Evaluation of electrical tongue stimulation for communication of audio information to the brain

Abstract

Non reparative solutions to damaged or impaired sensory systems have proven highly effective in many applications but are generally underutilized. For auditory disorders, traditional reparative solutions such as hearing aids and implant technology are limited in their ability to treat neurological causes of hearing loss. A method to provide auditory information to a user via the lingual nerve is proposed. The number of mechanoreceptors in the tongue exceeds the number of inner hair cells in the cochlea and the dynamic range of neurons in both systems is comparable suggesting that the achievable throughput of information in the lingual nerve is comparable to that of the auditory nerve. This supports the feasibility of transmitting audio information to the brain via the lingual nerve. Using techniques implemented in similar successful technology, the achievable throughput of the dorsal surface of the tongue using existing stimulation methods without additional innovation was estimated to be as high as 1,800 bits per second for an experienced user, in the same range required by many audio codecs used for spoken language. To make a more accurate estimation of achievable throughput, devices were developed to stimulate the tongue electrically, and an experiment to map the sensitivity of the tongue to a form of electrotactile stimulus was performed. For the population tested, discrimination ability of the tongue varied greatly. For all participants estimates for the immediately achievable throughput for the surface of the tongue was sufficient to communicate basic phonetic information to the participant. The estimated throughput for an experienced user was estimated to be as high as 1,400 bits per second. Lingual sensitivity maps were generated that will allow researchers and developers to manufacture electrode arrays that can reliably stimulate lingual nerve endings in a discriminatory manner. In another study we tested the feasibility of sending audio information to a person via the tongue. Preliminary data are presented on participants in a learning study that were able to discern stimuli generated from recorded voices, supporting our hypothesis on immediately achievable throughputs.