Kamakshi V. Gopal Ph.D.

1,
Jacek Kowalski Ph.D.2 Slope analysis of Auditory
Brainstem Responses in
children at risk of central
auditory processing disorders

1
Department of Speech and Hearing
Sciences, University of North Texas,
Denton, USA; 2Department of
Physics, University of North Texas,
Denton, USA
KEY WORDS:
auditory brainstem responses, central
auditory processing disorders, slope
vectors, orphological analysis, cluster
technique
RECEIVED/ACCEPTED:
July 30, 1998/September 22, 1998
ADDRESS FOR CORRESPONDENCE:
Kamakshi V. Gopal, Ph.D.,
Department of Speech and Hearing
Sciences, University of North Texas,
P.O. Box 305010, Denton TX 76203-
5010, USA. Tel: ‡ 940 565 2262,
fax: ‡ 940 565 4058, e-mail:
gopal@unt.edu
The method of slope vectors was used to quantify Auditory
Brainstem Responses (ABR) obtained from nine normal
children and nine children at risk for central auditory
processing disorders (CAPD) with language impairment, for
monaural and binaural stimulation conditions. Slopes thus
obtained were subjected to K-Means Cluster Analysis.
Distinction between the two groups was obtained only for
binaural stimulation conditions, wherein all normal children
were grouped under cluster 1 with higher slope values and 6
out of 9 CAPD children were grouped under cluster 2 with
lower slopes. The results suggest that there may be several
subcategories among children who are found to be at risk for
CAPD. One of the subcategories may comprise children who
exhibit poor ABR morphology, especially during binaural
stimulation conditions, which could be due to binaural
interference.
Scand Audiol 1999;28:85–90

Introduction ABR waveform morphology refers to the

In our previous article (Gopal & Pierel 1998),
which compared Auditory Brainstem Response
(ABR) waveforms between normal children
(control group) and children at risk for central
auditory processing disorders (CAPD group), we
noticed that the grand average ABR waveform
of the control group differed morphologically
from that of the CAPD group (Fig. 1). Individual
data indicated that monaural and binaural ABR
recordings from a majority of the children in the
control group exhibited sharp peaks, while a
number of subjects in the CAPD group displayed
shallower and less peaked waveforms. This
difference was observed in spite of no significant
differences between the two groups in hearing
sensitivity, middle ear function, IQ, and ABR
latency and amplitude values.
 1999 Scandinavian University Press
clarity, resolution and definition of the ABR
(Schwartz et al. 1994). The ABR morphology is
influenced by factors such as the stimulus
characteristics, recording montage, anatomical
variants, hearing loss and neuropathology
(Jacobson & Hall 1994). Morphological analysis
of ABR waveforms has not received wide
attention. For the most part, interpretation of an
ABR waveform is based on subjective criteria,
such as clarity of the waveform. However, when
an ABR waveform does not match the clin-
ician’s expectation of normal appearance despite
normal latency and amplitude values, the mor-
phology is judged to be poor (Hall 1992). The
morphological pattern of an ABR varies with
age, maturation and pathological conditions
even when stimulus and recording conditions
are controlled. Typically morphological analysis
86
Fig. 1. Grand average of right ear, left ear and
binaural ABR waveforms of the control group and the
CAPD group.
of ABRs is limited to the presence or absence of
waveforms. If reliable and valid methods of
morphological analysis are developed, gross and
subtle alterations in ABR waveform morphology
may lead to identification of differences in the
brainstem responses among subjects.
In a normal auditory system, binaural auditory
stimulation provides greater loudness (Haggard
& Hall 1982) and better speech understanding
(Harris 1965, Dirks & Wilson 1969, Kaplan &
Pickett 1981, Nabelek & Robinson 1982).
Speaks (1975) has indicated that binaural hear-
ing is superior to monaural hearing because of
stronger and more abundant contralateral path-
ways. Several studies have shown that, with
binaural stimulation, the amplitudes of various
auditory evoked potentials are enhanced com-
pared with those evoked by monaural stimula-
tion (Ainslie & Boston 1980, Debruyne 1984,
Kelly-Ballweber & Dobie 1984, McPherson et
al. 1989, McPherson & Starr, 1993, 1995, Jiang
& Tierny 1996a). These studies have also shown
that the binaural amplitude is smaller than the
algebraic sum of the two monaural responses in
people with normal hearing.
Not all people exhibit binaural advantage with
binaural stimulation. There is behavioral data,
primarily from speech discrimination test scores
and electrophysiological data obtained from
ABRs and middle latency responses indicating
that in some individuals auditory binaural
stimulation brings about interference (Arkebauer
et al. 1971, Hood & Prasher 1990, Jerger et
al. 1993, Silman 1995). These studies have
reported that in certain cases, during binaural
stimulation conditions, the presence of the
poorer ear input interferes with the better ear
input, leading to a substantially poorer binaural
⋅ ⋅ ⋅
Scandinavian Audiology Vol 28 No 2 1999
KV Gopal and J Kowalski
Fig. 2. Schematic representation of an ABR wave-
form with peaks labeled I–VII. The inset shows
quantities used in calculations of average ascending
and descending slopes of the i-th wave.
score compared with the score obtained from the
better ear.
Jerger et al. (1993) described binaural inter-
ference as a condition in which the poorer ear
response may hinder the better ear response,
leading to reduced binaural performance scores.
They further postulated that this condition could
be the result of a peripheral or central disorder
leading to binaural interference. Binaural inter-
ference, if central, could explain to some extent
central auditory processing disorders.
In this article we have used slope vectors to
quantify the differences in ABR waveform
morphology between monaural and binaural
conditions in normal children and children at
risk for CAPD, and further classified the slopes
using the K-Means Cluster Analysis technique.
Material and methods
Our data consisted of ABR waveforms obtained
from two groups of children between the ages of
7 and 13 years. The control group consisted of 9
normal children and the experimental group
(CAPD group) consisted of 9 children at risk
for CAPD. All children were screened for
normal hearing, normal middle ear function,
normal IQ and normal ABR thresholds. Children
in the experimental group had failed the SCAN
or the SCAN-A test and were considered to be
at risk for CAPD (Keith 1986). These children
also had low academic performance as reported
by their parents and/or teachers. They were
diagnosed with moderate to severe language
impairment associated with auditory processing
difficulties by certified speech-language pathol-
ogists, and were enrolled in language remedia-
tion therapy for a period of 2–3 years. Children
Slope analysis of ABRs 87

