DIGIT REPETITION PERFORMANCE IN PATIENTS

WITH FOCAL BRAIN DAMAGE

F. William Black and Richard L. Strub

(Department of Neurology, LSU Medical Center, New Orleans, Louisiana)

INTRODUCTION

The digit repetition test is commonly used to assess verbal short term
memory and is included in both standard intelligence tests and mental status
examinations (Ruesch and Moore, 1943; Heilbrun, 1958). The supposition
underlying the use of this test is that it reflects an important area of intellec­
tual functioning which is highly vulnerable to brain damage. There is, how­
ever, research evidence which casts doubt upon the usefulness of the digit
span test in differentiating organic patients from normals. Sterne (1969) was
unable to demonstrate significant differences in digit repetition between brain­
damaged and non-brain-damaged patients. As a specific clinical example, there
are patients with particular organic amnestic syndromes (Korsakoff's syndro­
me) who have been shown to have normal digit span performance (Milner,
1962, Drachman and Arbit, 1966 ).
Recent research interest has focused upon the possibility of localizing
focal lesions from digit span performance. The currently available data regard­
ing the digit repetition of patients with unilateral lesions is contradictory.
McFie (1969) and Newcombe (1969) reported that digit span performance
is impaired in patients with left hemisphere lesions, irrespective of the locus
of the lesion within the hemisphere. De Renzi and Nichelli (1975) stated
that verbal memory span (as tested by digit repetition) is directly dependent
upon the intactness of the left hemisphere, with impaired performance being
the result of aphasia. Digit repetition performance was not affected by damage
to the right hemisphere wherever localized. Weinberg, Diller, Gerstman
and Schulman ( 1972) found that patients with left hemisphere lesions per­
form like poorly educated normals on both forward and backward digit
repetition. In their right hemisphere lesion patients, digit span forward was
superior to that of left hemisphere lesion patients, and did not differ
significantly from that of normal controls. The right hemisphere lesion pa­
tients with visuospatial dysfunction (as assessed by a visual confrontation

Cortex (1978) 14, 12-21.

 Digit repetition p~formance 13

task) performed significantly poorly on digit span backward with no corres­

ponding deficit on digit span forward. This impairment was hypothesized to
be due to deficits in visual scanning or to difficulties in eye movements. Co­
sta (1975) reported no significant differences between samples of patients
with right and left hemisphere lesions and normals on a test of digit span
forward, but did find significantly lower scores for the brain-damaged patients
on digit span backward. This discrepancy in findings suggested that digit
span backward is more sensitive to the cognitive dysfunction resulting from
any brain lesion than is digit span forward. Although Costa found no signi­
ficant differences in either digit span forward or backward between left and
right hemisphere lesion samples, he did demonstrate statistically reliable dif­
ferences in digit span backward when the patients were categorized according
to the presence or absence of visual spatial deficits. This finding tends to
support Weinberg et al.'s (1972) hypothesis that the ability to repeat digits
backward is related to intact visual scanning.
In a study of learning disabled children, Rudel and Denckla ( 197 4) found
significantly more large digit span forward-backward discrepancies in patients
with presumable right hemisphere dysfunction than in patients with corres­
ponding left hemisphere dysfunction. The discrepancies in these patients were
due primarily to impaired digit span backward performance by the right
hemisphere patients.
Further clinical data is needed to document whether: (1) the digit repe­
tition test can effectively dichotomize brain-damaged patients from normals;
(2) immediate verbal recall as assessed by digit repetition is uniquely a left
hemisphere function, with deficits accounted for by the presence of aphasia
as has been suggested by De Renzi and Nichelli (1975); (3) digit span back­
ward is more highly vulnerable to the effects of brain damage in general
and right hemisphere damage specifically; ( 4) impaired digit span backward
is associated with defective visual spatial functions (Costa, 1975; Weinberg
et al., 1972); and (5) the impaired ability to repeat digits is a reflection
of general reduced intellectual ability or other associated cognitive deficits as
hypothesized by Costa (1975).
Accordingly, this study investigated the digit repetition (forward and
backward) performance of previously healthy young adult subjects with re­
cently acquired discrete brain lesions. Digit span performance was also cor­
related with general intelligence, memory, visual constructional ability, and
performance on previously used measures of verbal abstraction (WAIS Simi­
larities), and constructional ability (WAIS Block Designs).
14 F. W. Black and R. L. Strub

