JARO

JARO 16: 669–678 (2015)

DOI: 10.1007/s10162-015-0532-x
D 2015 The Author(s). This article is published with open access at Springerlink.com

Research Article Journal of the Association for Research in Otolaryngology

Delayed Auditory Brainstem Responses in Prelingually Deaf

and Late-Implanted Cochlear Implant Users
MARC J. W. LAMMERS,1,2 RUBEN H. M. VAN EIJL,1,2 GIJSBERT A. VAN ZANTEN,1,2 HUIB VERSNEL,1,2 AND
WILKO GROLMAN1,2
1
Department of Otorhinolaryngology and Head & Neck Surgery, University Medical Center Utrecht, Room
G.02.531P.O. Box 855003508 GA, Utrecht, The Netherlands
2
Brain Center Rudolf Magnus, University Medical Center Utrecht, Utrecht, The Netherlands
Received: 16 February 2015; Accepted: 15 June 2015; Online publication: 11 July 2015

ABSTRACT Keywords: auditory brainstem response, cochlear

implant, maturation, humans, prelingual deaf,
Neurophysiological studies in animals and humans hearing loss, deafness
suggest that severe hearing loss during early develop-
ment impairs the maturation of the auditory
brainstem. To date, studies in humans have mainly
focused on the neural activation of the auditory INTRODUCTION
brainstem in children treated with a cochlear implant
(CI), but little is known about the pattern of activation The duration of auditory deprivation before cochlear
in adult CI users with early onset of deafness implantation is an important predictor for hearing with a
(prelingual, before the age of 2 years). In this study, cochlear implant (CI). Whereas patients with late onset of
we compare auditory brainstem activation in deafness can obtain good speech perception in quiet, the
prelingually deaf and late-implanted adult CI users hearing performance of patients with long-term early-onset
to that in postlingually deaf CI users. Electrically deafness is generally poor (Teoh et al. 2004; Lammers et al.
evoked auditory brainstem responses (eABRs) were 2015). The large difference might be caused by an
recorded by monopolar stimulation, separately using impaired development of their auditory pathway in
a middle and an apical electrode of the CI. combination with cross-modal changes during a prolonged
Comparison of the eABR latencies revealed that wave period of auditory deprivation (Doucet et al. 2006; Lee
V was significantly delayed in the prelingually deaf CI et al. 2007; Kral and O’Donoghue 2010; Kral and Sharma
users on both electrode locations. Accordingly, when 2012; Lammers et al. 2015). Recently, we demonstrated
the apical electrode was stimulated, the III–V that prelingually deaf and late-implanted CI users display
interwave interval was significantly longer in the relatively early and large N1 peaks of the cortical auditory-
prelingually deaf group. These findings suggest a evoked potential (Lammers et al. 2015). This altered
slower neural conduction in the auditory brainstem, cortical activity raises the question regarding the extent to
probably caused by impairment of maturation during which the subcortical pathway, particularly the auditory
the long duration of severe hearing loss in infancy. brainstem, is affected in prelingually deaf CI users.
Shorter wave V latencies, reflecting a more mature It is well documented that after birth, auditory
brainstem, appeared to be a predictor for better brainstem response (ABR) wave latencies decrease and
speech perception. reach adult levels around the age of 2–3 years (Inagaki
et al. 1987; Eggermont and Salamy 1988). This decrease is
slower for wave V than for early waves and is hypothesized
Correspondence to: Huib Versnel & Department of Otorhinolaryngol- to result from increasing myelination and/or synaptic
ogy and Head & Neck Surgery & University Medical Center Utrecht &
Room G.02.531P.O. Box 855003508 GA, Utrecht, The Netherlands. efficacy within the auditory brainstem since these develop-
email: [email protected] ments lead to faster axonal conduction and synaptic
669
670 LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users

