Академический Документы
Профессиональный Документы
Культура Документы
A R T I C L E I N F O
A B S T R A C T
Article history:
Received 15 March 2013
Received in revised form 1 July 2013
Accepted 3 July 2013
Available online 1 August 2013
Objective: The difference of characteristics (latency and amplitude) between toneburst and narrow CEchirp1 stimuli on ABR recording was analyzed in normal hearing infants.
Methods: 500, 1000, 2000 and 4000 Hz toneburst and narrow band CE-chirp1 auditory brainstem
responses (ABRs) were recorded in 40 normal-hearing infants. The amplitude and latency parameters of
the ABR were collected for each of the four stimulus levels: 80, 60, 40, and 20 dB nHL. Both stimuli started
from 80 dB nHL using alternating polarity and the rates were both 27.1/s.
Results: The toneburst latencies are greater than narrow band CE-chirp latencies for all intensities at 500,
1000 and 2000 Hz (p < 0.001). However, at 4000 Hz this difference was not signicant. At 500 Hz, wave V
amplitude is larger for toneburst than narrow CE-chirp1 (p < 0.001) in 80 dB nHL. The difference
between the two stimuli in 60 dB nHL was not signicant (p = 0.495) and at 40 and 20 dB nHL the wave V
narrow band CE-chirp1 amplitude is greater than toneburst amplitude (p < 0.001). At 1000, 2000 and
4000 Hz there is no difference between the wave V toneburst and narrow band CE-chirp1 amplitudes at
80 dB nHL (p = 0.940; p = 0.776 and p = 0.217 respectively). On the other hand, in the levels to 60, 40 and
20 dB nHL, narrow band CE-chirp1 amplitudes are larger than toneburst amplitude (p < 0.001).
Conclusion: Narrow band CE-chirp1 ABRs generates shorter latencies than the toneburst ABRs,
especially to low frequencies. Higher amplitudes were found with narrow band CE-chirp1 stimuli for all
frequencies tested, except to high levels.
2013 Elsevier Ireland Ltd. All rights reserved.
Keywords:
Auditory brain stem responses
Evoked response audiometry
Hearing thresholds
Children
1. Introduction
The toneburst auditory brainstem response (ABR) is the current
gold standard test recommended by the Joint Committee on Infant
Hearing (JCIH) [1] for estimation of the audiogram for infants
between 1 and 3 months. The JCIH statement suggest that the
clinicians use air conduction toneburst stimuli to record ABR and,
when thresholds are elevated, bone-conduction toneburst stimuli
in order to differentiate among sensory, conductive and mixed
hearing loss and to determine the conguration of hearing loss in
each ear [1]. Despite those recommendations, many clinicians
today resist to use toneburst stimuli for ABR thresholds. One of the
most frequently reported problems is the difcult to identify the
wave V, particularly, using low frequency stimuli [2].
A recent advance in the ABR recording has been the use of chirp
stimuli, which compensates for frequency delay characteristics of
the basilar membrane. While with the click the stimulation of low
frequency cochlear regions occur after the high frequency regions,
the chirp stimulus was designed to produce simultaneous displacement maxima by canceling traveling-time differences along the
cochlear partition. The low frequency cochlear parts are stimulated
at the same time that the high frequency parts, so all parts on the
basilar membrane should reach maximum depolarization, producing a simultaneous neural response. ABR evoked by chirp stimulus
showed larger wave V amplitude than click stimulus [37].
Different chirps have been developed based on different models
delay to the cochlear traveling wave [4,5,810]. The CE-chirp1
patent is the result of studies by Claus Elberling and his
collaborators [5,11]. The broadband CE-Chirp1 is designed using
a delay model based on derived band ABR latencies. Decomposing
the broadband CE-Chirp1 into four components, ltered versions
of the CE-Chirp1 were created with the center frequencies 500,
1000, 2000, and 4000 Hz [5]. These stimuli, called narrow band CEChirps1 were obtained by decomposing the CE-Chirp1 and
constitute a subset of the CE-Chirp1 [12].
In frequency domain, Elberling and his collaborators [5] showed
that the chirps give shorter detection time and higher signal-tonoise ratio than the click when auditory steady state to chirp and to
1556
normal-hearing individuals in accordance with the recommendations put forward in ISO 389-9 (2009) [12].