in the control group had no reported difficulty at where Vi,max is the maximum (peak) potential
school, obtained normal scores on the SCAN/ observed in the i-th wave, Vi,l and Vi,r are the
SCAN-A test and had no speech-language potentials at the immediately preceding (left)
problems. and succeeding (right) minima, and Dti‡ and Dti‡
The basic instrumentation and procedures are durations of the i-th wave ascent and descent
adopted in the collection of ABRs have been (see inset in Fig. 2). We will call ri‡ and riÿ the
described in our previous article (Gopal & Pierel ascending and descending slopes of the i-th
1998). In brief, signed consent forms from wave. Each wave is then represented by an
subjects and their parents were obtained prior ordered pair (ri‡ , riÿ ) and a selected subsequence
to any testing. ABR recordings were made in an of n waves from a waveform is characterized
electrically shielded sound-treated room. Gold in terms of a 2n-dimensional slope vector
cup electrodes were placed on the vertex, nape with components (r1‡ ,r1ÿ ,r2‡ ,r2ÿ ,…,rn‡ ,rn‡ ). The
of the neck and forehead. Acoustic stimuli selected sequence may not include all observed
consisted of alternate clicks presented at 55 dB waves but it is always assumed that the
above the monaural ABR threshold, delivered components of the slope vector are chronologi-
through headphones at a repetition rate of cally ordered.
11.1/s. ABRs were recorded on a Modular The proposed reduced description of an ABR
Instruments Inc. system using standard protocols waveform is clearly invariant with respect to
for right, left and binaural stimulation condi- uniform re-scaling of the ABR potential values
tions. Computation of a reduced representation and conduction times. This reflects an assump-
of an ABR waveform by slope vectors is tion that much of the essential informational
described below. content of an individual ABR signal is predomi-
Consider an ABR signal with its typical nantly encoded in the instantaneous increase
sequence of constituent waves labeled in a and decrease rates and not in absolute value of
standard way as waves I to VII based on Jewett the ABR potential and conduction times which
& Williston (1971) nomenclature. may vary from individual to individual. Accord-
Assume that basic characteristics of the i-th ingly, the slope vectors of two individuals with
wave (i = I, II … VII) are its average rates of similar auditory processing features should be
increase, ri‡ , and decay, riÿ , defined as: represented as two close points in the slope
vector space. A possible classification scheme
Vi;max ÿ Vi;l ÿ Vi;max ÿ Vi;r into two or more types of ABR responses may
ri‡ ˆ ; ri ˆ

