W. Congress Parkway
Cochlear implants are electronic devices that are credited with restoring partial hearing to many otherwise profoundly deaf persons. They bypass many of the typical ear functions and apply direct electrical stimulation to the auditory nerve. While benefits of a cochlear implant for someone who is deaf can be tremendous, implantation is an invasive surgical procedure that involves many risks. In order to understand what information an implant can provide to its user under different settings, acoustic simulations can be used. In a simulation, input sounds are processed similarly to an actual implant. But then, rather than being delivered as electric charges to the auditory nerve, extracted signal parameters are transformed back into sounds. Thus, a normally hearing person can listen to these output sounds and try to make sense of them, just as an actual cochlear implant user would.
Cochlear implant simulations have been used effectively to address a number of important questions related to implant design and the nature of auditory perception. Simulations provided on this site demonstrate how changes in the number of spectral channels used by an implant can affect the resulting sound. Two different kinds of simulations are presented. In one, the noise band simulation, the input signal is divided into a chosen number of logarithmically spaced frequency bands (i.e., channels), then an intensity envelope is obtained for each band. Next, the envelopes are used to modulate white noise, which is later filtered using the original filter band parameters. Finally, all the modulated noise bands are added together to produced an output sound. In the second simulation, the processing is very similar to the first one, but instead of white noise sine waves are used. Their frequencies correspond to the center frequencies of each of the frequency bands used during the initial filtering.
Both types of simulations have been used to investigate how changes in the number of channels affect speech intelligibility. Results of these studies have consistently shown that speech intelligibility remains practically the same for a given number of channels, regardless of what type of simulation has been used. Another question that can be asked is whether the same would be true for environmental sounds. And, if so, how many channels would a listener require to accurately identify the source of an environmental sound? Well, take a listen...
The noise band simulation is in the left column of the table below, and the sine wave simulation is in the right. For either of the two simulations, first select what type of sound you would like to hear. Then select a number of channels to use in the simulation. Finally, press the Play Simulation button to hear it. When listening, keep in mind that you already know what the sources of these sounds are. Try to imagine what it would sound like to someone who does not know what produced these sounds. (NOTE: depending on the speed of your Internet connection some sounds may take several seconds to download before they start playing.)