Memoire Online - Contrainte Psycho-Physiques et Electrophysiologiques sur le codage de la stimulation électrique chez les sujets porteurs d'un implant cochléaire

L'objectif de cette pré-étude était de mettre au point un protocole d'enregistrement, de stimulation, de traitement et d'analyse des recueils de PEAEP.

Elle a permis de démontrer qu'il était possible d'obtenir des PEAEP de relativement bonne qualité chez plusieurs patients, à différents niveaux d'intensité, sur différentes zones de stimulation de la cochlée. Les ondes II, III et V sont très distinctes,

Elle a mis en évidence les faibles variations des latences des ondes en fonction de la zone et de l'intensité de stimulation. Cela peut s'expliquer, comparer à un PEAP par stimulation acoustique, par le shunt de la biomécanique cochléaire. Seules les amplitudes des ondes ainsi que des mesures de reproductibilité sont dépendantes de la zone et de l'intensité de stimulation.

S. GALLEGO,' C. MICHEYL,' C. BERGER-VACHON,' E. TRUY,² A. MORGON² and L. COLLET'

From 'CNRS URA 1447, and ²Department of Otorhinolaryngology, Edouard Herriot Hospital U, Lyon, France

Gallégo S, Micheyl C, Berger-Vachon C, Truy E, Morgon A, Collet L. Ipsilateral ABR with cochlear implant. Acta Otolaryngol (Stockh) 1996; 116: 228-233.

Ipsilateral ABR recording technique was developed with the MXM DIGISONIC DX10 cochlear implant, involving, firstly, setting of recording variables with regard to implant and ABR constraints, and, secondly, enhancement of recording quality by signal processing. The resulting recording quality then enabled us to characterize ABR latency, amplitude and wave reproductibility according to stimulus intensity and stimulation site (electrode stimulated). The findings agree with those of the literature on contralateral human and ipsilateral animal studies. Waves III and V amplitude increased with stimulus intensity. Waves III and V latency was insensitive to stimulus intensity or site. ABR quality diminished basally. Key words: ABR, cochlear implant, electrical stimulation, human, ipsilateral, wave characteristics.

The electrical stimulus used to evoke ABRs was generated using an MXMe DIGISTIM system connected to a PC computer via a serial port. ABR recording was performed using a NICOLET® PATHFINDER II. This system requires the fitting of 3 electrodes (one reference electrode, one anode and one cathode). The PATHFINDER was triggered externally by the DIGISTIM.

The following recording variables were determined after many preliminary trials, using a cochlear implant model and also taking into account electrophysiological constraints: 3 averagings of 256 sweeps; 100 /AT sensitivity; analog bandpass filtering from 0.28,000 Hz; 10 ms analysis time, 512 point window; 50 kHz sampling frequency.

After recording, ABR signals were processed using different processing stages. First of all, the first 400 fis of the ABR signal were excluded so as to eliminate stimulus artefact. Then, the signal was filtered with a digital bandpass of 300-3,000 Hz.

The 3 x 3 intercorrelation matrix was then computed in order to eliminate the most exceptional curves. Finally, the remaining curves were summed and, using an algorithm based on the detection of peaks and infiection points in the signal (5), wave amplitude and latencies were determined. Wave amplitudes were determined on the basis of the length of the wave duration (6).

Fig. 1 shows typical ABR traces recorded from one
electrode in one patient (SC). The traces were

The recording of ABRs is far more difficult to perform in cochlear implant patients than in non-implanted persons. The amplitude of the electrical stimulation, which is about 10,000,000 times greater than that of the ABR signal, causes saturation of the amplifier. One means of reducing the stimulation artefact consists of contralateral recording: ABRs are recorded in the opposite ear (1). However, another solution could reside in the determination of optimum stimulation variables, which would allow considerable reduction in stimulus amplitude, duration, and leakage across the scalp.

The first aim of the present study was to establish the feasibility of ipsilateral ABR recording; then, once the technique was ready, to elaborate a signal processing technique allowing good-quality ABRs; and finally, to characterize ABRs as a function of 2 variable-stimulation intensity (2) and electrode position (tonotopy) (3).

