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

Article 15 :

EFFECT OF STIMULATION INTENSITY AND INTRACOCHLEAR SITE ON ELECTRIC AUDITORY
BRAINSTEM RESPONSES IN HUMAN USING A MULTICHANNEL COCHLEAR IMPLANT WITH A
VARIABLE DURATION PULSE

S. Gallégo, J Durrant, E. Truy, C Berger Vachon, L. Collet
Article soumis

Dans un premier temps, l'article décrit un protocole qui permet d'étudier les PEAEP en fonction de l'intensité et du site de stimulation électrique de la cochlée. Après avoir démontré l'intérêt de l'utilisation de la durée de l'impulsion pour coder l'intensité, l'article montre qu'il est possible d'obtenir des PEAEP sur tous les sujets testés et que les caractéristiques des ondes (II, III et V) sont similaires à celles obtenues avec d'autres types d'implants cochléaires.

L'étude statistique montre que l'amplitude des ondes croit avec l'intensité de stimulation (codée par la durée de l'impulsion) et croit lorsque l'on se dirige vers des stimulations apicales.

Par contre les latences sont des paramètres peu dépendants de l'intensité de stimulation (en durée de l'impulsion).

La latence de l'onde V (et l'intervalle III-V) varie en fonction du site de stimulation. Elle décroît légèrement lorsque l'on se dirige vers des stimulations apicales.

L'estimation par les PEAEP des fonctions de sonies des sujets implantés et en particulier des seuils de détection et de confort est un sujet intéressant pour l'aide au réglage de l'implant.

EFFECT OF STIMULATION INTENSITY AND INTRACOCHLEAR SITE ON
ELECTRIC AUDITORY BRAINSTEM RESPONSES IN HUMAN USING A
MULTICHANNEL COCHLEAR IMPLANT WITH A VARIABLE DURATION PULSE.

S Gallégo (1,2), JD Durrant (1,3), E Truy (4), C Berger-Vachon (1), L Collet (1,4).

1- UPRESA 5020 CNRS Laboratory

2- MXM Laboratory

3-Departments of Communication Science and Disorders and Otolaryngology, Pittsburgh

4- ORL Dept, Ed. Herriot Hospital

ABSTRACT : The Electrical Auditory Brainstem Response (EABR) elicited via a cochlear implant is an interesting tool for analysis of retrocochlear mechanisms, i.e. the impontine brainstem. In this paper, EABRs were studied with respect to stimulation intensity and the excited part of the cochlea. The parameter used to control the stimulus intensity was the pulse duration. Latencies were found to be rather stable, which was not seen in all previously reported studies (namely variations of waves III and V). This tact can be explained by the shape of the stimulating pulse. By judicious choice of the stimulus waveform, synchronization of the fibers is enhanced by making it more difficult for firing on the negative phase of the pulse. Amplitudes of waves II, III, and V were slightly increased with the stimulation intensity. EABR latencies and intentais (mostly wave V and III-V interval) decreased from base to apex. This might reflect a compensatory mechanism to the propagation delay from the base to apex in a normal cochlea. The amplitudes of wave II, III, and V also were increased from base to apex; this may be linked to cells'density along the cochlea which progressively increases from the base to the end of the first cochlear turn.

Key words: Cochlear implant, EABR, characterization, pulse duration, level effect, site effect.

INTRODUCTION

Auditory Brainstem Responses (ABRs) recorded, on the surface of the scalp, presumably reflect (primarily) the propagation of nerve action potentials through the lower auditory pathway, involving multiple synaptic delays, decussation(s) of fibers, and discontinuities of the volume conductor. First described in the literature in 1967 (Sohmer and Feinmesser, 1967), the ABR comprises up to 7 waves with in a time window of approximately 10 milliseconds (Jewett and Williston, 1971). Each wave cornes from the activity of one or more generators distributed along the afferent auditory system, from the cochlea to the middle geniculate body (Moore, 1987a,b). The electrically elicited ABR has met with increased interest. The Electric Auditory Brainstem Response (EABR) has been studied in animais and humans alike and compared systematically to acoustically evoked ABRs (e.q. see Gyo and Yanagihara, 1980 and Starr and Brackman, 1979, respectively). Wave latency has been found to vary little with stimulus intensity, unlike the acoustic ABR. The acoustically elicited response is well-known to demonstrate a pronounced latency-intensity function which, in turn, is considered to reflect the influence of cochlear propagation (Don and Eggermont, 1978, 1993, 1994, Gorga et al, 1988). In any event, the place encoding mechanism precludes scrutiny of place-specific stimulation without the influence of wave-propagation-based latency effects. The EABR potentially allows the characterization of the activation of the auditory system without such `distortion'.

The development of cochlear implant (House et al, 1976; Michelson, 1971; Simmons, 1966) opened the door to the study of the human EABR since the implanted electrode can be used as a stimulator (Starr and Brackman, 1979). Van den Honert and Stypulkowski, 1986 characterized the effects of the stimulus intensity in patients fitted with a single-channel cochlear implant. Results showed that EABRs are generated in a similar manner to acoustical ABR with no significant difference found for the interpeak intervals 11411, III-V, II-V. Recent comparisons between acoustically and electrically evoked ABR - in the same patients with brainstem implant (stimulation at the level of the cochlear nucleus) confirms this result for wave III, and V (Waring, 1992, 1995). Nevertheless, authors emphasize the technical difficulties of recording the EABR. First, electrical stimulation can introduce myogenic and facial nerve responses (Fifer and Novak, 1990; Van den Honert and Stypulkowski, 1986;

Waring, 1992). Furthermore, the stimulation artifact itself is very difficult to remove from EABR recordings. Nevertheless, studies performed in animais (Van den Honert and Stypulkowski, 1984, 1986) have shown correlations between EABR and ABR recordings and revealed the relation between the number of surviving cells in the spiral ganglion and EABR magnitude (Hall, 1990), as had been assumed in the human (Brightwell et al, 1985; Smith and Simmons, 1983).

The arrivai of a new generation of cochlear implants (Clark et al, 1981; Eddington, 1980) made possible EABR recordings utilizing place-specific stimulation (Abbas and Brown, 1988), albeit constrained by the spread of electrical excitation due to the transmission-cable-like electroanatomy of the cochlea. Results, however, were not remarkably different when different parts of the cochlea were stimulated (Abbas and Brown, 1991a). In any event, it proved difficult to reliably record EABR for basal stimulation of the cochlea (Shallop et al, 1993), and it appeared that wide intervals on the cochlea were being stimulated (Abbas and Brown, 1991a) or for recording problem (locking of the fibers and signal to noise ratio) to study only wave V and not waves II and III (Miller et al, 1993).