To ensure the best possible outcome of the toneburst and
narrow band ABR testing, all recordings were repeated to ensure
reproducibility and a running estimate of the physiological
background noise was made [13]. The ABRs were obtained by
weighted averaging [14] and the averaging was terminated when
the estimated residual background noise was 40 nV or less after
800 sweeps [15].
ABR records began in the intensity of 80 dB nHL with toneburst
stimuli and then responses were recorded in the intensities of 60,
40 and 20 dB nHL. After recording toneburst ABR in one ear, the
narrow band CE-chirp1 ABR recordings were initiated in the same
ear, following the same intensities recorded in the toneburst ABR.
Then this was repeated the same order in the other ear when
possible.
ABR component amplitudes (peak to following trough) and
latencies were determined by visual inspection for waveform V of
the ABR. In each recorded ABR the most salient response peak wave
V replicable was identied by an experienced clinician and its
latency measured. Wave V peak-to-peak amplitude was analyzed
in all stimulus and level conditions. The amplitude was measured
by reading the voltage between the computers cursor at the
negative peak and a second curser at the initial positive peak. The
amplitude was dened as the difference between the voltages
obtained from the two cursor positions.
3. Results
At 500, 1000, 2000 and 4000 Hz, for each condition level and
stimulus, the mean latencies and the corresponding standard
deviations and the difference between the latencies recorded with
both stimuli are calculated and shown in Table 1.
Observing the difference between two stimuli latency (toneburst latency narrow band CE-chirp1 latency) at each intensity,
all differences are positive, indicating that the latency is greater
when using the toneburst stimulus than when using narrow band
CE-chirp1 stimulus.
Latency data were analyzed using a two-factor (toneburst and
narrow band CE-chirp1) analysis of variance (ANOVA) using a
within-subject repeated measures (stimuli and intensity) and is
possible to conclude that tonebursts latencies is greater than
narrow band CE-chirp1 latencies for all intensities (p < 0.001) at
500, 1000 and 2000 Hz. But, contrary to occurred with the
frequencies at 500, 1000 and 2000 Hz, at 4000 Hz this difference
was no signicant. At 4000 Hz narrow band CE-chirps and
toneburst latencies are similar. However, the median differences
are positive, meaning that at least 50% of the ears toneburst latency
is greater than narrow band CE-chirp1 latency.
The effects of stimulus type and latency intensity function of
the wave V are illustrated in Fig. 1.
The ABR latencies for the low frequency stimuli (500 and
1000 Hz), are shorter than normally seen in corresponding
toneburst ABRs, especially at higher levels (80 dB nHL).
At 500, 1000, 2000 and 4000 Hz, for each condition level and
stimulus, the mean amplitude and the corresponding standard
deviations and the difference between the amplitude recorded
with both stimuli were calculated in Table 2.
Pairwise T-test was performed and the wave V 500 Hz
toneburst amplitude was greater than narrow band CE-chirp1
amplitude (p < 0.001) at 80 dB nHL. The difference between the
two stimuli in 60 dB nHL was not signicant (p = 0.495) and at 40
and 20 dB nHL the wave V narrow band CE-chirp1 amplitude was
greater than toneburst amplitude (p < 0.001).
At 1000 Hz, pairwise T-test showed that there is no difference
between the wave V toneburst and narrow band CE-chirp1
1557
Table 1
Means and standard deviations and the difference between both stimuli for wave V latency (ms) for 500, 1000, 2000 and 4000 Hz toneburst and narrow band CE-chirp1.