transmission (Eggermont and Salamy 1988; Moore et al. 2012 were consented to participate in a study which
1995; Thai-Van et al. 2007). Long periods of deafness affect included eABR and cortical auditory-evoked potential
the subcortical pathway, resulting in gradual spiral gangli- (CAEP) recordings. Twenty-three adults, with at least
on cell degeneration (Spoendlin 1975; Versnel et al. 2007), 6 months experience with their CI, agreed to participate
and a volume reduction of the cochlear nucleus and its in this study. In 20 subjects, eABRs could be recorded,
cells (Moore 1990; Leake et al. 2008; Ryugo et al. 2010). while in the remaining three postlingual subjects, no
On the other hand, electrically evoked ABRs (eABRs) in clear eABRs could be evoked due to electrical artifacts
congenitally deaf cats demonstrate latencies decreasing contaminating the waveforms. Their data were therefore
with age, similarly to normal-hearing cats (Tillein et al. not included in the analyses. Due to time constraints, in
2012). This suggests that auditory brainstem structures and one subject eABRs without CAEPs were obtained (post
pathways develop even in the absence of auditory 1). Prelingually deaf subjects were selected based on the
stimulation. following criteria: onset of severe to profound binaural
In humans, development of the auditory brainstem hearing loss before the age of 2 years (based on medical
following deafness has been studied by recording charts including diagnostic audiometry and self-reported
eABRs in children with CI (Gordon et al. 2006, 2008; patient information) and insufficient residual hearing
Thai-Van et al. 2007; Sparreboom et al. 2010). These during childhood for normal speech and language
studies demonstrated that in children with early-onset development. Based on these criteria, the diagnosis of
deafness, eABR wave latencies decrease after implan- prelingual deafness was confirmed by the multidisciplin-
tation, irrespective of age at implantation, like they do ary CI team prior to implantation. Eleven adults met
in normal-hearing children. On the contrary, in these criteria and were thus labeled as prelingually deaf.
bilaterally implanted children, when a response is Nine adults became deaf during adolescence or adult-
evoked using a second CI implanted much later than hood (915 years of age) and were categorized as
the first, the wave V latency is longer than the postlingually deaf. All participants were users of Nucleus
responses evoked by the first CI (Gordon et al. 2008; multi-channel CIs, and in all subjects a full insertion of
Sparreboom et al. 2010). This suggests impaired the electrode array was achieved. Table 1 summarizes
maturation of the auditory brainstem of the later detailed patient characteristics. The data of the CAEPs
implanted ear (Gordon et al. 2008; Sparreboom et al. (recorded in 22 subjects, including all 11 prelingual
2010), or altered neuronal connections induced by subjects and 8 postlingual subjects enrolled in this study)
the period of unilateral hearing with the first CI. have been reported in a separate paper (Lammers et al.
Whereas above-mentioned studies were performed 2015).
in children and in animal models, we address the effect
of early deafness on the auditory brainstem in adults. We
Speech Perception
compare eABRs in prelingually deaf subjects who had
little or no auditory stimulation for more than 20 years Speech perception scores were obtained using the
to eABRs in postlingually deaf CI users. According to Dutch Society of Audiology standard consonant-vowel-
various studies (Eggermont and Salamy 1988; Moore consonant (CVC) word list at 65 dB SPL (Versfeld
et al. 1995; Thai-Van et al. 2007; Leake et al. 2008; Ryugo et al. 2000). In this open-set test, only auditory cues
et al. 2010), myelination and synapses in the brainstem were available. Speech perception was scored based
of the postlingual group should have developed nor- on the number of phonemes correctly identified. For
mally because of sufficient auditory input during each subject, the most recent scores prior to the
childhood. In contrast, we expect the coarse structures evoked potential recordings were used (time intervals
of the brainstem to develop in the prelingual group but between 0 and 9 months).
sensory-driven maturation to be impaired reducing
axonal myelination and synaptic efficacy among others.
Procedure and Stimuli
Any effect by auditory stimulation after cochlear im-
plantation could only have occurred in the adult system Participants were seated in a comfortable reclining
and is expected to be negligible. Therefore, we hypoth- chair in an electrically shielded, sound attenuated
esize typical eABR waveforms in both groups, but longer booth and were asked to keep their eyes closed and
wave V latencies in the prelingually deaf. minimize movements. The electric stimulus consisted
of a biphasic pulse, with a phase width of 25 μs and an
inter-phase gap of 8 or 58 μs. A monopolar stimula-
METHODS tion electrode configuration was applied, and two
positions of the active electrode were used: at the
Participants
apical end of the array (typically electrode no. 20) and
All adult users of a Cochlear® CI who visited the a central position (typically electrode no. 11). A basal
outpatient clinic from December 2011 to December electrode was also used (as it was for CAEPs, Lammers
LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users 671

TABLE 1
Subject demographics
Subject Group Side Etiology Age at Age at Age at Implant Pre-op CVC Post-op CVC Primary mode of
CI test onset CI experience score score communication
(years) deafness (years) (years) (%) (%)a (pre-operative)
(years)

Pre 1 Prelingual Right Meningitis 23 0.5 21 1.7 0 22 (3) Lip reading

Pre 2 Prelingual Right Congenital 38 0 37 0.6 0 0 (1) Lip reading and
sign language
Pre 3 Prelingual Left Rubella 56 0.3 54 2.5 0 82 (5) Lip reading
Pre 4 Prelingual Left Varicella 55 2 55 0.6 0 26 (5) Lip reading and
sign language
Pre 5 Prelingual Left Rubella 47 0 46 1.5 48 66 (4) Oral and lip
reading
Pre 6 Prelingual Left Meningitis 55 0.8 49 5.7 12 18 (0) Lip reading and
sign language
Pre 7 Prelingual Left Meningitis 47 0.8 41 5.7 0 28 (5) Lip reading and
sign language
Pre 8 Prelingual Right Congenital 43 0 40 2.9 0 0 (9) Lip reading and
sign language
Pre 9 Prelingual Left Unknown 31 0 27 3.2 0 15 (8) Lip reading and
sign language
Pre 10 Prelingual Right Rubella 42 0 41 1.4 0 0 (1) Lip reading and
sign language
Pre 11 Prelingual Left Unknown 36 0 27 8.9 28 77 (0) Lip reading and
sign language
Mean 43 0 40 3.2 8 30
Post 1 Postlingual Left Progressive 75 68 68 6.5 0 78 (0) Oral
Post 2 Postlingual Right Otitis media 80 64 75 5.3 37 89 (2) Oral
Post 3 Postlingual Right Otitis media 58 27 43 14.4 – 60 (3) Oral and lip
reading
Post 4 Postlingual Right Unknown 73 50 59 13.8 – 72 (0) Oral
Post 5 Postlingual Right Progressive 57 29 50 6.9 28 95 (7) Oral and lip
reading
Post 6 Postlingual Left Trauma 22 16 16 6.5 0 90 (0) Oral
Post 8 Postlingual Right Progressive 59 43 45 13.3 25 85 (3) Oral and lip
reading
Post 9 Postlingual Right Progressive 53 40 41 11.9 11 84 (8) Oral
Post 11 Postlingual Left Meningitis 33 15 25 8.2 0 97 (0) Oral
Mean 57 39 47 9.6 14 83
– not performed
a
Months between eABR recording and CVC measures are displayed between brackets in the post-op CVC score column