In these studies, the pulse amplitude was adjusted to modify the stimulation intensity. Consequently, when the intensity of electrical current was raised, the portion of the cochlea stimulated also was increased, i.e. via the spread of excitation. It is a wellestablished property of neurons in general (see Colombo and Parkins, 1987, for review and model) that stimulus duration also can used to increase the stimulation intensity. This has also be demonstrated specifically for auditory neurons (Pfingt et al, 1991). The inherent advantage of manipulating stimulus pulse duration, versus intensity, is that the length of the stimulated interval on the cochlea can be held constant. The use of the pulse duration to modify the stimulus intensity allows better phase locking of the auditory nerve fibers, at least when stimulus intensity is kept low. In the case, the shorter the duration of the pulse is, the better the synchronization becomes. Poor locking of the fibers in very low loudness, when constant duration pulse are used, can explain the poor correlations seen between the perceptual threshold and EABR appearance in human cochlear implantees (Shallop et al, 1991; Mason et aI, 1994; Brown et al, 1994) and in the cat (Smith et aI, 1994).

The purpose of this report is to characterize in detail the EABR obtained in such subjects with short pulse duration, generally below 100 ps, according to stimulus intensity (i.e. duration, with amplitude held constant) and the putative cochlear segment excited. This method of stimulation (i.e. short duration with amplitude constant) enhances EABR recording (measure of waves II, III and V) by better synchronization.

MATERIAL & METHODS Subjects

Eleven implantes, fitted with the DIGISONIC cochlear implant (CI), participated in this study (8 males and 3 females, aged from 3.5 to 69 years).

The French DIGISONIC MXM cochlear implant

The DX10 DIGISONIC (Beliaeff et al, 1994, Chouard et al, 1995, Gallégo et al, 1997c) is a 15-channel, transcutaneous, cochlear implant device made by the French firm, MXM (06, Vallauris). It comprises an external and an internai part. The external part incorporates the emitting antenna which is juxtaposed the internai receiving antenna (i.e. forming the transcutaneous link between the two parts). The internai part is thus totally implanted. It is composed of the receiver package and a 15electrode array. The electrodes (0.5 mm) are equally spaced (0.7 mm) along the first turn of the cochlea and typically are situated from 5 to 20 mm, with respect to the round window. Indeed, performance of patients implanted with this 15-channel device demonstrates its effectivenessin stimulating the neurons of the spiral ganglion, namely (for interest of this report) to permit them to accurately scale pitch (Truy et al, 1995).

The mode of stimulation is common ground (Gallégo et al, 1997d,e) --the activated electrode goes positive initially white the others are set to ground. Capacitive coupling ensures a net zero current. Pulse duration ranged from 5 to 310 ps, but was generally bellow 100 ps. The stimulating current ranged from 0.25 to 3 mA but was generally below 1 mA. These specifications (possibility to modify pulse duration or

pulse amplitude) make this device adapted to EABR recording. As shown in figure 1, by the 'in vitro' recording of a stimulus pulse (i.e. in normal saline solution), the stimulus pulse demonstrates a large asymmetry. Adjustment of the coupling capacities, however, permits perfect equalization of positive and negative charges, averting the problem of electrolysis of the metallic electrode in an ionic solution (Gallégo et ai, submittted a). Most of the fibers are synchronized by the positive phase of the stimulation, facilitated also by the sharp onset (Rattay and Motz, 1987). The response to the negative phase (usually below 100 ps) is minimal (Moxon, 1971; Javel et al, 1986, 1987). This is due to the asymmetry, specifically the short duration of the positive phase, which is much less than the nerve refractory period (Abbas and Brown, 1991b; Kasper, 1991).

			C=150 nF
_IL								R=1 kohm

Figure 1: Electrical model of the DIGISONIC stimulation inside the cochlea. Cochlea was represented by a resistance only (Clopton and Spelman, 1982); the admitted value was about 1 kohm. The voltage average value, on the resistance, is equal to zero. In vitro measurements showed that positive and negative phases of the stimulating wave were not symmetrical. Fibers are likely to be mostly synchronized by the positive phase of the wave.

In order to perform EABR recording, special equipment (DIGISTIM) was developed by MXM which supplants the external part of the implant. DIGISTIM is powered by batteries, PC controlled through an optoisolated serial port. This system generates pulses with adjustable parameters (electrode selection, pulse amplitude, pulse duration, and stimulation frequency). It is also possible to synchronize an external device, such as an evoked response test system.

EABR recording

A commercial evoked response measurement system (Nicolet Path Finder II was used to record the EABR. The recording montage was as follows: forehead at hair line connected to the noninverting input of the recording amplifier; earlobe connected to the inverting input; the contralateral earlobe connected to ground. Recording parameters were similar to those utilized for conventional ABR recordings. Full scale sensitivity was +1- 50 pV. To minimize distortion of the tracing due to the stimulation artifact, a wide bandpass (0.2-8000 Hz) was employed (Van den Honert and Stypulkowski, 1986). The analysis window was 10 ms (512 points). The sampling frequency was more than twice the maximum frequency seen in the ABR, avoiding aliasing, and allowed an efficient rejection of the noise (Gronfors and Juhola, 1995). Each time-ensemble average derived from over 512 repetitions of the stimulus. In order to assess the reproducibility of the response averages, three replications were made per stimulus condition. Stimulation repetition rate was 60 Hz. (Note: the line frequency in France is 50 Hz; in this case, therefore, the repetition rate discouraged, rather than encouraged, phase coherence with line-frequency noise.). The frequency of stimulation also is elevated ( relative to pates that optimize ABR waveform definition under acoustic stimulation) but, contrary to the situation for acoustic ABR, this does not pose a significant problem. Adaptation under electric stimulation actually is more likely a rate whose period is that of neural refractory period (Abbas and Brown, 1991; Kasper, 1991). Each electrode was tested over 16 stimulus magnitudes (i.e. pulse durations) in descending order from the most comfortable. Let Min be the perception threshold, and Max, the comfortable magnitude (at 300 Hz). The stimulus was never over 1.5x(Max-Min)+Min. When the 16x3 recordings were completed, the data were transferred to a floppy disk, after ASCII conversion. The data were fed to a personal computer for further processing.

