Speech Perception with Cochlear Implants

TLDR: Cochlear implants, besides restoring hearing sensation to otherwise deaf individuals, provide an excellent tool with which to investigate how the human central nervous system (CNS) processes complex patterns of sensory information.

Abstract: Cochlear implants, besides restoring hearing sensation to otherwise deaf individuals, provide an excellent tool with which to investigate how the human central nervous system (CNS) processes complex patterns of sensory information. Throughout the lifetime of normal-hearing persons, the auditory CNS has been continually trained to extract meaningful speech (and other meaningful sounds) from a constant barrage of auditory sensory information. The CNS establishes networks to process auditory sensory information; for complex pattern recognition tasks, these networks can take as long as 10 to 12 years to fully develop (see Hartmann and Kral, Chapter 6). Once these networks are fully mature, auditory pattern recognition is highly robust to degradations in the sensory signal, as revealed by decades of speech perception research. For example, military cryptologists in the 1940s searched for a type of signal degradation that would render speech unintelligible during transmission (but could be decoded at the receiving end by reversing the degradation, thereby restoring intelligibility). To their amazement, even severe alterations to the speech signal did not destroy its intelligibility. One of the most well-known examples is the work of Licklider and Pollack (1948), who eliminated all amplitude information of the speech signal by means of “infinite clipping” (the signal waveform was simply absent or present, according to an amplitude threshold).