et al. 2015), but due to stimulation artifacts, the signal- using a Medelec Synergy T-10 Evoked Potential
to-noise ratio of the eABRs was too low to obtain system. The ground was placed on the forehead and
reliable and reproducible waveforms in several pa- the contralateral mastoid was used as reference.
tients. Stimuli were generated using the Cochlear Recordings were filtered from 100 Hz to 5 kHz and
Custom Sound EP 3.1 software and presented at a rate recorded with a sampling rate of 20 kHz. Sweeps
of 35 Hz at the individual’s maximum comfortable containing signals of 950 μV at any electrode were
loudness level (C-level). For each subject, 1,500 rejected and not included in the average signal.
accepted sweeps were averaged. Multiple additional Electrode impedances were kept below 5 kΩ.
replications at the same level and lower stimulus levels
were recorded to confirm the response. Stimulation
levels were decreased until wave V could not be
Data Analyses
distinguished anymore. For data analyses, only the
response obtained at C-level was used for each subject. Averaged eABR data were analyzed using custom
scripts in MATLAB (version 7.11.0, Mathworks). For
each subject, the analysis of the wave III and wave V
Evoked Potential Recording latencies and the III–V interval was based on the
Responses were recorded by Ag/AgCl electrodes eABRs obtained at maximum comfort level. Waves III
placed according to the 10–20 system at Cz and Fz and V were manually identified by two authors (MJWL
672 LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users

and RHMvE) independently. Disagreements were and the postlingual group differed significantly in implant
resolved by discussion. experience (mean difference: 6.5 years; unpaired, two-
Statistical analyses were completed using SPSS tailed t test, t18=4.6, PG0.001). Age at implantation did not
version 22.0 software. Repeated measures ANOVAs differ significantly (unpaired, two-tailed t test, t18=1.1,
with the two different intracochlear stimulus locations P=0.305), neither did age at testing (unpaired, two-tailed
(i.e., middle and apical) as within-subjects factor and t test, t18=2.0, P=0.057). Preoperative CVC phoneme scores
group (i.e., prelingual or postlingual) as between- were not different between the two groups.
subjects factor were used. Significant main effects and
interactions (pG0.05) were followed with Bonferroni
eABR Waveforms in the Pre- and Postlingual
post hoc tests, and the Greenhouse–Geisser correc-
Groups
tion was applied to compensate for violations of the
sphericity assumption. Group differences in peak In all nine postlingual subjects, reproducible eABR
latencies were calculated with unpaired, two-tailed t waveforms could be obtained on both electrodes. In 9
tests or the Mann–Whitney test for independent out of 11 prelingual subjects, eABR waveforms could be
samples. Stepwise multiple linear regression analyses obtained on both electrodes, whereas in two subjects no
were performed to evaluate the influence of other clear waves III and V could be identified on the middle
variables on wave latencies. Linear regression analyses electrode. Figure 2 shows the individual eABR waveforms
were performed to evaluate the relationship between evoked at an apical electrode (typically electrode no. 20)
speech perception and eABR latencies. for prelingual (left) and postlingual (right) subjects in
order of CVC score. The waveforms tended to be
relatively small for the prelingual subjects with poor
RESULTS speech perception. Depending on factors such as
location of recording and stimulation electrodes, head
Patient Demographics and Speech Perception size, and skull thickness, eABR wave amplitudes were
highly variable among subjects and thus less reliable for
The monosyllabic open-set speech perception scores group comparisons. Therefore, only wave latencies were
varied widely among the patients, ranging from 0 to considered for group comparisons.
82 % in the prelingual group and from 60 to 97 % in the
postlingual group (Fig. 1). Median speech perception
scores were significantly higher in the postlingual group Wave Latencies
(median 85 % correct) than in the prelingual group (22 % Grand averages of the eABRs evoked at apical and middle
correct; Mann–Whitney U=6.0, PG0.0001). Patient demo- electrodes (Fig. 3) indicate that waves III of the two patient
graphics are presented in Table 1. The prelingual group groups coincide, while wave V starts and peaks considerably
later for the prelingually deaf patients when compared to
100 the postlingually deaf. Accordingly, wave V latency was
CVC phoneme score (%)