Digital filtering and signal processing

To improve the quality of the recorded curves several digital-signal processing were algorithms were applied. The details of the filter function applied are proprietary (MXM), but the principles of design of the filter and its performance are described by - Gallego et al- (1997b, submitted b). A hrigh-quality EABR recording is desirable for

the most reliable measurements of latencies and amplitudes. In order to evaluate EABR reproducibility (for a given electrode and stimulation intensity) a 3x3 crosscorrelation matrix was constructed with the time-ensemble averages. This provided a statistical basis by which reject a response which appeared to be an outiiner, relative to the others (i.e. for a given subject and condition). Criteria for selection and other details are discussed below. Following analysis of reproducibility, a grand average was computed from the accepted (individual) averages.

Literature	LII (ms)	LIII (ms)	LV (ms)	dll-Ill (ms)	dIII-V (ms)
Van den Honert and	1.20	2.10	4.09	0.95	1.83
Stypulkowski, 1986	(s.d. 0.14)	(s.d. 0.15)	(s.d. 0.26)	(s.d. 0.16)	(s.d. 0.17)
Abbas and Brown, 1988	1.36 (s.d. 0.19)	1.99 (s.d. 0.23)	3.99 (s.d. 0.37)	0.63	2.00
Kasper et al, 1992	1.38	2.16	3.94	0.75	1.79
	(s.d. 0.09)	(s.d. 0.18)	(s.d. 0.22)		(s.d. 0.17)
Gallégo et al, 1996	NG	1.96	3.91	NG	1.95
		(s.d. 0.10)	(s.d. 0.16)

Table I: Average values and standard deviations in EABR latencies (Van den Honert and stypulkowski, 1986; Abbas and Brown, 1988; Kasper et al, 1992 ; Gallégo et al, 1996). NG=not given.

Assessment of EABR reproducibility is important in deciding whether or not a given wave is present. ln conventional/clinical ABR assessments, reproducibility typically is based upon a visual comparison of test and retest averages (Arnold, 1985), but this method subjective is completely. To augment the decision process, namely by supplementing visual inspection with a statistical tool, cross-correlation coefficients were calculated, and zero-crossing of the first derivative of the waveform was used for picking peak per se. Thus involved several rules or constraints based upon preliminary studies (Gallégo et al, 1996, 1997a) and results obtained by other EABR researchers (Van den Honert and Stypulkowski, 1986; Abbas and Brown, 1988; Shallop et al, 1991; Kasper et al, 1992). Collectively, these results led to the rules for the detection and evaluation of waves II, III, and V latencies and interpeak intervals (table I), as follows: the latency of wave II must be from 0.8 to 1.6 ms; the latency of wave III, between 1.5 to 2.5 ms; latency of wave V, from 3.4 to 4.5 ms; interval 11411, from 0.6 to 1.3 ms; interval III-V, from 1.4 to 2.2 ms. Finally, wave amplitudes were

measured peak-to-peak, from a given vertex-positive peak to the following negative peak or through.

Figure 2 shows examples of response averages that appear to be highly reproducible upon visual inspection. Indeed, latencies of waves Ila, Ilb, III, and V appear to be essentially invariant across test runs. The 3x3 cross-correlation matrix confirms the visual impression. Wave amplitudes also are significantly correlated (p<0.001) if the cross-correlation coefficient is higher than 0.2. Intercorrelation is over 0.2 (p<0.001) and the following values have been obtained: 1-2:0.911; 1-3:0.862; 2-3:0.847. The mathematical processing shows a slight shift between the three curves: 1-2:Ops; 1- 3:20ps; 2-3:20ps. Consequently visual examination and mathematical processing lead to a reliable and objective assessment of EABR reproducibility.

0 1 2 3 4 5 6 7 8

Figure 2: EABR reproducibility with three identical stimulations. Intercorrelation and time shift between
the curves are: 1-2: 0.911, dT 1-2: 0.00ms ; 1-3: 0.862, dT, 1-3: 0.02ms ; 2-3: 0.0.847, dT 2-3: 0.02ms.

Amplitude- Intensity Trade-off

Intensity of stimulation, again, can be varied by adjusting amplitude and/or duration of the stimulus pulse (Columbo and Parkins, 1987). These parameters can be controlled independently by the DIGISONIC (fig 3a). Figure 3a shows EABRs evoked by pulses eliciting the same loudness percept in the test subject. In this case, several combinations of amplitudes and durations; the EABRs are quite similar. However,

when amplitude alone is adjusted (figure 3b), an additional wave at 5 ms can be seen with pulses of high amplitude (short duration). In figure 3c, EABRs are plotted for different pulse durations. In figure 3b, no added waves are seen for high intensities. (Note: Loudnesses were the same for corresponding plots of figures 3b and 3c.)

4

3

0 I 2

b:

8

8

3

7

1 2

C.

a:

Figure 3: (3a) Stimulation loudness was equivalent for the 4 curves; with different amplitude and pulse duration. (3b) Thresholds for several stimulation levels (the amplitude of the pulse was modified; stimulation decreased from 1 to 4). (3c) Thresholds for several stimulation levels when the duration of the pulse was modified. Stimulation had the same loudness for corresponding recordings on 3b and 3c (eg 1 and 1).

LII

LIII
LV
d11-111

dIII-V

dll-V
dllp-Iln

dIllp-Illn

dVp-Vn

Ail
A111
AV

Mean

	Std. dev.	1	2	3
1.278	0.172	NS	NS	NS
2.050	0.180	NS	NS	NS
3.864	0.276	***	NS	NS
0.772	0.062	***	NG	NG
1.814	0.139	NS	NG	NG
2.586	0.156	NG	NG	NG
0.340	0.026	NG	NG	NG
0.571	0.075	NG	NG	NG
1.074	0.102	NG	NG	NG
0.249	0.099	NG	NG	NG
0.442	0.157	NG	NG	NG
0.590	0.163	NG	NG	NG

Table II: EABR mean values and standard deviations on 11 subjects, and comparison with other studies (1:Van den Honert and Stypulkowski, 1986; 2:Abbas and Brown, 1988; Kasper et al, 1992). In same cases, our results were compared with other findings using a comparison of mean values (NS: difference being not significant; ***: statistical difference p<0.05; NG: not measured).

