h’euroscience Vol. 13, No. 2, PP. 553-562, 1984 Printed in Great Britain 0306-4522/84 $3.00 + 0.08 Pergamon Press Ltd 0 1984IBRO A MODEL FOR THE ELECTROSTIMULATION OF THE NERVUS ACUSTICUS I. J. HOCHMAIR-DESOYER,E. S. HOCHMAIR, H. MOTZ* and F. RATTAY Technical University, Vienna, Austria; *Clarendon Laboratory, Oxford, U.K. Abstract-Electrostimulation of the nervus acusticus has been successfully used to achieve speech understanding. A model of nerve excitation due to Fitzhugh [(1961) Biophys. J. zyxwvutsrqponmlkjihgfedcb 1, 445-4461 has been explored to show that the stochastic response of single acoustic mammalian fibers observed by Rose, Brugge, Anderson and Hind [(1967) J. Neurophysiol. 30,769-7931 may be due to the non-linear interaction of sub-threshold electrical nerve activity with the signal. Computations by means of an analogue computer also shows that a phase-locked response with a frequency dependence resembling behaviour observed with electrostimulation is obtained with the model. A dynamic range for single fiber excitation is produced by the non-linear interaction. This work tries to set up a theoretical background for the results that are achieved with electrical stimu- lation of the auditory nerve in deaf patients by using a system called a “cochlear prosthesis”. The system contains stimulation electrodes4 which are surgically placed either in the inner ear in close proximity to the excitable structures of the auditory nerve or external to the chochlea, thus stimulating a larger portion of the auditory nerve. The prosthesis can be used in cases of severe damage to the inner ear structures. In such cases the incoming sound, which is transformed into fluid motion inside the inner ear, can no longer be transformed into neuronal activity. An electrically stimulating electrode serves to evoke the neuronal activity. The basic components of a cochlear prosthesis are demonstrated in Fig. 1. They are: the electrode, implantable electronic receiver/stimulator circuitry and an external sound-processor/transmitter. In our original experiments different frequency bands were supplied to the four bipolar channels. It has, how- ever, proved preferable to stimulate with a single optimally sited, bipolar electrode using a broadband signal. Many different coding schemes for the simu- lation signals have been tried. Recently it was found that most patients can distinguish between sinewaves, rectangles and pulses6 and that, furthermore, the use of analogue stimulation signals (a direct analogue of the sound wave adequately compressed in dynamic range and adjusted for isoloudness response) leads to better speech comprehension5 than the use of pul- satile signals. We have found that considerable im- provement of performance is obtained by careful adjustment of the amplitude as a function of fre- quency of the wideband signal supplied to the single bipolar channel to suit individual requirements of the *Address for correspondence: Professor H. Motz, 16 Bed- ford Street, Oxford OX4 ISU, U.K. Abbreviations: AP, action potential; RF, range factor. patients. Very good results have also been obtained by stimulating with a window outside the cochlea, e.g. near the round window. The patients were tested for residual hearing and their pure tone audiograms showed thresholds worse than or equal to 95 dB SPL at all frequencies above or at 250 Hz except for one case. They all had no measurable threshold up to 120dB for frequencies above or at 1000 Hz. The possibility of hearing in the normal acoustic sense is excluded in our tests. A direct cable connects the output of the tape machine supplying the speech signals to an input jack in the patient’s speech processor which disables its micro- phone. The tape machine was in a different room and the stimuli were thus not present in the test room in acoustic or optical form. In the most favourable cases, patients achieved an almost 100% understanding of sentences from an open list. The understanding of one-syllable words from an open list reached 70%. We have so far not been able to improve our results by the use of larger numbers of electrodes distributed along the scala tympany designed to stimulate nerve endings selec- tively with different frequencies in a manner resem- bling what we believe to be the working of the normal ear. Most of our patients have implants with four bipolar electrodes and we are exploring new strategies for their use. The good results obtained with a single bipolar electrode require explanation. BASIC CONSIDERATIONS In this paper we offer a theoretical model of the transmission of the signal to the nerve. The signal emanates from one of the dipoles embedded in the cochlea, or from an extra-cochlear electrode. In both cases currents flow in the slightly conducting liquid surrounding the nerve fibers or dendrites which are irregularly oriented with respect to the lines of electric 553