¥ significantly longer in the nine prelingual subjects than in

80 the nine postlingual subjects across electrode locations
§ (F(1,16)=6.86, P=0.019; Fig. 4A). Analyses for the two
60 electrode locations separately revealed that on the apical
electrode the average wave V latency in the 11 prelingual
40 subjects was 3.9 ms, whereas in the postlingual subjects this
¥ ¥ was 3.6 ms (unpaired, two-tailed t test, t18=2.6, P=0.020).
¥
20 ¥ On the middle electrode, wave V latencies were on average
3.9 ms for the prelingual subjects and 3.7 ms for the
0 postlingual subjects (unpaired, two-tailed t test, t16=2.4,
Prelingual Postlingual P=0.030). If the three best performing prelingual users, pre
3, 5 and 11, were excluded from the analyses, wave V
FIG. 1. Consonant-vowel-consonant (CVC) phoneme scores of the
prelingual and postlingual group taken prior to the evoked potential latency remained significantly longer than in the
recordings. Each dot represents an individual subject. The horizontal postlingual group (F(1,13)=13.48, P=0.003).
lines represent the median scores (prelingual group, 22 %; Analysis of the wave III latency did not reveal
postlingual group, 85 %). Subjects with hearing experience (pre 1, differences between the prelingual group and the
3, 4, 6, and 7) are marked with yen sign; subject pre 5, who used
postlingual group (F(1,16)=1.14, P=0.301). Average
both visual speech cues and oral communication preoperatively, is
marked with section sign. This plot is a partial replot of Figure 1 in wave III latencies were around 2.0 ms in the
Lammers et al. (2015); all 11 prelingual data points and 8 of 11 prelingual group and 1.9 ms in the postlingual group
postlingual data points of that paper are plotted here. on both electrode locations (apical electrode: un-
paired, two-tailed t test, t18=1.1, P=0.307; middle
LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users 673

Prelingual Postlingual

82%¥
Pre 3 97%‡
Post 11
Pre 11 77%§
95%‡
1 µV Post 5

Pre 5 90%‡
Post 6
66%§
89%
Pre 7 Post 2
28%¥
Pre 4
26%¥
85%
Post 8

22%‡¥
Pre 1 84%‡
18%¥§‡ Post 9
Pre 6

15%‡
Pre 9 78%‡
Post 1
Pre 2 0%

72%
Post 4
0%
Pre 8
60%‡
Pre 10 0%‡ Post 3

1 2 3 4 5 6 1 2 3 4 5 6
Time (ms) Time (ms)
FIG. 2. Individual eABR waveforms evoked at an apical electrode. shown to facilitate comparisons between subjects. At the right side of
In cases in which large stimulus artifacts partially obscured the the individual waveforms, the CVC phoneme scores are presented. In
measurements recorded on Cz, waveforms measured at electrode Fz the outer margins of the figure the patient numbers are presented.
are presented here (denoted with double dagger sign). Waveforms Subjects with hearing experience (pre 1, 3, 4, 6, and 7) are marked
were corrected for stimulus artifact by fitting a first-order polynomial with yen sign; subjects who had a preoperative CVC phoneme score
and subtracting it from the signal. In almost all subjects in both higher than 0 % correct (pre 5, 6 and 11) are marked with section
groups, wave III (indicated with first upward arrowhead) and wave V sign.
(indicated with second upward arrowhead) could be identified. The
vertical lines drawn near the peak latencies of waves III and V are

electrode unpaired, two-tailed t test, t16=1.2, P=0.237). Fig. 4B). If the three best performing prelingual users, pre
Wave III latency remained similar for both groups, if 3, 5 and 11, were excluded from the analyses, the III–V
the three best performing prelingual subjects were interval in the prelingually deaf CI users, was significantly
excluded from the analyses (F(1,13)=1.71, P=0.214). longer than in the postlingual subjects across both
The interwave III–V interval was longer for electrode locations (F(1,13)=7.05, P=0.020).
prelingually than for postlingually deafened, which was
nearly significant when analyzed across both electrode
locations (F(1,16)=4.43, P=0.052). On the apical electrode Factors Related to Wave V Latency and III–V
the III–V interval was significantly longer in the prelingual Interval
subjects (average III–V interval prelingual group: 1.9 ms, In order to assess if other factors, besides age at onset of
postlingual group 1.7 ms; unpaired, two-tailed t test, deafness, could have contributed to the differences in
t18=2.7, P=0.015). On the middle electrode, the difference wave V latency, stepwise multivariate regression analyses
in III–V interval between groups was smaller and not were performed. On the apical electrode, these analyses
significant (unpaired, two-tailed t test, t18=1.5, P=0.143; revealed that group was the only significant predictor of
674 LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users

Middle Apical

Prelingual
Postlingual

0.4

1 2 3 4 5 6 1 2 3 4 5 6
Time (ms) Time (ms)
FIG. 3. Grand average eABR waveforms measured at Cz for all polynomial and subtracting it from the signal and then they were
subjects in both groups, presented for the two stimulation electrode normalized by dividing the signal by the difference in amplitude
locations separately. The blue and red traces represent the waveforms between the top of wave V and its preceding trough.
of the prelingual and postlingual groups, respectively. Waveforms
were first corrected for stimulus artifact by fitting a first-order