MATERIALS AND METHODS

Subjects
Clinical subject samples were composed of patients with discrete brain lesiom
secondary to high velocity small shrapnel fragments penetrating the skull and
dura. All injuries were war inflicted. For the purpose of this study, the brain
was divided into quadrants following the schema of Benson and Barton ( 1970 ),
with the fissure of Rolando designated as the anterior-posterior line of demarca­
tion. Seventy-one subjects with lesions restricted to one of the four quadrants
were available from the population of all patients studied with war-related brain
injuries. Fourteen patients were assigned to the right frontal, 20 to the left
frontal, 19 to the right posterior, and 18 to the left posterior samples. The locus
and limitation of the lesions were documented by neurosurgical exploration in
all cases. Considerable care was exercised to exclude any patient with evidence
of bilateral or diffuse lesions. This was accomplished by the use of skull x-ray,
brain scan, EEG, and other neurodiagnostic procedures when deemed necessary
by the referring neurosurgeon. To preclude any contamination of results, poten­
tial subjects with clinically significant vascular or infectious complications of the
initial brain lesion were also excluded from the subject pool. Patients with
significant wounds, infections, or other illness were not considered for selection.
As has been recommended (De Renzi and Nichelli, 1975), aphasia was not used
as an exclusion criterion. On clinical neurological examination, three of the left
frontal quadrant sample, two of the left posterior sample, and one of the right
posterior sample showed some evidence of aphasia. In all such cases, the aphasia
was mild and near-normal communication was possible. All aphasic patients were
easily able to meet Goodglass et al.'s criterion that the patient be able to ade­
quately repeat at least one verbal item (digit) (Goodglass, Gleason and Hyde,
1970).
The diameter of entrance wounds in the four clinical samples was approxi­
mately equal, with a total sample range from .5 to 4.2 em and a median dia­
meter of 2 em. All selected subjects were right handed and were healthy prior
to injury without history of psychiatric or neurological problems.
For purpose of comparison, a 25 subject control sample was selected from
patients hospitalized because of war-related peripheral injuries. None of the con­
trol sample had wounding of the head or upper extremities and showed no
evidence of brain lesions of any etiology.
The age of clinical and control subjects ranged from 18 to 31 years, with
a total sample mean of 21.5 years (SD. = 2.1). Variations in mean age for
the five samples was nonsignificant (F = 1.1). The mean educational level for
the total sample was 11.8 years (S.D. = 1.8), with no significant differences
among sample means (F = .86 ). The recency of injury (time in weeks between
dates of wounding and evaluation) in the five samples also did not differ appre­
ciably (total sample mean= 1.2 months; S.D. = .9).

Tests and procedure

All patients received the W AIS Digit Span subtest under standard condi­
tions. Two trials were given at each length of digit strings, with the test being
terminated upon a failure to accurately reproduce both trials at a given length.
Scores utilized for this study were the maximum number of digits accurately
reproduced forward and the maximum number reproduced backward.
 Digit repetition performance 15

All patients were also administered the WAIS, Wechsler· Memory Scale
(Form II) and the Bender Gestalt Test under standard conditions.
Following the hypothesis that defective backward digit repetition may be
related to a difficulty in revisualization (Weinberg et al., 1972), the total brain­
damaged sample was divided into a group with and a group without visual
constructional deficits. Bender Gesta,h protocols were scored for integration,
distortion, perseveration, and rotation errors according to the Koppitz ( 1964)
system. Twenty-two patients ·made more than one of the above errors and
were classified as demonstrating visual constructional deficits; while the remaining
49 patients made either no errors (N = 31) or only one scoreable error (N =
18) and were classified as not demonstrating visual constructional deficits.

RESULTS

Means and standard deviations for digit span forward and backward
in the four brain-damaged and the control samples appear in Tab~e I.

TABLE I

Means and Standard Deviations of Digit Span Forward and Backward

Digit span forward Digit span backward

Sample
Mean S.D. Mean S.D.