Intensity (dB nHL)
Stimuli
500 Hz
1000 Hz
2000 Hz
4000 Hz
Average
SD
Average
SD
Average
SD
Average
SD
80
Toneburst
NB CE-chirp1
Difference
36
36
36
8.57
2.91
5.66
0.51
0.53
0.44
34
34
34
6.90
6.45
0.45
0.30
0.34
0.36
38
38
38
8.46
5.18
3.29
0.37
0.39
0.55
37
37
37
7.17
5.57
1.60
0.49
0.74
0.45
60
Toneburst
NB CE-chirp1
Difference
36
36
36
9.76
3.57
6.19
0.69
0.70
0.74
34
34
34
7.48
7.02
0.46
0.38
0.30
0.34
38
38
38
9.40
5.91
3.49
0.52
0.31
0.62
37
37
37
7.82
6.48
1.34
0.57
0.66
0.29
40
Toneburst
NB CE-chirp1
Difference
36
36
36
11.54
5.26
6.28
0.84
0.97
1.30
34
34
34
8.13
7.80
0.33
0.42
0.29
0.35
38
38
38
10.68
7.37
3.31
0.50
0.47
0.68
37
37
37
8.51
7.42
1.09
0.64
0.76
0.27
20
Toneburst
NB CE-chirp1
Difference
36
36
36
13.19
7.05
6.14
0.89
0.61
1.16
34
34
34
9.05
8.67
0.39
0.59
0.28
0.55
38
38
38
12.32
8.81
3.51
0.58
0.62
0.76
37
37
37
9.56
8.47
1.09
0.49
0.69
0.34
Fig. 1. Latency intensity function of the wave V observed in tonebursts and narrow band CE-chirps1 ABR at 500, 1000, 2000 and 4000 Hz.
Table 2
Means and standard deviations and the difference between both stimuli for wave V amplitude (mV) for 500, 1000, 2000 and 4000 Hz toneburst and narrow band CE-chirp1.
Intensity (dB nHL)
Stimuli
500 Hz
1000 Hz
Average
2000 Hz
SD
Average
4000 Hz
SD
Average
SD
Average
80
Toneburst
NB CE-chirp1
Difference
36
36
36
0.304
0.225
0.079
0.094
0.062
0.092
38
38
38
0.232
0.233
0.001
0.080
0.102
0.089
37
37
37
0.290
0.263
0.028
0.083
0.115
0.138
34
34
34
0.215
0.235
0.021
SD
0.075
0.087
0.096
60
Toneburst
NB CE-chirp1
Difference
36
36
36
0.181
0.189
0.008
0.054
0.076
0.071
38
38
38
0.145
0.205
0.060
0.063
0.064
0.091
37
37
37
0.240
0.306
0.066
0.073
0.072
0.070
34
34
34
0.148
0.257
0.110
0.050
0.073
0.076
40
Toneburst
NB CE-chirp1
Difference
36
36
36
0.133
0.168
0.035
0.028
0.041
0.045
38
38
38
0.113
0.163
0.050
0.042
0.032
0.049
37
37
37
0.153
0.250
0.097
0.034
0.062
0.051
34
34
34
0.122
0.209
0.087
0.036
0.061
0.064
20
Toneburst
NB CE-chirp1
Difference
36
36
36
0.109
0.148
0.039
0.019
0.044
0.043
38
38
38
0.100
0.142
0.042
0.014
0.039
0.041
37
37
37
0.125
0.166
0.041
0.032
0.043
0.043
34
34
34
0.102
0.156
0.055
0.035
0.046
0.048
1558
0,60
0500 Hz
1000 Hz
2000 Hz
4000 Hz
NB CE-chirp
Toneburst
0,45
Amplitude(uV)
0,30
0,15
0,00
0,60
0,45
0,30
0,15
0,00
Intensity (dBnHL)
80
60
40
20
80
60
40
20
Fig. 2. Box-plots to narrow band CE-chirp1 and toneburst wave V amplitude by intensity at frequencies of 500, 1000, 2000 and 4000 Hz.
Fig. 3. 500, 1000, 2000 and 4000 Hz narrow band CE-chirp1 and toneburst ABR recording from an infant with 2 months (left ear).
500 Hz than higher frequencies and this is consistent with the low
frequency response arising from a more apical region of the
cochlea [1619]. However, when the narrow band CE-chirp1
stimulus was used, the opposite occurred. With narrow band CEchirp1 stimulus the ABR latencies are shorter than toneburst
latencies, especially at low frequencies. So as the frequency
decreases, so does the latency. The low stimuli showed shorter
latencies than the high stimuli. Similar results showed click
latencies longer than chirp latencies at 50 dB nHL [20] and, with
our recent study that the CE-chirp1 latencies were shorter than
click latencies [21].