wave V latency (r=0.516, F(1,18)=6.535, P=0.020). Implant Wave V latency was not correlated to age at
experience was not associated with wave V latency on this implantation on either tested electrode. Moreover,
electrode (P=0.846; Fig. 5B). Contrarily, implant experi- the other preoperative patient characteristics present-
ence was found to be the only significant predictor of wave ed in Table 1 were not a factor determining wave V
V latency on the middle electrode (r=−0.665, F(1,16)= latency. Within the prelingual or postlingual group
12.693, P=0.003; Fig. 5A). Group was not a significant the wave V latency was not significantly correlated to
predictor of wave V latency on this electrode (P=0.887). age at onset of deafness.
When examining both groups separately, wave V latency The III–V interval was not associated with age at
was significantly correlated to implant experience in the implantation, age at onset of deafness, or implant
prelingual group (r=−0.706, P=0.034; Fig. 5A), but not in experience on either electrode location.
the postlingual group (r=−0.324, P=0.395). Previous
studies on eABRs in children demonstrated that the wave
V latency as function of CI experience could be best
Relationship Between eABRs and Speech
described by an exponential decay (Gordon et al. 2006;
Perception
Thai-Van et al. 2007). If we describe the wave V latencies As demonstrated in Figure 2, typical eABRs exhibiting
on the middle electrode in the prelingual subjects as an both waves III and V could be recorded in almost all
exponential, it would yield a time constant of about 4 years. subjects irrespective of their speech perception scores

A 5.0
B 3.0
Middle Middle
Apical Apical
* * *
Wave III-V interval (ms)
Wave V latency (ms)

4.5
2.5

4.0

2.0
3.5

3.0 1.5
e

st

st
st

st
Pr

Pr

Pr
Pr
Po

Po
Po

Po

FIG. 4. Wave V latencies (A) and III–V interwave intervals (B) of the prelingual (pre) and postlingual (post) groups presented for the two stimulation
electrode locations. The box plots represent the lower and upper quartile with the median. Whiskers indicate the 5–95 percentiles. *PG0.05.
LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users 675

A Middle electrode B Apical electrode

4.6 4.6
Prelingual
4.4 Postlingual 4.4

wave V latency (ms)

4.2 4.2

4.0 4.0

3.8 3.8

3.6 3.6

3.4 3.4

3.2 3.2
0 2 4 6 8 10 12 14 16 0 2 4 6 8 10 12 14 16
Implant experience (years) Implant experience (years)
Fig. 5. Wave V latency on the middle (A) and apical (B) electrode ence for the total group (dotted line, r=−0.665, P=0.003, n=18) and
as a function of implant experience. Each red square represents a for the prelingually deaf subjects (solid blue line, r=−0.706,
subject in the postlingual group. Each blue dot represents a subject in P=0.034).
the prelingual group. On the middle electrode, there was a
significant correlation between wave V latency and implant experi-

or group (pre or postlingually deaf). Waveform when both groups were analyzed separately. Stepwise
morphology was however less clear in the prelingually multivariate regression analyses confirmed that III–V
deaf and poor performing subjects. Besides, in two interval was not a significant predictor besides group.
prelingual subjects no clear eABRs could be recorded
on the middle electrode (Fig. 2).
Since postlingually deafened subjects show shorter
DISCUSSION
latencies and better speech perception, a correlation
is expected between these two measures. If all subjects
In this study, we evaluated differences in auditory
were included in a linear regression analysis and if
brainstem activation between pre- and postlingually
group was omitted, wave V latency was significantly
deaf subjects who received a cochlear implant in
associated with speech perception on both electrodes
adulthood. The extensive duration of auditory depri-
(Apical r=−0.640, P=0.002; Middle r=−0.706, P=0.001;
vation in the CI users with prelingual deafness
Fig. 6 dotted lines). When the groups were analyzed
appeared to have resulted in a delayed wave V while
separately a negative correlation between phoneme
wave III was virtually unaffected. This delayed wave V
score and wave V latency was present on the apical
activation may reflect a hindered neural and synaptic
electrode for the postlingual group (r=−0.833,
development especially in the more rostral part of the
P=0.005; Fig. 6B solid line) and on the mid electrode
brainstem. To our knowledge, this is the first study to
for the prelingual group (r=−0.728, P=0.026; Fig. 6A
examine whether long durations of early onset
solid line).
auditory deprivation leads to delays in human audito-
Besides group, implant experience was significantly
ry brainstem activation. These findings reveal that in
associated with speech perception (r=0.580, P=0.007).
the absence of auditory input throughout childhood,
Stepwise multivariate regression analyses revealed that
development of the rostral auditory brainstem seems
on the apical electrode wave V latency and implant
to be hampered, due to degraded myelination and/or
experience were not predictive for speech perception,
synaptic efficacy.
and only group remained a significant predictor
(r=−0.755, F(1,19)=23.89, PG0.001). On the middle
electrode, group and wave V latency were significant
Auditory Brainstem Development and Hearing
predictors of speech perception explaining for 69 %
Impairment
of the variance (r=−0.828, F(1,19)=16.37, PG0.001;
variance explained by group: 37%, P=0.009 and by Neuropathological studies in autopsied fetuses and
wave V latency: 32 %, P=0.018). The III–V interval was infants revealed that the auditory brainstem rapidly
significantly associated with postoperative speech matures during the perinatal period. It is assumed
perception when assessed over all subjects (Apical that during the first two trimesters of pregnancy the
r=−0.632, P=0.003; Middle r=−0.494, P=0.037), but not anatomical structures involved in the auditory path-
676 LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users