Three patients implanted with an MXIVI' DIGISONIC DX10 participated in the study. This cochlear implant is a non-conventional auditory prosthesis composed of 15 electrodes (4), the function of which is to replace the peripheral part of the auditory system. These electrodes stimulate the ganglion of Corti and elicit auditory sensations.

X-ray photographe revealed that electrode positioning was roughly the same across patients. Thus, it could be considered that stimulation of a given electrode excited roughly the same frequency region from one patient to another.

`Performance' is percentage recognition on a Lafon word-list (17 words of 3 phonemes each) by lip-reading plus cochlear implant

recorded at 16 different stimulation levels linearly decreasing from a maximum (80 in arbitrary units determined by the equipment i.e. 0.104 p C/cycle) to 0. The maximum was set just below discomfort level.

Fig. 2 shows the latency and amplitude of waves III and V as a function of stimulation intensity. In each patient, ABRs were recorded for 6 to 9 different electrodes. Because of the large amount of data-- about 50 matrices, as shown in Fig. 2 --findings were clustered by patient. The relationships between stimulation intensity and main ABR characteristics (latency, amplitude and reproducibility) were studied using linear regression. The results of regression analyses (mean, slope, correlation coefficient and signifiance level) are shown in Table II.

In order to compare ABR characteristics between patients, data were then normalized and a regression analysis was performed, the results of which are represented by the scatter plots in Fig. 4a and also given in Table II.

Fig. 3 shows typical ABR traces obtained from 9 electrodes in one patient (SC). The electrodes were numbered 1-15, basally to apically. Each curve was obtained by averaging about 8,000 traces. Relationships between electrode number and ABR characteristics were studied by means of regression analyses performed on the data clustered by electrode number. As in the previous section, data were then normalized and averaged across patients and regression variables were computed. The obtained relationships are shown in Fig. 4b and the corresponding variables are given in Table II.

The results obtained in this study demonstrate the
possibility of recording ipsilateral ABRs on a coch-
lear implant (MXM DX10). The good quality of the

Fig. 1. Ipsilateral ABRs from electrode 14 in patient SC for 16 decreasing intensity levels. (The arrow on curve 10 shows patient's subjective threshold).

Table II. ABR variable data according to stimulus intensity and electrode number (position), per patient and average

Variable	Patient	Mean	S.D.	Variation according to intensity				Variation according to electrode number
Variable	Patient	Mean	S.D.	Slope	n	R	p	Slope	n	R	p
L III (ms)	BO	1.90	0.0899	2.60E-4	68	0.059	NS	3.85E-3	68	0.184	NS
L III (ms)	LE	2.30	0.117	8.40E-4	68	0.184	NS	-8.00E-3	68	0.285	<0.02
L III (ms)	SC	1.69	0.0905	-1.62E-3	108	0.455	<0.001	7.30E-4	108	0.030	NS
L III (ms)	Total	1.96	0.0966	-4.90E-4	244	0.123	NS	-7.40E-4	244	0.030	NS
L V (ms)	BO	3.86	0.176	-1.16E-3	68	0.133	NS	-0.0231	68	0.563	<0.001
L V (ms)	LE	4.15	0.190	-1.58E-3	68	0.206	NS	-0.0130	68	0.275	<0.03
L V (ms)	SC	3.74	0.128	-9.50E-4	108	0.190	<0.05	-5.59E-3	108	0.162	NS
L V (ms)	Total	3.91	0.161	-1.12E-3	244	0.171	<0.01	-0.0133	244	0.329	<0.001
A III (pV)	BO	0.362	0.152	3.89E-3	67	0.509	<0.001	0.0150	67	0.419	<0.001
A III (pV)	LE	0.375	0.232	3.87E-3	68	0.411	<0.001	0.0165	68	0.285	<0.02
A III (pV)	SC	0.483	0.337	6.34E-3	106	0.475	<0.001	0.0437	106	0.484	<0.001
AIII (pV)	Total	0.407	0.2668	4.83E-3	241	0.440	<0.001	0.0266	241	0.394	<0.001
A V (pV)	BO	0.312	0.149	2.49E-3	66	0.334	<0.01	0.0117	66	0.341	<0.01
A V (pV)	LE	0.440	0.238	-3.70E-3	68	0.385	<0.001	0.0136	68	0.229	0.060
A V (pV)	SC	0.326	0.178	-3.38E-3	107	0.488	<0.001	0.0202	107	0.423	<0.001
AV (pV)	Total	0.359	0.189	3.14E-3	240	0.409	<0.001	0.0153	240	0.322	<0.001
COR (%)	BO	26.7	29.4	0.769	68	0.527	<0.001	0.954	68	0.139	NS
COR (%)	LE	27.6	30.0	0.802	65	0.649	<0.001	-0.521	65	0.067	NS
COR (%)	SC	30.9	28.6	0.399	109	0.360	<0.001	2.71	109	0.355	<0.001
COR (%)	Total	28.4	29.1	0.551	242	0.464	<0.001	1.28	242	0.173	0.007

