Application of triphasic pulses with adjustable phase amplitude ratio (PAR) for cochlear ECAP recording: I. Amplitude growth functions

This study describes the use of triphasic electrical stimulation pulses with an adjustable phase amplitude ratio (PAR) for the reduction of electrical stimulus artifacts. It is hypothesized that the setting of a certain PAR can facilitate a nearly artifact-free recording of electrically evoked compound action potentials (ECAP) in the cochlea. Artifact reduction with triphasic pulses using single epochs is expected to prevent latency or polarity effects, which are seen in standard forward masking or alternating polarity strategies. Although the application of a third phase is already implemented in implants manufactured by MED-EL (Zierhofer, 2003) and Cochlear (Sydney, Nucleus 5 System: van Dijk et al. (2007)) for the reduction of stimulation artifacts generated with these stimulators in ECAP measurements, an elaborate systematic evaluation of PAR for artifact reduction has not yet been conducted (compare evaluation for one subject Schoesser et al. (2001)). In the present paper, the effect of PAR variation on human ECAP recording and the feasibility of amplitude growth function recording with triphasic pulses and an optimized PAR are evaluated. Measurements were accomplished in five subjects, whereby more detailed test series were carried out in one subject. All subjects were implanted with devices from the company MED-EL, Innsbruck. A comparison of PAR optimized triphasic pulses was carried out against two other measurement techniques (biphasic alternating polarity stimulation and biphasic stimulation according to Miller) for apical, middle, and basal electrodes. ECAP thresholds were estimated by means of amplitude growth functions. However, recording of ECAP with triphasic pulses showed drawbacks: additional artifacts depending on stimulation and/or recording parameters are introduced, the ratio between the additional artifact and improved detectability of neural responses is dependent on PAR, and response thresholds obtained with triphasic pulses - although similar in shape - are in most cases substantially higher compared to thresholds measured with the Miller method. Higher thresholds most probably occur because the triphasic pulse patterns seem to less effectively stimulate neural structures compared to biphasic pulses since measured response thresholds are higher. For certain electrode groups threshold profiles obtained with triphasic pulses were found to be similar compared to stimulation with biphasic pulses. (C) 2011 Elsevier B.V. All rights reserved.