A Middle electrode B Apical electrode

100 100
Prelingual
Postlingual
CVC phoneme score (%)
80 80

60 60

40 40

20 20

0 0
3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6
Wave V latency (ms) Wave V latency (ms)
FIG. 6. CVC phoneme score as a function of wave V latency for solid line indicates a significant correlation within the prelingual
the middle (A) and apical (B) electrodes. Each red square represents a group for the middle electrode (r=−0.728, P=0.026), and the red
subject in the postlingual group. Each blue dot represents a subject in solid line indicates a significant correlation within the postlingual
the prelingual group. The univariate analyses for the total group group for the apical electrode (r=−0.833, P=0.005).
revealed significant correlations on both electrodes (dotted lines,
Middle r=−0.706, P=0.001; Apical: r=−0.640, P=0.002). The blue

way are already developed, independent of sound- with these findings, the results of our study show that
evoked activity (Moore and Linthicum 2007). The after early onset of deafness and years of auditory
subsequent final maturation of the auditory brainstem deprivation, the wave V latency (and the wave III–V
is believed to result from sensory-driven processes of interval) is longer than in deaf subjects who experi-
myelination, an increase in size of auditory neurons enced auditory stimulation in early development.
and rapid growth and branching of dendrites (Thai- In children with early-onset deafness, it has been
Van et al. 2007; Moore and Linthicum 2007). shown that wave V latencies decrease over the first years
Myelination of the auditory pathway starts around following cochlear implantation, in similar fashion as the
the 26th to 29th week and continues in the brainstem exponential decrease with age observed in normal-
during the fetal and postnatal period (Inagaki et al. hearing children (Gordon et al. 2006; Thai-Van et al.
1987; Moore et al. 1995; Moore and Linthicum 2007). 2007). As suggested by our middle electrode recordings,
By the age of 1 year, the myelin density of the duration of cochlear implant use may have been a factor
cochlear nerve is comparable to that of an adult contributing to the wave V latency in addition to onset of
(Moore et al. 1995). deafness (prelingual vs. postlingual). Fitting a decaying
This myelination process and increase in synaptic exponential function shows the time constant of this
efficacy of the auditory brainstem is probably reflected effect to be 4 years, which would be much longer than
by the decreasing ABR wave latencies within the same previously reported constants of 68 weeks (Thai-Van et al.
period (Inagaki et al. 1987; Eggermont and Salamy 2007) or 5 months (Gordon et al. 2006) observed in
1988; Ponton et al. 1992, 1996). In the presence of children. Although not consistently found over both
normal auditory stimulation, ABR wave latencies electrode locations, this effect of implant experience
mature over the first years of life (Inagaki et al. might suggest a possible sensory driven maturation of the
1987; Eggermont and Salamy 1988; Ponton et al. auditory brainstem, which is still present even in adult-
1992). In normal-hearing children, wave V latencies hood after long durations of auditory deprivation.
decrease following an exponential model, reaching Acquiring longitudinal data will be necessary to clearly
adult values of 5–6 ms around the age of two. In the identify the role of chronic electrical stimulation in adults.
absence of normal auditory stimulation, congenitally Animal models of deafness show that the major
deaf children show a significantly longer wave V structures and pathways in the brainstem develop, but
latency when compared to children who became deaf auditory input is required for the refinement of neuronal
after 1 to 4 years of age (Thai-Van et al. 2007). Among connections. Hence, in the absence of auditory stimula-
prelingually deaf children implanted between the tion, the cochlear nucleus becomes smaller and the
ages of 1 to 17 years, the III–V wave interval recorded projections received from the auditory nerve are broad-
at activation of the CI does not depend on age at ened (Moore 1990; Leake et al. 2008), furthermore, cells
implantation (Gordon et al. 2006). In concordance shrink and synapses loose vesicles (Ryugo et al. 2010). The
LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users 677

