|
|
|
 |
 |
What Does the Cochlea Tell the Brain? |
Another important stream of basic research that has led to refinements in cochlear implants was initiated in 1965 by Nelson Kiang of Harvard University. Kiang examined the impulses traveling down the auditory nerve in response to sound and learned a great deal about how sound information is encoded in the nerve and in the brain. He discovered, for example, that any nerve fiber produces a proportional number of nerve impulses in response to an increasing frequency of sound, although in a random pattern. A single nerve fiber can produce impulses at most only 200 to 300 times a second. Yet speech involves sounds at frequencies of up to 4,000 hertz (cycles per second), and humans can hear frequencies of up to 20,000 hertz. Taken together, the random nature of nerve impulses and their highest rate must mean that an entire population of nerve fibers, all responsive to sound in the same frequency range, must be required to fully encode a single frequency of sound. Donald Eddington of the University of Utah (now at MIT) and Merzenich and his team have attempted to directly simulate these distributed response patterns in their cochlear implant model.
Beginning in the mid-1970s, Murray Sachs and Eric Young, of Johns Hopkins University, studied the responses of the auditory nerve to complex stimuli, such as speech. They determined that the brain is not merely analyzing the various frequencies but making sophisticated use of the temporal patterns of nerve impulses. This sophisticated processing probably underlies our ability to pick out a single conversation in a noisy room and to localize sounds in three dimensions.
These insights have yet to be incorporated into the design of cochlear implants, but a separate line of research has been taken. Blake Wilson, of the Research Triangle Institute in North Carolina, noticed that, because the cochlea is filled with a conductive fluid, the stimulation at one electrode in a cochlear implant spreads to nerve fibers far from its intended target. This cross-talk, as it is called, tends to make a sound muddy and difficult to interpret. He reasoned that the problem might be reduced if the electrodes in a cochlear implant were stimulated sequentially instead of simultaneously. When this scheme, known as interleaving, was introduced into the external speech processors that are part of every cochlear implant. Implant wearers reported greatly increased satisfaction with the devices (modern implants contain up to 22 electrodes--two of the 24 channels observed by Zwicker and colleagues in 1957 are considered unimportant in speech perception).
|
|
|
|
|
|
|