RESULTS

EABR--General Characteristics

Figure 4: EABR recordings on 11 subjects. Stimulation was on the most apical electrode in the cochlea. Waves II, III and V can be seen for each subject. Y-scale was changed for each subject.

EABRs were recorded successfully for all implantees and all the electrodes leading to an auditory sensation. Ail in all, 58 electrodes were considered, and more than 1000 EABR recording were examined. Figure 4 gives an example of the typical EABR recording obtained. Ail EABR demonstrated similar characteristics with peak latencies as follows: wave II from 0.97 ms to 1.53 ms; wave III from 1.80 and 2.36 ms; wave V from 3.50 and 4.32 ms. An 11x11 correlation matrix of the recordings shown on figure 4 indicates that, apart from some shifts between the subjects, EABRs turn out to be similar across implantees.t If subject SC is Leen as_ a_

reference, following shifts and correlations have been observed: LA:dT=40ps, R=0.83; DN:dT=20ps, R=0.93; DE:dT=260ps, R=0.77; BO:dT=300ps, R=0.89; RO:dT=300ps, R=0.94; LE:dT=580ps, R=0.74; MO:dT=400ps, R=0.86; MA:dT=340ps, R=0.54; FA:dT=700ps, R=0.91; RI:dT=650ps, R=0.65.) Thus, EABR recordings appeared to be reliable and similar among subjects (R>0.2; p<0.001). Average values of latencies, amplitudes, and intervals between wave II, III, and V are indicated on table II. A comparison with the values obtained by others researchers did not show significant differences (Van den Honert and Stypulkowski, 1986; Abbas and Brown, 1988; Kasper et al, 1992).

Effects of stimulation lntensity (Duration)

2

3

4

5

6

7

8

Figure 5: Recordings for one subject: 16 decreasing stimulation intensities were taken. The arrow shows the patient's subjective auditory threshold. Stimulation intensity was controlled by the pulse duration.

In this study again, the primary parameter of stimulus intensity of interest was the
pulse duration. Figure 5 provides exemplary data. EABRs could be seen with a very

low stimuli. EABR wave latencies seemed not to be affected by the stimulus intensity. On the other hand, amplitudes were closely linked to the stimulus intensity, as expected. Indeed, the visual detection level of the EABRs was found to correspond well with the threshold of the auditory perception of the stimulus (indicated by the arrow in Fig. 5).

150 200

100

52

3.5 3.0 2.5 2,0 1.5 1.0 0.5

0.0

0

Fig 6 b:

0.5

2.0

0.0

1.5

1.0

0.8

0.6

0.4

0

0 53

Fig 6 a:

1.2

0.2

1.0

0.0

6

5

3

4

2

"le

100

150 200

100 150 200

53

Intensity (us/cycle) Fig 6 f:

4.0

·

0.0

0

0 50

Fig 6 d:

100

150

200

3.0

2.5

2.0

1.5

1.0

0.5

0 50 1C0 150 200

Fig 6 c:

2.0 1.5 1.0 0.5 0.0

52 100 150 200

Intensity (us/cycle) Fig 6 e:

Figure 6: Subject BO: EABR modification with the stimulation intensity given in ps.

(6a) wave II, III, and V latencies.

(6b) d11-111, dIII-V, and dll-V intervals.

(6c) wave II amplitude.

(6d) wave III amplitude.

(6e) wave V amplitude.

(6f) PII-N11, PIII-N111, PV-NV intervals.

Evolution of different EABR parameters, with respect to the stimulation intensity, for a given subject (BO) and on several electrodes is represented on figure 6. Below, p (positive) will indicate the (positive) peak of the wave and n (negative) the valley. Absolute latencies and interpeak intervals were nearly independent of the stimulus magnitude, particularly Ilp-Iln and Illp-Illn. The Vp-Vn interval was observed to increase slightly with increased stimulus intensity.

For each subject, we carried out multiple-linear regression analyses among EABR parameters as a function of stimulus intensity and electrode site. The second column of Table III provides the means and standard deviation (latencies, interpeak intervals and amplitudes) for a change of 100 ps in pulse duration (corresponding to the approximate average dynamic range of stimulation). The third column indicates which values are significantly different from zero. It can be seen, for intense, that lengthening of 100 ps of the stimulus pulse led to decrease of 106 ps of Pll latency. The changes in amplitude for wave II, latency for wave II, and interpeak intervals II- III, III-V, and II-V and not statistically significant. A significant increase of the amplitude of wave III, and V also was seen. Latencies of wave III, & V slightly, but significantly decreased with the stimulation (-106, -200, and -156 ps respectively).

	Effect of stimulation Intensity		Effect of Intracochlear Site
	Variation per 100 psof pulse duration	p	Variation per 15 mm of cochlea	p
LII (ms)	-0.212	NS	0.026	NS
LIII (ms)	-0.400	0.023	-0.022	NS
LV (ms)	-0.312	0.013	-0.181	<0.001
dl1-111(ms)	-0.106	0.053	-0.047	NS
dIII-V (ms)	0.008	NS	-0.157	0.004
dll-V (ms)	-0.102	NS	-0.161	NS
All (pA)	0.582	0.087	0.166	0.019
AIII (pA)	0.848	0.019	0.363	0.040
AV (pA)	0.912	0.011	0.366	0.075

Table III: Columns 3 and 5: Means and standard deviations (latencies, interpeak intervals and amplitudes) by multiple linear regression for a change of 100 ps in pulse duration and for a change of 15 mm in stimulation site. Columns 4 and 6: Comparison from zero of EABRs variations.

Effects of Stimulated part of the Cochlea

Figure 7: EABR recordings for several electrodes stimulated with the same loudness, for 2 patients. Electrode 1 is basal and electrode 15 is apical.