cochleotopic organization is however maintained despite mechanisms of brainstem and cortex responses may
deafness, but these broader connections might affect also explain that eABR wave V latencies tend to
frequency resolution (Leake et al. 2008). In ferrets with decrease with implant experience, whereas such
bilateral cochlear ablation, the number of neurons tendency is not found for CAEP latencies.
projecting from the cochlear nucleus to the inferior
colliculus was not different than that of normal-hearing
animals (Moore 1990), suggesting that the coarse structure
Methodological Considerations
of the pathway is not affected by deafness. This is
supported by a longitudinal study comparing eABRs of In our study population, a significant correlation between
congenitally deaf cats with those with normal hearing speech perception and wave V latency was found when
(Tillein et al. 2012). This study showed that in the absence eABRs were evoked at the middle electrode (Fig. 6). This
of auditory input, the normal eABR waves develop and agrees with a report by Gallégo et al. (1998) who also
wave latencies decrease at the same rate as in normal- found an association between speech perception and wave
hearing cats. Although wave latencies were comparable, V latencies. Although wave V latency on the middle
eABR waveform morphology was less clear in the congen- electrode was found to be an independent predictor of
itally deaf cats. Moreover, the responses revealed a speech perception, these results should be interpreted
reduction in the amplitude of wave III and an increase of with caution. Larger patient series should be investigated
wave V amplitude (Tillein et al. 2012). The ongoing to clarify the predictive role of eABR wave latencies,
decrease of wave V latency in congenitally deaf cats after especially since there is no clear trend in the current
birth does not coincide with human data. This might be literature studying relative small sample sizes (Abbas and
due to a discrepancy between the two species or due to a Brown 1991; Brown et al. 1995; Gallégo et al. 1998;
difference in stimulation (e.g., lower level in cats than in Makhdoum et al. 1998; Firszt et al. 2002; Kim et al. 2008;
humans). Gibson et al. 2009).
Variability among the CI patients was typically substan-
tial. Notably, the speech perception scores among the
Changes in Brainstem Responses and Cortical
patients classified as prelingually deaf varied widely (Fig. 1).
Potentials in Prelingually Deaf
Three of those subjects had high CVC scores and short
Combining the results of the present study with the wave V latencies, both characteristics being shared by
cortical potentials recorded in the same subjects postlingually deaf subjects. Since judgment of early onset
(Lammers et al. 2015) gives an insight in the of deafness is partly based on self-reported information
spontaneous development of the auditory pathway. (see BMethods^ section), one cannot exclude significant
In the prelingually deaf subjects, both the brainstem use of residual hearing (also discussed in Lammers et al.
and cortical waveforms showed a normal morphology 2015), which might have contributed to a better develop-
indicating that a coarse neural network from cochlea ment of the brainstem and eventually to a better speech
to cortex develops regardless of sensory input. Thus, perception with their CI. On the other hand, sensitivity
even in the absence of sensory input throughout analyses excluding the three best performing prelingually
childhood, the development of this innate auditory deaf subjects did not change the found effects. Moreover,
pathway is not disrupted. the early onset of deafness in these three subjects is
Remarkably, whereas the brainstem responses were supported by the cortical potentials which had latencies in
delayed, the cortical responses in the prelingually deaf the range of the other prelingually deaf subjects, signifi-
subjects showed significantly shorter latencies than cantly shorter than postlingual subjects (Lammers et al.
the postlingually deaf subjects. We should stress that 2015).
the cortical latency differences (~15 ms) are almost
100 times larger than the brainstem latency differ-
ences (~0.2 ms). Whereas myelination and synaptic
development explain the relatively subtle ABR latency CONCLUSION
differences, other mechanisms play a role in cortical
potentials. Sensory input representing biologically Electrically evoked auditory brainstem responses in
relevant sounds is crucial for refinement of the prelingually deaf late-implanted CI users demonstrate
cortical network, in particular during childhood increased wave V latencies and interwave III–V
(Buonomano and Merzenich 1998; Innocenti and intervals, suggesting delays in neural conduction
Price 2005; Ohl and Scheich 2005; Kral 2013). Thus, within the auditory brainstem. These results indicate
without such input, the cortex maintains a rather that long durations of hearing impairment directly or
coarse organization containing fewer cortico-cortical shortly after birth may lead to impaired neuronal
connections with less top-down inhibition (Kral and connections within the innate and elementary parts of
Sharma 2012; Lammers et al. 2015). The different the auditory brainstem.
678 LAMMERS ET AL.: eABRs in Prelingually Deaf CI Users