ti‡
tiÿ then be based on one of the well-developed
clustering algorithms, where each cluster in the
Table I. (A) Mean right ear ABR slopes for peak I descending, and peaks II through V for ascending and
descending conditions. (B) Mean left ear ABR slopes for peak I descending, and peaks II through V for
ascending and descending conditions. (C) Mean binaural ABR slopes for peak I descending, and II through V
for ascending and descending conditions. Note: peak I ascending slope was not calculated since it was difficult
to identify the correct trough preceding peak I in many cases
(A)
Groups I des. II asc. II des. III asc. III des. IV asc. IV des. V asc. V des.
Control 0.469 1.016 1.018 0.780 0.910 0.804 0.552 0.612 0.947
CAPD 0.425 0.470 0.450 0.400 0.613 0.678 0.600 0.459 0.711

(B)
Groups I des. II asc. II des. III asc. III des. IV asc. IV des. V asc. V des.
Control 0.544 0.733 0.741 0.690 0.720 0.647 0.452 0.658 1.025
CAPD 0.381 0.613 0.512 0.600 0.615 0.660 0.363 0.418 0.859

(C)
Groups I des. II asc. II des. III asc. III des. IV asc. IV des. V asc. V des.
Control 0.739 1.662 1.613 1.494 1.754 1.022 0.761 0.592 1.473
CAPD 0.545 1.210 0.585 0.732 1.007 0.878 0.644 0.780 1.373

⋅ ⋅ ⋅
Scandinavian Audiology Vol 28 No 2 1999
88 KV Gopal and J Kowalski

Table II. Grouping of ascending and descending slopes for right ear, left ear and binaural ABRs using K-Means
Cluster Analysis in normal children from the control group (numbered 1–9) and children at risk for CAPD
(numbered 10–18)
Cluster
(group)
Data selected for cluster analysis number Data points (subjects) in the cluster
Right ear ascending and descending slopes 1 1, 3, 8, 9, 12, 15
2 2, 4, 5, 6, 7, 10, 11, 13, 14, 16, 17, 18
Left ear ascending and descending slopes 1 1, 3
2 2, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18
Binaural ascending and descending slopes 1 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12, 15
2 11, 13, 14, 16, 17, 18

partition contains responses of the same type,
and where distant clusters correspond to radi-
cally different responses.
Furthermore, we restricted analysis to cluster-
ing techniques used in the bi-partition problem,
i.e. an optimal subdivision of a general multi-
variate data set into two, distinct groups. For this
we analysed lumped slope vector data from all
18 subjects (9 normal subjects and 9 subjects at
risk for CAPD) with a widely used K-Means
Cluster Analysis algorithm, a standard compo-
nent of many statistical software packages. This
cluster algorithm was applied to ascending and
descending slopes obtained for left ear, right ear
and binaural ABRs from all 18 subjects. Chil-
dren from the control (normal) group were
assigned a number from 1 to 9. Children from
the CAPD group were assigned a number from
10 to 18.
Results
Representation of ABR waveforms by slope
vectors indicated that children in the control
group generally had higher slopes than children
in the CAPD group. There were however,
several exceptions when individual data was
analysed for all the different conditions. Table I
shows group mean ABR slopes obtained for
right ear, left ear and binaural stimulation
conditions. A total of 27 conditions comprising
peak I descending slopes, peaks II through V
ascending and descending slopes from both
groups, were used for further analysis.
Table II depicts the results from K-Means
Cluster Analysis performed on monaural (right
and left) and binaural ascending and descending
slopes. Results indicated that for the right and
left monaural conditions the cluster analysis
failed to group the data in any systematic
fashion. We obtained the best, though not crisp,
partition of our small database for nine-dimen-
sional slope vectors for peaks I through V for
binaural ABR waveforms. With subjects num-
bered 1 to 9 in the control group and 10 to 18 in
the CAPD group, all subjects from the control
group were lumped by the algorithm in the same
cluster. The complementary cluster was com-
posed of six children from the CAPD group.
Three children from the CAPD group were
misclassified as “normal” (i.e. put into the first
cluster).
Discussion
Jerger et al. (1993) presented case studies
exhibiting behavioral and electrophysiological
auditory binaural interference. In their electro-
physiological measures, they did not show any
evidence of interference in the auditory brain-
stem responses. However, middle latency re-
sponses showed binaural interference. Jerger et
al. (1993) have also discussed a similar effect in
the visual system. Although binocular stimula-
tion is believed to provide improvements in
people with normal vision, it is shown to bring
about binocular rivalry in patients with strabis-
mus, congenital cataract or astigmatism, leading
to amblyopia. Hence, in people with amblyopia,
binocular responses bring about poorer re-
sponses compared with monocular stimulation
due to binocular interference or binocular
inhibition.
Analysis of group data from monaural and
binaural slopes for ascending and descending
conditions, showed that 23 out of 27 conditions
had lower slopes for the CAPD group compared
with the normal group. Based on the slope vector
analysis, in any given case, the slope gets lower
when the amplitude of the peak potential is low,