Figure 7a and 7b demonstrate typical trends in the effects of stimulation electrode/channel, for stimili judged to have the same loudness. The most basal electrode is 2 and the most apical is 14. It can be seen that EABR waveforms were not modified by electrode number or, in effect, stimulus site along the cochlea. Also, wave II, III, and V latencies were similar regardless electrode activated. However, EABR amplitudes were systematically increased when the electrode was moved from base to apex. Data presented in Figure 8 serve to characterize the effects electrode number, for all patients' latencies, amplitudes, and interpeak intervals for waves (II, III, and V). Each point represents the average of the values obtained for different stimulation intensities, (one subject per electrode). The following additional trends are evident:

-wave II latency was not modified with respect to the stimulated zone (fig 8a); -wave V increases from apex to base (fig 8a);

-intervals dIII-V and dll-V decreased with stimulation Gloser to the apex (fig

8b),

-wave II, III, and V amplitudes increased from base to apex (fig 8c, 8d, 8e), as noted before;

-PN intervals for waves II, III, and V were, roughly constant (fig 8f).

dll-V	3.5 3.0
2.5
				·
2.0		: ·
		· .
1.5		:	·			dIll-V
1.0
0.5
0.0
		6 8 10	12 14	16
0
	Fig 8 b:
1.4	- AIII
1.2
1.0
0.8		.	· ·
0.6		. · · .
	· a		· .	·
0.4
	·	. . · ·
0.2
0.0
		6 8 10	12 14	16
Fig 8 d:
2.0
		O PIII-NIII	O ru-sli
1.5
1.0
	· :	· · ·	·
	· ·	· :	·	;
0.5	- · ·	: ;	· · ·	·
	· :	· :		1
			· ;
	0.0

0 2 4 6 8 10 12 1⁴ 15

Electrode number

16

2 4 6 8 10 12 14

0.0

0 2 4

Fig 8 C:

2.0

6 8 10 12 14 16

Fig 8 a:

0.8

0.6

0.4

0.2

1.6

1.2

0.8

0.4

0.0

0 2 4 6 8 10 12 14 16

Electrode number

.								.	·
: ·	:
	i ·
·	· :	·	. ·	:	· '	:		- .
		.
		.	.					° ° · 8
°	· :			°	: '	!
	i 8	. .			· · .	i
· 8	° .

0.5

0

4.5

4

Pt' 4.0

3.5
3.0
2.5
PM 2.0
1.5
1.0

.tV

Fig 8 e: Fig 8 f:

Figure 8: EABR values, with respect to the electrode, for all patients. (8a) waves II, III, and V latencies. (8b) d11-111, dIII-V, and dll-V intervals. (8c) wave 11 amplitude. (8d) wave III amplitude. (8e) wave V amplitude. (8f) PII-NII, PIII-N111, PV-NV intervals.

Here too linear regressions were calculated. There fourth column of table III shows the mean and standard deviations of the EABR measures (again, as a function of stimulus duration and electrode site), for a change of 15 mm in stimulation site (corresponding to the depth of penetration of the electrode array. The fifth column indicates which resuits were significantly different from zero. It can be seen, for instance, that lengthening of 15 mm of cochlea led to statistically significant increase

of 26 ps of Pli latency. A significant increase of the amplitude of wave II, III also was seen. Latencies of wave V and interval of waves III-V significantly decreased with the stimulation site(-181, -157 ps). However, the changes of amplitude of wave V, latencies of waves II and III, and interpeak intervals of waves II-III and II-V were not statistically significant.

DISCUSSION

Latencies, amplitudes and intervals measured from waves II, III, and V observed in this study are in good agreement with the values obtained by other researchers (table I). EABRs are composed of at Ieast 3 waves occurring within a time shorter than 6 ms. Wave II, with its bifid shape (Van den Honert, 1986), was seen in all the subjects. Wave Ila is probably wave I because its latency is similar to the auditory nerve compound action potential latency (Brown and Abbas, 1990; Gantz et al, 1994). Wave III was observed to have a relatively high amplitude and short duration. The IV- V complex was wider than the other waves, as characteristic of the acoustically evoked response. It has been suggested (Abbas and Brown, 1988, 1991a) that the stimulation of the basal part of the cochlea triggers a wide spread excitation. This was not seen with our technique.

Although EABRs stimulated in base of the cochlea are difficult to record (Shallop et al, 1993), it was possible in this study to obtain and label EABRs for all electrodes (within and across subjects) capable of eliciting an auditory sensation. Consequently waves II, III, and V latencies, amplitudes, and intervals were measured. We believe that the high quality of our EABR recordings was due to the shape of the stimulation pulse (asymmetry, sharp positive phase and short duration pulse for low-level), leading to good synchronization, and our method of digital processing of the data, leading to robust suppression of artifacts.

Average values calculated in our study were in agreement with those given by other investigators. However, they may be far different from usual values measured on healthy subjects as it seems likely that wave latency reflects the effects of limited populations of primary auditory neurons and/or auditory deprivation prior to implantation (Gallégo et al, submitted c). For healthy subjects, waves latencies should be close of those seen for SC, DN, and LA (L11=1.10ms, 1111=1.85 ms,

LV=3.55 ms). Regardless of the variability among subjects, EABR shapes were reasonably stationary in time; intersubject correlation was always higher than 0.2 (p<0.001).

Choice of Stimulus Parameters and Effects of Stimulus Intensity/Duration

As stated earlier, electrical stimulation can be controlled by two means, pulse amplitude (figure 3b) and duration (figure 3c). Amplitude is the most commoniy used parameter in examining the EABR. The problem with adjusting amplitude is that at high intensities, the spread of excitation within in cochlea effectively increases and may even corne to involve the vestibule and facial nerve. The use of the pulse duration bears several advantages, as suggested earlier and verified by the results reported here. The results in Figure 3 demonstrate the comparability of results obtained with stimuli of the same intensity, regardless of parameters. However, for EABR, when amplitude is high and duration short, a wave, probably myogenic (Fifer and Novak, 1990), is seen with a latency of 5 ms approximately (e.g. see figure 3b). Consequently, the use of pulse duration, rather than magnitude to control stimulation intensity is likely to improve the quality of the recording and to simplify the interpretation. The results suggest further that, ineed, using pulse duration, the stimulated zone on the cochlea remains constant. Even with low-intensity stimulation (i.e. decreased duration), synchronization is facilitated (fig. 3c).

Several authors (Allum et al, 1990; Abbas and Brown, 1991a; Shallop et al, 1990, 1991, 1993) have found an increase in latency for low levels of stimulation; others (Van den Honert and Stypulkowski, 1986; Kasper et al, 1991, 1992) found latencies to be essentially invariant. None of the studies considered pulse duration in order to control the stimulation. In our work, stimulation duration had an average of 54.5ps (s.d. 36.25ps). The mean range was around 100 ps. We observed a constant of latencies and interpeak intervals with stimulation intensity (i.e. ranges were -212ps=- 16.5% for wave II, -400ps= -19.5% for waves Ill, and -312ps= -8.1% for wave V).