Considering the CE-Chirp1 design, and the fact that the 500,
1000, 2000 and 4000 Hz narrow band CE-chirp1 constitute a
subset of the CE-Chirp1, it is expected that the narrow band CEchirp1 latencies are shorter than toneburst latencies, especially for
the low frequency stimuli. The narrow band CE-chirps1 were
presented in such a way that their temporal references (0 ms)
corresponded to the estimated time of arrival at the tympanic
membrane of the 8000 Hz component. This way, these frequency
components will arrive earlier, resulting in shorter response
latencies [20]. The latency differences between narrow band CEchirps1 when compared with tonebursts stimuli are also consistent with the expected travel times of the narrow band CE-chirp1
will cause maximum stimulation into the cochlea. Therefore, each
of the narrow band CE-chirp1 has a determined timing in order to
compensate for the cochlea traveling delay at centered regions at
500, 1000, 2000 and 4000 Hz and this fact, combined with the
latency reference (0 ms), the observed ABR latencies for the low
frequency stimuli (500 and 1000 Hz) are much shorter than those
evoked by toneburst stimuli [20]. Fig. 3 can illustrate these latency
differences.
Few studies have investigated frequency specic chirps. Bell
et al. [22] generated two chirps derived from the Dau et al. chirp
[3]. The high frequency chirp (30006000 Hz) and a low-frequency
chirp (375750 Hz) were presented at sensation levels between 10
and 50 dB to 10 adult subjects. Their results were similar to this
study, showing that wave V latency decreased as stimulus
frequency decreased. From a clinical perspective, latencies shorter
than those usually observed in recording the toneburst ABRs does
not constitute a problem for clinical applications in infants,
however, normative latency-intensity data to narrow band CEchirps1 should be established.
The narrow band CE-chirps1 evoked larger wave V amplitudes
for all frequencies and levels tested, except at 80 dB nHL and at
60 dB nHL for 500 Hz, when the difference between the two stimuli
was not signicant. Fig. 3 shows the ABR evoked by the tonebursts
and narrow band CE-chirps1. The waves V were clearly detectable
in all conditions stimuli and are larger than the toneburst ABR
amplitude. These results are in accordance with previous studies
comparing click and chirp amplitudes [35]. In our recent study
comparing CE-chirp1 and click wave V amplitudes, we observed
larger amplitudes for CE-chirps1 than for clicks for all conditions
level, except at 80 dB nHL [21]. So, as the narrow band CE-chirps1
constitute a subset of the CE-Chirp1, as expected that narrow band
CE-chirps1 evoked larger wave V amplitudes. The presence of the
larger wave V found in the current study may be a consequence of
the simultaneous depolarization of frequency specic region of the
cochlea, which would be expected from the frequencydelay
characteristics of the narrow band CE-chirps1 stimuli.
At high levels, in some times the toneburst stimulus evoked
higher amplitude responses than narrow band CE-chirps1 (i.e. at
80 dB nHL for 500 Hz) or the amplitude difference between the two
stimuli was not signicant (i.e. at 60 dB nHL for 500 Hz and at
80 dB nHL for 1000, 2000 and 4000 Hz). Similar results were
presented by Wegner and Dau [23]. A low-frequency chirp and a
250 Hz tone pulse with comparable duration and magnitude
1559
1560
[10] M. Don, C.W. Ponton, J.J. Eggermont, B. Kwong, The effects of sensory hearing loss
on cochlear lter times estimated from auditory brainstem response latencies, J.
Acoust. Soc. Am. 104 (1998) 22802289.
[11] E. Sturzebecher, M. Cebulla, C. Elberling, T. Berger, New efcient stimuli for
evoking frequency-specic auditory steady-state responses, J. Am. Acad. Audiol.
17 (2006) 448461.
[12] C. Elberling, M. Don, A direct approach for the design of chirp stimuli used for the
recording of auditory brainstem responses, J. Acoust. Soc. Am. 128 (5) (2010)
29552964.
[13] C. Elberling, M. Don, Quality estimation of averaged auditory brainstem
responses, Scand. Audiol. 13 (1984) 187197.
[14] M. Don, C. Elberling, Evaluating residual background noise in human auditory
brainstem responses, J. Acoust. Soc. Am. 96 (1994) 27462757.
[15] M. Don, C. Elberling, Use of quantitative measure of ABR peak amplitude and
residual background noise in the decision to stop averaging, J. Acoust. Soc. Am. 99
(1996) 491499.
[16] O. Yamada, H. Ashikawa, K. Kodera, H. Yamane, Frequency-selective auditory brainstem response in newborns and infants, Arch. Otolaryngol. 109 (1983) 7982.