⋅ ⋅ ⋅
Scandinavian Audiology Vol 28 No 2 1999
Slope analysis of ABRs
or if the duration of the wave ascent and descent
is lengthened. Inspection of our data set suggests
that both factors, i.e. reduced amplitude of peak
potentials and increased time duration of waves
may have contributed towards the low slope
values obtained in the CAPD group.
In sensory evoked potential recordings the
initial positive response is thought to reflect
excitation of neurons produced by input from the
lower centers. The later negative response is a
result of the elaboration of excitatory input in the
surrounding area followed by inhibition (Martin
1985). In ABRs a decrease in the amplitude of a
peak potential may occur due to:
1. a decrease in the number of neurons con-
tributing to the response;
2. fewer fibers conducting the volley;
3. a decrease in synchrony of active neurons;
4. a decrease in the field potential generated in
the auditory pathway (Boettcher et al. 1993,
Jiang and Tierny 1996b).
It is interesting to note that reduced slope
values in the CAPD group were obtained for
ABR waves I through V. We speculate that the
reduced ABR slopes observed in our CAPD
group could most likely be due to one or more of
the following reasons:
1. reduced number of auditory nerve cells
synchronously firing in response to the
incoming signal at the level of the cochlea,
which in turn results in reduced amplitude in
succeeding waves;
2. reduced number of fibers conducting the
volley through the brainstem;
3. abnormality in the volume conduction
mechanism, which is the underlying theory
in evoked potential recordings.
Individual data analysed using K-Means
Cluster Analysis for right ear and left ear slopes
separately did not show any systematic classifi-
cation. However, when the data for binaural
ABR slopes were analysed, the classification of
subjects was closer to the original grouping of
our subjects. All normal subjects were grouped
into cluster 1. Among the subjects at risk for
CAPD, six out of nine were grouped into cluster
2. The remaining three from the CAPD group
were misclassified.
We interpret our results to suggest that even in
our limited data sample of the children at risk for
CAPD, a subgroup may exist demonstrating
89
binaural interference, which may be contributing
to their auditory processing difficulties. Further-
more, these children may be identified by their
poor ABR morphological characteristics exhib-
ited during binaural stimulation condition. These
differences, if replicated, may lead to further
exploration of CAPD subgroup differences using
other methodological analyses.
Morphological analysis proposed in this paper
based upon bi-partition of slope vectors, looks
promising as a diagnostic tool. We believe that
clustering techniques, if applied to a sufficiently
large lumped database obtained from both
groups, may give a useful diagnostic algorithm.
When two large, well-separated clusters are
formed in the data set, any new data point can
be assigned to one of the groups by simply
comparing its distances from both clusters in the
slope space.
Acknowledgement
Supported in part by a faculty research grant
from the University of North Texas, Denton,
Texas.
References
1. Ainslie PJ, Boston JP. Comparison of brain stem
auditory evoked potentials for monaural and binaural
Stimuli. Electroencephalogr Clin Neurophysiol 1980;
49:291–302.
2. Arkebauer HJ, Mencher GT, McCall C. Modification of
speech discrimination in patients with binaural asym-
metrical hearing loss. J Speech Hear Disord 1971;36:
208–12.
3. Boettcher FA, Mills JH, Norton BL. Age-related
changes in auditory evoked potentials of gerbils. 1.
Response amplitudes. Hear Res 1993;71:137–45.
4. Debruyne F. Binaural interaction in early, middle and
late auditory evoked responses. Scand Audiol 1984;
13:293–6.