Still other investigators have considered the influence of shape of the stimulus pulse
(Rattay, 1987), the evoked response elicited by the negative phase (Moxon, 1971;
Javel, 19861 and refractory period (Abbas and Brown, 1991 b; Kasper et al, 1992).

Rattay showed that the steeper the slope of the electrical pulses, the better the synchronization of the 8th nerve response. Both Moxon and Javel noted that when the electrical stimulation biphasic pulse was wide, auditory neurons could respond on the second phase. Durational effects also relate to the refractory period of neurons. One of the consequences of this phenomena is that EABR thresholds were correlated to perceptual thresholds (Gallégo et al, 1996,1997a; Truy et al, 1997a,b) than with other stimulation strategies (Shallop et ai, 1991, 1993; Brown et al, 1994; Mason et al, 1994; Smith et al, 1994). Propagation speed appeared to be increased for strong stimulation, but the variation is compared to ABR amplitude, and intervals between the waves turned out to be unrelated to the stimulation intensity.

Effect of Cochlear Region of stimulation

It has been noted (Black et al, 1983; O'Leary et al, 1985) that with bipolar stimulation, the stimulated segment on the cochlea can be very small (voltage distribution is divided by three for a distance from 2 to 4 mm). Studies in animais show a high correlation between the response and the frequency-place excitation parameters in the inferior colliculus (Black and Clark, 1980). Variation of latency, interpeak intervals and amplitudes of waves II, Ill, and V versus the stimulated electrode were measured in the present study (fig 6). The latency of wave II was not found to be dependent upon the stimulated zone. Therefore there is no tonotopic influence upon the timing of neural discharges reaching (i.e. unlike the case of acoustic ABR). This result is in agreement with previous histological studies (Hinojosa et al, 1985; Spoendlin et al, 1972, 1988 ; Moore, 1987a ; some extend, Spoendlin, 1989) which failed to show any substantive/systematic variations in nerve fiber diameter as a function of tonotopic origin in the cochlea. This also is consistent with electrically stimulated single-unit responses (Kiang and Moxon, 1972).

Generally, wave V latency (in agreement with Miller et al, 1993) and the III-V interval was found to decrease from base to apex. Also, information processing appeared to be faster, from the exit of the cochlear nucleus to the last nuclei of the brain stem, for low-frequency compared to the high-frequency places of stimulation. This apparent increased efficiency of transmission along the brain stem amounted to roughly 0.18 ms (4.7%) for wave V tatencies (cochlea to inferior colliculus) and 0.16 ms (8.8%) for

waves III-V interval (cochlear nucleus to inferior colliculus) for a corresponding distance of 15 mm along the cochlea.

Wave amplitudes in this study were found to be higher with more apicalward stimulation (end of the first turn); width of the waveform, however, was roughly constant. It has been observed that the transfer functions of the auditory nerve fibers is essentially the same from base and apex (Kiang and Moxon, 1972). The modification of the amplitude cannot be explained by desynchronization of the fibers. An explanation, however, can be offered via results of the following studies: -in the human and in the cat a correlation has been reported between EABR amplitude, and the number of cells in spiral ganglion (Smith and Simmons, 1983; Hall, 1990); -in both the cat (Spoendlin, 1972) and in normal humans (Hinoja et al, 1985; Spoendlin and Schrott, 1988, 1989), histological studies have showed an increase of cell density in the spiral ganglion, from the base to the end of the first cochlear turn. The low amplitude of the waves, at the base, could be directly connected to a lower number of cells on the spiral ganglion. For some subjects, however, the amplitude did not vary monotonically from base to apex. The effect thus may depend on the etiology of deafness, as since some authors noted a correlation between number of spiral ganglion cells and the nature of the auditory loss (e.q. Otte et al, 1978; Schmidt, 1985).

ACKNOWLEDMENTS

The authors acknowledge persons and institutions who supported their work: the MXM company, the CIFRE, the Hospices Civils of Lyon, the CNRS, the University of Lyon, professor Alain Morgon (head of the ORL department), and the eleven implantees who participated in this study.

REFERENCE

Abbas PJ, Brown CJ. Electrically evoked brainstem potentials in cochlear implant patients with multi-electrode stimulation. Hear Res 1988 36:153-62

Abbas PJ, Brown CJ. Electrically evoked auditory brainstem response: Growth of the response with current level. Hear Res 1991a 51:123-37

Abbas PJ, Brown CJ. Electrically evoked auditory brainstem response: Refractory properties and strength duration functions. Hear Res1991b 51:138-48.

Allum JHJ, Shallop JK, Hotz M, Pfaltz CR. Characteristics of electrically evoked 'auditory' brainstem responses elicited with the Nucleus 22-electrode intracochlear implant. Scand Audiol 1990 19:263-7

Arnold SA. Objective versus Visual Detection of the Auditory Brain Stem Response. Ear Hear 1985 6(3):144-50

Beliaeff M, Dubus P, Leveau J-M, Repetto J-C, Vincent P. Sound Signal Processing and

Stimulation Coding of the DIGISONIC DX10 15-Channel Cochlear Implant. Advances in Cochlear Implant Ed Hochmair-Desoyer IJ & Hochmair ES Innsbruck 1994 198-203

Black R, Clark GM. Differencial electrical excitation of the auditory nerve. J Acous Soc Am 1980 67:868-74.

