(Reprinted from Nature, Vol. 241, No. 5390, pp. 468- 469 February 16 1973)
Defects of Non-Verbal Auditory Perception in Children
with Developmental Aphasia
SOME otherwise normal children fail to learn to speak and are designated developmental aphasics.
Several authors have suggested that auditory perceptual deficits, particularly of sequencing, may be the
primary dysfunction1-4. Efron5 suggested that the left temporal lobe mediates temporal analysis and that it
is the disruption of this function which is central to adult aphasia. We examined children with
developmental aphasia and demonstrated inferior discrimination of sound quality to which a sequencing
difficulty could be secondary,
The subjects were nine boys and three girls, aged 6.8 to 9.2 years, from the John Horniman School for
Aphasic Children and twelve normal children matched for age, sex and nonverbal intelligence. The
aphasics were free from other signs of neurological dysfunction, had normal hearing, and a mean
non-verbal IQ of 108 (85-127).
Two different sounds were presented sequentially and subjects were examined for their ability to
indicate (a) the order in which the sounds occurred or (b) whether the two sounds were the same or
different. Stimuli were two 75 ms complex tones of frequencies within the speech range (fundamental
frequency: Tone 1-100 Hz, Tone 2-305 HZ; rise/fall time 40µs-1). Normal speech rates approach 80 ms
per phoneme. Stimuli were generated by a 'Speech Synthesizer' and 'Modular One' computer, recorded
onto a 'Uher' tape recorder and played to subjects through earphones. Subjects responded on two identical
depressible panels.
We used two test methods. In the repetition method subjects were trained to respond to each tone
separately by pushing the left panel to Tone I and the right to Tone 2. Discrimination training continued
until criterion was reached (20 of 24 consecutive responses correct, P <0.001 binomial test). Subjects
were then trained to respond to each of the four possible two-tone sequences (1-1,1-2,2-1,2-2) by pushing
the panels in the corresponding order. The inter-sound-interval (ISI) was 428 ms. Correct responses were
demonstrated four times by the experimenter followed by eight training trials in which subjects received
knowledge of results and error correction. Next, 24 trials without knowledge of results were given.
Subjects were then tested on a total of 48 two-tone sequences with ISI’s of 8, 15, 30, 60, 150, 305, 947,
1,466, 1,985, 3,023, 3,543 and 4,062 ms, all presented in random order. A similar procedure was used
with the same two tones for sequences of three and four elements (ISI 428 ms),
Because a subject may perceive the elements of a temporal sequence but not be able to reproduce a
corresponding motor pattern, a same-different method was also used6. The response panel was turned
through 90° to avoid confusion between methods. Subjects were initially presented with the two tones
(ISI 428 ms) and trained to press the top panel if the tones were the same and the bottom panel if different.
Training continued until the criterion described above was reached.
The same series of two-tone sequences, 24 with ISI 428 ms and 48 with ISI varied as above, were again
presented and the subject indicated whether the two tones were the same or different. Half of the subjects
in each group performed the repetition task first; half performed the same-different task first.
On the repetition task, all the aphasics and controls were able to reach criterion both on the initial
auditory discrimination and on the reproduction of the two tones separated by 428 ms. However, when
the interval between the two tones was decreased, the aphasics' performance was markedly impaired
(Table I). On the repetition test, whereas the controls as a group performed significantly better than
chance (P< 0.001) at the shortest ISI (8 ms), this level of performance was not
achieved by the aphasics at ISIs less than 305 ms. By contrast, increasing the ISI had no adverse effect on
aphasics' performance. They performed with undiminished efficiency (P < 0.001) when the interval was
increased up to 4 s. When the number of elements in the sequence was increased to three, only two of the
aphasics reached criterion. All controls reached criterion on both three and four element sequences.
Results for the same-different task showed a similar pattern.
Both groups performed accurately (P< 0.001) at 428ms ISI but, when the interval between tones was
decreased, controls remained virtually errorless at all intervals whilst the aphasics again required at least
305 ms to perform at the level of accuracy defined above.
Correct performance on the same-different task requires discrimination of the two tones but not
identification of the order in which they occur. The ISI at which the aphasics fail to reach the criterion of
P < 0.001 on a binomial test is the same for both the repetition method (which does require perception of
order) and the same-different method (which does not). Accordingly, Efron's and Lowe and Campbell's
observations that aphasics exhibit impairment of auditory perception of rapidly presented sequences may
be explicable in terms of a failure of discrimination of auditory quality.
PAULA TALLAL
P. T. is supported by the American Association of University Women.
M. PIERCY
Department of Experimental Psychology ,
Downing Street,
Cambridge
Received July 31; revised November 3, 1972.
1
Eisenson,J. J Speech Hearing Dis.,33,3(1968).
Benton, A. L., Cortex, 1, 40 (1964).
2
3
Stark, J., J. Commun. Dis., 1, 31 (1967).
Lowe, A. D., and Campbell, R. A., J. Speech Hearing Res., 8, 313 (1965).
5
Efron, R., Brain, 86, 403 (1963).
6
Konorski, J., Bull. Acad. Pol. Sci., 7, 115 (1959).
4