5. Dirks D, Wilson R. Binaural hearing for aided and
unaided conditions. J Speech Hear Res 1969;12:650–
64.
6. Gopal KV, Pierel K. Binaural interaction component in
children at risk for central auditory processing dis-
orders. Scand Audiol 1999;28:77–84.
7. Haggard M, Hall J. Forms of binaural summation and
the implications of individual variability for binaural
hearing aids. Scand Audiol 1982;5 (Suppl):47–63.
8. Hall JW Handbook of auditory evoked responses.
Boston: Allyn and Bacon, 1992.
9. Harris J. Monaural and binaural speech intelligibility
and stereophonic effects based upon temporal cues.
Laryngoscope 1965;75:428–46.
10. Hood J, Prasher D. Effect of simulated bilateral
⋅ ⋅ ⋅
Scandinavian Audiology Vol 28 No 2 1999
90
cochlear distortion on speech discrimination in normal
subjects. Scand Audiol 1990;19:37–41.
11. Jacobson JT, Hall JW. Newborn and infant auditory
brainstem response applications. In: Jacobson KT,
editor. Principles and applications in auditory evoked
potentials. Boston: Allyn and Bacon, 1994.
12. Jerger J, Silman S, Lew HL, Chmiel R. Case studies in
binaural interference: converging evidence from beha-
vioral and electrophysiologic measures. J Amer Acad
Audiol 1993;4:122–31.
13. Jewett DL, Williston JS. Auditory evoked far fields
averaged from the scalp of humans. Brain 1971;94:
681–96.
14. Jiang ZD, Tierney TS. Binaural interaction in human
neonatal auditory brainstem. Pediatr Res 1996a;39:
708–14.
15. Jiang ZD, Tierney TS. Long-term effect of perinatal
and postnatal asphyxia on developing human auditory
brainstem responses: brainstem impairment. Pediatr
Oto Rhino Laryngol 1996b;34:111–27.
16. Kaplan H, Pickett J. Effects of dichotic/diotic versus
monotic presentation on speech understanding in noise
in elderly hearing-impaired listeners. Ear Hear 1981;
2:202–7.
17. Keith RW SCAN: a screening test for auditory
processing disorders. San Antonio: The Psychology
Corporation, 1986.
18. Kelly-Ballweber D, Dobie RA. Binaural interaction
measured behaviorally and electro-physiologically in
young and old adult. Audiology 1984;23:181–94.
⋅ ⋅ ⋅
Scandinavian Audiology Vol 28 No 2 1999
KV Gopal and J Kowalski
19. Martin JH. Cortical neurons, the EEG, and the mech-
anisms of epilepsy. In: Kandal ER, Schwartz JH,
editors. Principles of neural science. II Ed. New York:
Elsevier, 1985.
20. McPherson DL, Tures C, Starr A. Binaural interaction
of the auditory brain-stem potentials and middle latency
auditory evoked potentials in infants and adults.
Electroencephalogr Clin Neurophysiol 1989;74:124–
30.
21. McPherson DL, Starr A. Binaural interaction in
auditory evoked potentials: Brainstem, middle- and
long-latency components. Hear Res 1993;66:91–8.
22. McPherson DL, Starr A. Auditory time-intensity cues
in the binaural interaction component of the auditory
evoked potentials. Hear Res 1995;89:162–71.
23. Nabelek A, Robinson P. Monaural and binaural speech
perception in reverberation for listeners of various ages.
J Acoust Soc Am 1982;71:1242–8.
24. Schwartz DM, Morris MD, Jacobson JT. The normal
auditory brainstem response and its variants. In:
Jacobson JT, ed. Principles and applications in auditory
evoked potentials. Boston Massachusetts: Allyn and
Bacon, 1994.
25. Silman S. Binaural interference in multiple sclerosis:
Case study. J Amer Acad Audiol 1995;6:93–6.
26. Speaks CE. Dichotic listening. A clinical or research
tool? In: Sullivan MD, editor. Proceedings of a
symposium on central auditory processing disorders.
Omaha Nebraska: University of Nebraska Medical
Center, 1975.
Copyright of Scandinavian Audiology is the property of Scandinavian University Press
International Division and its content may not be copied or emailed to multiple sites or posted
to a listserv without the copyright holder's express written permission. However, users may
print, download, or email articles for individual use.