Black R, Clark G, Tong Y, Patrick J. Current distributions in cochlear stimulation. Ann NY Acad Sci 1983 405:137-45

Brightwell A, Rothera M, Conway M, Graham J. Evaluation of status of the auditory nerve: psychophysical test and ABR. Eds RA Schindler & MM Merzenich Cochlear Implants Raven Press New York 1985 343-9

Brown CJ, Abbas PJ. Electrically evoked whole nerve action potentials: Data from human cochlear implant users. J Acoust Soc Am 1990 88(3):1385-91

Brown JB, Abbas PJ, Fryauf-Bertschy H, Kelsay D, Gantz BJ. Intraoperative and Postoperative Electrically Evoked Auditory Brain Stem Responses in Nucleus Cochlear Implant Users: Implications for the Fitting Process. Ear Hear 1994 15:168-76

Bruce JG, Brown CJ, Abbas PJ. Intraoperative measures of electrically evoked auditory nerve compound action potential. Am J Otology 1994 15(2):137-44

Busby PA, Tong YC, Clark GM. Electrode position, repetition rate, and speech perception by early- and late-deafened cochlear implant patients. J Acoust Soc Am 1993 93(2):1058-67 Chouard CH, Ouayoun M, Meyer B, Fugain C. Speech Coding Strategies of the DIGISONIC Fully Digitized Cochlear Implant. Acta Otolaryngol (Stockh) 1995 115:264-8

Clark GM, Tong YC, Martin LFA. A multiple-channel cochlear implant: an evaluation using closed-set spondaic words. J Laryngol 1981 95:461-4

Clark G, Shepherd R, Franz B, Dowell R, Tong Y et al. The histopathology of the human temporal bone and auditory central nervous system following cochlear implantation in patient. Acta Otolaryngol (Stockh) 1988 (suppl 448):1-65

Clopton B, Spelman F. Neural mechanisms relevant to the design of an auditory prothesis: location and electrical characteristics. Ann Otol Rhinol Laryngol 1982 91(sup98):9-14 Columbo J, Perkins CW. A model of electrical excitation of the mammalian auditory nerve neuron. Hear Res 1987 31:287-312

Don M, Eggermont JJ. Analyses of click-evoked brainstem potentials in man using high-pass noise masking. J Acoust Soc Am 1978 63(4):1084-98

Don M, Ponton CW, Eggermont JJ, Masuda A. Gender differences in cochlear response time: An explanation for gender amplitude differences in the unmasked auditory brain-stem response. J Acoust Soc Am 1993 94(4):2135-48

Don M, Ponton CW. Auditory brainstem response (ABR) peak amplitude variability reflects

individuel differences in cochlear response times. J Acoust Soc Am 1994 96(6):3476-91 Durrant JD. Fundamentals of sound generation. Bases of Auditory Brain-Stem Evoked Responses (Ed EJ Moore) Grune et Stratton Inc New York 1983 15-50

Eddington DK. Speech discrimination in deaf subjects with cochlear implants. J Acoust Soc Am 1980 51:885-91

Eggermont J. Analysis of click-evoked brainstem potentials in humans using high-pass noise masking.lI. Effect of click intensity. J Acoust Soc Am 1980 68:1671-5

Fifer RC, Novak MA. Myogenic Influences on the Electrical Auditory Brainstem Response (EABR) in Humans. Laryngoscope 1990 100:1180-4

Gallégo S, Micheyl C, Berger-Vachon C, Truy E, Morgon A, Collet L. Ipsilateral ABR with cochlear implant. Acta Otolaryngol (Stockh) 1996 116:228-33.

Gallégo S, Truy E, Morgon A, Collet L. EABRs and surface potentials with a transcutaneous multielectrode cochlear implant. Acta Otolaryngol (Stockh) 1997a 117:164-8.

Gallégo S, Collet L., Berger-Vachon C. Electrically auditory brainstem responses (EABR) : contribution of a filter adapted to the auditory system. J international federation for medical & biomedical engineering, 1997b, 35 (suppl 1), 304.

Gallégo S, Garnier S, Berger-Vachon C. Traitement du signal de parole pour sourds-profonds: l'implant cochléaire. Revue d'électricité et d'électronique. 1997c, 8, 50-53.

Gallégo S, Collet L., Berger-Vachon C. A model of electrical stimulation delivered by the Digisonic cochlear implant. J international federation for medical & biomedical engineering, 1997d, 35 (suppl 1), 300.

Gallégo S, Lu BL, Berger-Vachon C. Modelling of the electrical stimulation delivered by the DIGISONIC Multichannel cochlear implant. Advance in Modelling & Simulation Eng 1997e, in press.

Gallégo S., Frachet B., Truy E., Collet L. Cochlear implant performance and electrically auditory brainstem response characteristics. (in press)

Gallégo S, Beliaeff M, Frachet B, Berger-Vachon C, Collet L. Long-term change in threshold and

comfort levels and dynamics in Digisonic cochlear implant bearers. Submitted a.

Gallégo S, Durrant J, Collet L., Berger-Vachon C. Numeric time-variant filters adapted to the

recording of electrically auditory brainstem responses (E-ABR). Submitted b.

Gallégo S, Truy E, Berger-Vachon C, Collet C. Electrically auditory brainstem responses in cochlear implant assessment: possibility and interest. Submitted c.

Gantz BJ, Brown CJ, Abbas PJ. Intraoperative measures of ellectrically evoked auditory nerve compound action potential. Am J Otol 1994 15:2 137-44

Gorga MP, Kaminski JR, Beauchaine KA, Jesteadt W. Auditory brainstem responses to tone bursts in normally hearing subjects. J Speech Hear Res 1988 31:87-97

Greenwood DD. A cochlear frequency-position function for several species-29 years later. J Acoust Soc Am 1990 87(6):2592-605

Gronfors T, Juhola M. Effect of sampling frequencies and averaging resolution on medical

parameters of auditory brainstem responses. Comput Biol Med 1995 25, 5:447-54

Gyo K, Yanagihara N. Electrically and acoustically evoked brainstem responses in guinea pig.

Acta Otolaryngol (Stockh) 1980 90:25-31.

Hall RD. Estimation of surviving spiral ganglion cells in the deaf rat using the electrically evoked auditory brainstem response. Hear Res 1990 45:123-36

Hinojosa R, Seligsohn R, Lerner S. Ganglion cell counts in the cochleae of patients with normal audiograms. Acta Otolaryngol (Stockh) 1985 99:8-13

House WF, Berliner K, Crary W, Graham M, Luckey R, Norton N, Selters W, Tobin H, Urban J, Wexler M. Cochlear implants. An Otol Rhinol Laryngol 1976 85(suppl 27):1-93

Javel E. Basic response properties of auditory nerve fiber. Neurobiology of Hearing: The

Cochlea. 1986 Ed Altschuster RA, Bobbin RB, Hoffman DW. Raven Press. New York.

