Temporal integration in cochlear implants and the effect of high pulse rates

Although cochlear implants (CIs) have proven to be an invaluable help for many people afflicted with severe hearing loss, there are still many hurdles left before a full restoration of hearing. A better understanding of how individual stimuli in a pulse train interact temporally to form a conjoined percept, and what effects the stimulation rate has on the percept of loudness will be beneficial for further improvements in the development of new coding strategies and thus in the quality of life of CI-wearers. Two experiments presented here deal on the topic of temporal integration with CIs, and raise the question of the effects of the high stimulation rates made possible by the broad spread of stimulation. To this effect, curves of equal loudness were measured as a function of pulse train length for different stimulation characteristics. In the first exploratory experiment, threshold and maximum acceptable loudness (MAL) were measured, and the existence and behaviour of the critical duration of integration in cochlear implants is discussed. In the second experiment, the effect of level was further investigated by including MAL measurements at shorter durations, as well as a line of equal loudness at a comfortable level. It is found that the amount of temporal integration (the slope of integration as a function of duration) is greatly decreased in electrical hearing compared to acoustic hearing. The higher stimulation rates seem to have a compensating effect on this, increasing the slope with increasing rate. The highest rates investigated here lead to slopes that are even comparable to those found in persons with normal hearing and hearing impaired. The rate also has an increasing effect on the dynamic range, which is otherwise taken to be a correlate of good performance. The values presented here point towards larger effects of rate on dynamic range than what has been found so far in the literature for more moderate ranges. While rate effects on threshold, dynamic range and integration slope seem to act uniformly for the different test subjects, the critical duration of integration varies strongly but in a non-consistent way, possibly reflecting more central, individual-specific effects. Additionally, measurements on the voltage spread of human CI-wearers are presented which are used to validate a 3D computational model of the human cochlea developed in our group. The theoretical model falls squarely inside of the distribution of measurements. A single, implant dependent voltage-offset seems to adequately explain most of the variability