ACKNOWLEDGMENTS INNOCENTI GM, PRICE DJ (2005) Exuberance in the development

of cortical networks. Nat Rev Neurosci 6:955–965
KIM AH, KILENY PR, ARTS HA, EL-KASHLAN HK, TELIAN SA, ZWOLAN TA
We gratefully acknowledge all participants and the members
(2008) Role of electrically evoked auditory brainstem response
of the cochlear implant team for their time and support. We in cochlear implantation of children with inner ear
thank Aren Bezdjian for his valuable suggestions on the malformations. Otol Neurotol 29:626–634
manuscript. This study is supported by an unrestrictive KRAL A (2013) Auditory critical periods: a review from system’s
research grant from Cochlear Ltd. perspective. Neuroscience 247:117–133
KRAL A, O’DONOGHUE GM (2010) Profound deafness in childhood. N
Conflict of Interest Engl J Med 363:1438–1450
KRAL A, SHARMA A (2012) Developmental neuroplasticity after
cochlear implantation. Trends Neurosci 35:111–122
Wilko Grolman received an unrestrictive research LAMMERS MJW, VERSNEL H, VAN ZANTEN GA, GROLMAN W (2015)
grant from Cochlear Ltd. for this study. Wilko Altered cortical activity in prelingually deafened cochlear
Grolman received unrestrictive research grants from implant users following long periods of auditory deprivation. J
MED-EL GmbH and Advanced Bionics. No compet- Assoc Res Otolaryngol 16:159–170
LEAKE PA, HRADEK GT, BONHAM BH, SNYDER RL (2008) Topography
ing interests declared by the other authors. of auditory nerve projections to the cochlear nucleus in cats
after neonatal deafness and electrical stimulation by a cochlear
Open Access This article is distributed under the terms of implant. J Assoc Res Otolaryngol 9:349–372
the Creative Commons Attribution 4.0 International License LEE HJ, GIRAUD AL, KANG H, KIM CS, LEE DS (2007) Cortical activity at rest
(http://creativecommons.org/licenses/by/4.0/), which predicts cochlear implantation outcome. Cereb Cortex 17:909–917
permits unrestricted use, distribution, and reproduction in MAKHDOUM MJ, GROENEN PA, SNIK AF, VAN DEN BROEK P (1998) Intra- and
any medium, provided you give appropriate credit to the interindividual correlations between auditory evoked potentials and
original author(s) and the source, provide a link to the speech perception in cochlear implant users. Scand Audiol 27:13–20
Creative Commons license, and indicate if changes were MOORE DR (1990) Auditory brainstem of the ferret: bilateral
made. cochlear lesions in infancy do not affect the number of neurons
projecting from the cochlear nucleus to the inferior colliculus.
Brain Res Dev Brain Res 54:125–130
MOORE JK, LINTHICUM FH JR (2007) The human auditory system: a
REFERENCES timeline of development. Int J Audiol 46:460–478
MOORE JK, PERAZZO LM, BRAUN A (1995) Time course of axonal myelination
ABBAS PJ, BROWN CJ (1991) Electrically evoked auditory brainstem in the human brainstem auditory pathway. Hear Res 87:21–31
response: growth of response with current level. Hear Res 51:123–137 OHL FW, SCHEICH H (2005) Learning-induced plasticity in animal
BROWN CJ, ABBAS PJ, BERTSCHY M, TYLER RS, LOWDER M, TAKAHASHI G, and human auditory cortex. Curr Opin Neurobiol 15:470–477
PURDY S, GANTZ BJ (1995) Longitudinal assessment of physiolog- PONTON CW, EGGERMONT JJ, COUPLAND SG, WINKELAAR R (1992)
ical and psychophysical measures in cochlear implant users. Ear Frequency-specific maturation of the eighth nerve and brain-
Hear 16:439–449 stem auditory pathway: evidence from derived auditory brain-
BUONOMANO DV, MERZENICH MM (1998) Cortical plasticity: from stem responses (ABRs). J Acoust Soc Am 91:1576–1586
synapses to maps. Annu Rev Neurosci 21:149–186 PONTON CW, MOORE JK, EGGERMONT JJ (1996) Auditory brainstem response
DOUCET M, BERGERON F, LASSONDE M, FERRON P, LEPORE F (2006) generation by parallel pathways: differential maturation of axonal
Cross-modal reorganization and speech perception in cochlear conduction time and synaptic transmission. Ear Hear 17:402–410
implant users. Brain 129:3376–3383 RYUGO DK, BAKER CA, MONTEY KL, CHANG LY, COCO A, FALLON JB,
EGGERMONT JJ, SALAMY A (1988) Maturational time course for the SHEPHERD RK (2010) Synaptic plasticity after chemical deafening
ABR in preterm and full term infants. Hear Res 33:35–47 and electrical stimulation of the auditory nerve in cats. J Comp
FIRSZT JB, CHAMBERS RD, KRAUS N (2002) Neurophysiology of cochlear Neurol 518:1046–1063
implant users II: comparison among speech perception, dynamic SPARREBOOM M, BEYNON AJ, SNIK AFM, MYLANUS EAM (2010) Electrically
range, and physiological measures. Ear Hear 23:516–531 evoked auditory brainstem responses in children with sequential
GALLÉGO S, FRACHET B, MICHEYL C, TRUY E, COLLET L (1998) Cochlear bilateral cochlear implants. Otol Neurotol 31:1055–1061
implant performance and electrically-evoked auditory brain- SPOENDLIN H (1975) Retrograde degeneration of the cochlear nerve.
stem response characteristics. Electroencephalogr Clin Acta Otolaryngol 79:266–275
Neurophysiol 108:521–525 TEOH SW, PISONI DB, MIYAMOTO RT (2004) Cochlear implantation in
GIBSON WP, SANLI H, PSARROS C (2009) The use of intra-operative adults with prelingual deafness. Part I. Clinical results. Laryngo-
electrical auditory brainstem responses to predict the speech scope 114:1536–1540
perception outcome after cochlear implantation. Cochlear THAI-VAN H, COZMA S, BOUTITIE F, DISANT F, TRUY E, COLLET L (2007)
Implants Int 10(Suppl 1):53–57 The pattern of auditory brainstem response wave V maturation
GORDON KA, PAPSIN BC, HARRISON RV (2006) An evoked potential in cochlear-implanted children. Clin Neurophysiol 118:676–689
study of the developmental time course of the auditory nerve TILLEIN J, HEID S, LANG E, HARTMANN R, KRAL A (2012) Development
and brainstem in children using cochlear implants. Audiol of brainstem-evoked responses in congenital auditory depriva-
Neurotol 11:7–23 tion. Neural Plast. 182767
GORDON KA, VALERO J, VAN HOESEL R, PAPSIN BC (2008) Abnormal VERSFELD NJ, DAALDER L, FESTEN JM, HOUTGAST T (2000) Method for
timing delays in auditory brainstem responses evoked by the selection of sentence materials for efficient measurement of
bilateral cochlear implant use in children. Otol Neurotol the speech reception threshold. J Acoust Soc Am 107:1671–1684
29:193–198 VERSNEL H, AGTERBERG MJH, DE GROOT JCMJ, SMOORENBURG GF, KLIS
INAGAKI M, TOMITA Y, TAKASHIMA S, OHTANI K, ANDOH G, TAKESHITA K SFL (2007) Time course of cochlear electrophysiology and
(1987) Functional and morphometrical maturation of the morphology after combined administration of kanamycin and
brainstem auditory pathway. Brain Dev 9:597–601 furosemide. Hear Res 231:1–12