Javel E, Tong Y, Shepherd B, Clark G. Responses of cat auditory nerve fibers to biphasic electrical current pulses. Ann Otol Rhinol Laryngol 1987 96(suppl 128):26-30

Jewett DL, Williston JS. Auditory evoked far-fields averaged from the scalp of humans. Brain 1971 94:681-96

Kasper A, Pelizzone M, Montandon P. Electrically Evoked Auditory Brainstem Responses in Cochlear Implant Patients. ORL 1992 54:285-94

Kasper A. Electrically Evoked Activity in Human Auditory System. Thesis 1991 Genève.

Keitley EM. Frequency map of the spiral ganglion in the cat. J Acoust Soc Am 1987 81(4):103642

Kiang NYS, Moxon EC. Physiological considerations in artificial stimulation of the inner ear. Ann Otol Rhinol Laryngol 1972 81:714-31

Leake PA, Snyder RL. Topographic organisation of the central projections of the spiral ganglion in cats. J Comp Neurol 1989 281:612-29

Mason S, Sheppard S, Garnham C, Lutman M, O'Donoghue, Gibbin K. Improving the relationship of intraoperative EABR threshold to T-level in young children receiving the Nucleus cochlear implant. Advances in Cochlear Implant Ed Hochmair-Desoyer IJ & Hochmair ES 1994 44-49.

Michelson RP. Electrical stimulation of human cochlear: A preliminary report. Arch Otolaryngol 1971 93(3):317

Miller CA, Abbas PJ, Brown CJ. Electrically evoked auditory brainstem response to stimulation of different sites in the cochlea. Hear Res 1993 66:130-42

Moore JK. The human auditory brainstem: A comparative view. Hear Res 1987a 29:1-32

Moore JK. The human auditory brainstem as a generator of auditory evoked potentials. Hear Res 1987b 29:33-43

Moxon EC. Neural and mechanical responses to electric stimulation of the cat's inner ear. Thesis. 1971 MIT Cambridge, Mass.

O'Leary S, Black R, Clark G. Current distributions in the cat cochlea: a modelling and electrophysiological study. Hear Res 1985 18:273-81

Otte J, Schuknecht H, Kerr A. Ganglion cell populations in normal and pathological human

cochleae. Implications for cochlear implantation. Laryngoscope 1978 88:1231-46

Pfingst BE, De Haan DR, Holloway LA. Stimulus features affecting psychophysical detection thresholds for electrical stimulation of the cochlea.1: Phase duration and stimulus duration. J Acoust Soc Am 1991 90(4):1857-66

Rattay F, Motz H. Stimulation of a multichannel nerve array to pulse shapes produced by a single channel electrostimulation. Perception 1987 16:769-76.

Rosbe KW, Burgess BJ, Glynn RJ, Nadol JB. Morphologic evidence for three cell types in the human spiral ganglion. Hear Res 1996 93:120-7

Schmidt J. Cochlear neuronal populations in developmental detects of the inner ear:

implications for cochlear implantation. Acta Otolaryngol (Stockh) 1985 99:14-20

Shallop JK, Beiter AL, Goin DW, Mischke RE. Electrically Evoked Auditory Brain Stem Responses (EABR) and Middle Latency Responses (EMLR) Obtained from Patients with the Nucleus Multichannel Cochlear Implant. Ear Hear 1990 11,1:5-15

Shallop JK, VanDyke L, Goin DW, Mischke RE. Prediction of behavioral threshold and confort values for Nucleus 22-channel implant patients from electrical auditory brain stem response test results. Ann Otol Rhinol Laryngol 1991 100:896-8

Shallop JK. Objective electrophysiological measures from cochlear implants patients. Ear Hear 1993 14:58-63

Simmons FB. Electrical stimulation of the auditory nerve in man. Arch Otolaryngol 1966 84:2-54 Smith DW, Finley CC, Van den Honert C van den, Olszyk VB, Konrad KEM. Behavioral and electrophysiological responses to electrical stimulation in the cati. Absolute thresholds. Hear

Res 1994 81:1-10

Smith L, Simmons FB. Estimating eighth nerve survival by electrical stimulation. Ann Otol Rhinol Laryngol 1983 92:19-23

Sohmer H, Feinmesser M. Cochlear action potentials recorded from the external ear in man. Ann Otol Rhinol Laryngol 1967 76:427-35

Spoendlin H. Innervation densities of the cochleae. Acta Otolaryngol (Stockh) 1972 73:235-48 Spoendlin H, Schrott A. The Spiral Ganglion and the Innervation of the Human Organ of Corti. Acta Otolaryngol (Stockh) 1988 105:403-10

Spoendlin H, Schrott A. Analysis of the human auditory nerve. Hear Res 1989 43:25-38

Starr A, Brackman DE. Brainstem potentials evoked by electrical stimulation of the cochlea in human subjects. Ann Otol Rhinol Laryngol 1979 88:550-60

Truy E, Gallégo S, Berger-Vachon C, Morgon A. Multichannel cochlear implant in patients with ossified cochlea. Interest of the optimization of frequential selectivity of speech signal with Digisonic device. Proc IlIrd int Symp Transplants & implants in otology. Bordeaux (France), June 10-14 1995. M portman, P. Boudard, D. Portman, pp. 399-402.

Truy E, Gallégo S, Channal JM, Collet L, Morgon A. Correlation between electrical auditory brainstem response and perceptual thresholds in Digisonic cochlear implant. Laryngoscope 1997a, 117: in press.

Truy E, Gallégo S, Channal JM, Collet L, Berger-Vachon C, Morgon A. Intérêt des potentiels évoqués auditifs électriques du tronc cérébral chez les patients porteurs d'un implant cochléaire digisonic. Ann Otolaryngol Chir Cervicofac, 1997b, 114, 116-124.

Van den Honert C, Stypulkowski PH. Physiological properties of the electrically stimulated auditory nerve.11. Single fiber recording. Hear Res 1984 14:225-43

Van den Honert C, Stypulkowski PH. Characterization of the electrically evoked auditory brainstem response (ABR) in cats and humans. Hear Res 1986 21:109-26

Waring MD. Electrically evoked auditory brainstem response monotoring of auditory brainstem

implant integrity during facial nerve tumor surgery. Laryngoscope 1992 102:1293-5

Waring MD. Auditory brain-stem responses by electrical stimulation of the cochlear nucleus in

human subjects. Electroenceph clin Neurophysiol 1995 96:338-47

Zapala DA, Gould HJ, Mendel MI. Place specific influence on the wave I to V interpeak latency of the auditory brain-stem response. J Acoust Soc Am 1992 92(6):3174-84