Fig. 2. Wave III and V latency [(a) and (6), respectively] and amplitude [(c) and (d, respectively] variation according to stimulus intensity in electrode 14 in patient SC (intensity in arbitrary equipment-determined units).

Fig. 3. Average ABR from 9 electrodes in patients SC (electrode 3 being the most basal, electrode 14 the most apical).

recordings allowed precise measurement of wave III and V amplitude and latency. Thus, these characteristics were able to be characterized as a function of stimulation intensity and electrode number.

The validity of the present results can be assessed by comparing them to those obtained using acoustic stimulation in normally-hearing patients. Although a direct comparison between wave latencies is impossible, due to the fact that the cochlear filtering stage no longer exists in cochlear implant patients, relative latencies can be compared. The differences between wave III and V latencies obtained in this study (1.95 for patient BO, 1.84 for LE and 2.05 for SC) were within the limits usually obtained using acoustical ABRs in normally-hearing subjects (1.85-2.15 ms (7)). Another correspondence between acoustic and electrical ABRs consisted in the fact that ABRs vanished at a stimulation level corresponding to the behavioral threshold.

The relationships observed in the present study between stimulation level and ABR amplitude and latency are in agreement with the data from previous studies (8, 2, 9). Wave III and V amplitudes increased with stimulation level (the correlation was highly significant).

Wave III and V latencies remained more or less constant for low to moderate intensities and then increased when the stimulation intensity reached threshold (from 0 to 80, about 60 ps for wave III and 195 for wave V). Crosscorrelation increased with stimulation intensity.

The relationships observed between electrode position and ABR amplitude and latency are also in

11, 3). The more apical the site, the greater the amplitude and reproducibility of waves III and V,

and the smaller the latencies (10 ps for wave III latency, 200 ps for wave V latency).

The results of the present study are of relevance not
only to physiologists but also to audiologists regard-

ing objective assessment of auditory performances in
the cochlear implanted. These results merit further

investigation in a larger number of patients in order
to obtain a reliable database. There is here a possible

The authors wish to thank those people and institutions that have supported this study: the MXM company, the Hospices Civils of Lyon, the CNRS, the University of Lyon, the three patients.

Fig. 4. Variation in wave III and V latency and amplitude and in ABR correlations according to stimulus intensity (a) and electrode number (position) (b).

L'utilisation des filtrages numériques post-traitement permettent d'améliorer considérablement les PEAP lors d'une stimulation acoustique (Grônfors, 1994). Dans le cas d'une stimulation électrique, ces filtrages spéciaux ne peuvent pas être utilisés car les artefacts dus à la stimulation électrique viennent détériorer le tracé. Comme le montre la figure 93, certaines fois, il n'est pas nécessaire d'utiliser un post-traitement des tracés. Cela n'est malheureusement pas toujours le cas. Afin d'utiliser les PEAEP pour la clinique et particulièrement pour le réglage de l'implant cochléaire, la qualité des tracés doit être suffisante et stable pour que l'on puisse en extraire les latences et amplitudes des ondes, même celles très précoces.

Pour cela, nous avons développé un traitement des courbes adapté aux contraintes de la stimulation électrique lors du recueil des PEAEP.

Contrainte Psycho-Physiques et Electrophysiologiques sur le codage de la stimulation électrique chez les sujets porteurs d'un implant cochléaire

Article 13 :

IPSILATERAL ABR WITH COCHLEAR IMPLANT