Controls 7.0 1.0 5.0 ·1.3

Left frontal lesion 5.8 1.5 4.2 1.2
Left posterior lesion 4.9 1.3 4.1 1.5
Right frontal lesion 5.5 .9 4.3 1.3
Right posterior lesion 6.2 1.5 4.2 1.3

Digit span forward differed significantly from that of controls in the right
frontal (F = 19.75, p < .001), left frontal (F = 8.82, p < .001), and left
posterior (F = 28.41, p < .001) samples. The mean digit span forward scores
of controls and right posterior lesion samples did not differ significantly (F
= 3.65). The left and right hemisphere lesion samples did not differ appre­
ciably on digit span forward (F = 1.03) nor were there significant differences
among the four quadrant brain-damaged samples (F = .87). Then mean digit
span forward scores of subjects with visual constructional deficits and those
without such deficits did not differ appreciably (F = .29).
The mean digit span backward scores of controls did not differ signifi­
cantly from those of any of the brain-damaged samples (right frontal lesion
sample: F = 2.19; left frontal lesion sample: F = 3.73; right posterior
lesion sample: F = 3.39; and left posterior lesion sample: F = 3.00). The
16 F. W. Black and R. L. Strub

right and left hemisphere lesion samples also did not show significant differ­
ences on digit span backward (F = .43). There were no significant differences
in digit span backward among the four quadrant samples (F = .31). Subjects
with visual constructional deficits did not differ appreciably from subjects
without such deficits (F = .98).

Incidence of significantly impaired digit span performance

Because of the general lack of significant mean differences in digit span
performance (either forward or backward) among the samples studied, the
incidence of .significant (- 2 standard deviations from the control sample
mean) impairments in digit repetition was examined in all samples. The results
appear in Table II.

TABLE II
Incidence of Significantly Impaired DSF and DSB Performance (- 2 Standard Deviations)

Samples Digit span forward Digit span backward

(%) (%)

Controls 0 0
Right frontal lesion 50 0
Left frontal lesion 55 6
Right posterior lesion 45 8
Left posterior lesion 67 5
All brain-damaged 57 5

From inspection, it is apparent that the frequency of impaired per­

formance was significantly higher for digits repeated forward than
backward (chF = 54.6, p < .001). The incidence of impaired digit
repetition both forward and backward were essentially equal in the
four quadrant brain-damaged samples and showed no appreciable statistical
differences among samples.
~~~~ .. f;~.f1' t l. / ! .,~.
~~~·- ~ .
'

Incidence of forward-backward digit repetition discrepancies

Rudel and Denckla (197 4) reported a higher incidence of large ( + 3
point) forward-backward digit repetition discrepancies in learning disabled
children with left sided soft neurological signs (i.e., a presumable right
hemisphere dysfunction). Accordingly, normal (0-2) and large ( + 3) for­
ward-backward digit repetition discrepancies were studied in the carrent sam­
ples. The incidence of large discrepancies ranged from 35% of the control
.subjects to 25% of the right posterior and left frontal lesion subjects, 11%
 Digit repetition performance 17

of the right frontal lesion subjects, and 7% of the left posterior subjects.
The incidence of large discrepancies was significantly lower than that in the
control sample for the right frontal lesion sample (chi2 = 10.81, p < .01)
and left posterior lesion sample (chi2 = 23.62, p < .01). The incidence
of such discrepancies in the right posterior and left frontal lesion samples,
although less than that in the control sample, were not significantly different.
The right and left hemisphere lesion samples showed no significant differ­
ence in the incidence of forward-backward discrepancies (chi2 = 1.42 ).

Correlations with other tests

Correlations of digit span forward and backward with performance on
the WAIS, Wechsler Memory Scale, WAIS Verbal Similarities, WAIS Block
Designs, and Bender Gestalt error scores appear in Table III.

TABLE III

Correlations of Digit Span With Other Cognitive Tests

WAIS WMS WAIS WAIS Bender

Sample I.Q. M.Q. Simil. B.D. Gestalt

Left frontal lesion

DSF .57** .55** .01 .12 -.19
(N = 20)
DSB .32 .52** .38 -.01 -.13
Left posterior lesion
DSF .57** .48* .35 .01 -.30
(N = 18)
DSB .29 .54** .18 .04 -.23
Right frontal lesion
DSF .61* .37 .29 .40 -.20
(N = 14)
DSB .58* .68** .40 .56 -.04
Right posterior lesion
DSF .64',;' .46* .57* .49'' .07
(N = 19)
DSB .84** .95** .60'"' .61** -.23
All brain-damaged
DSF .55** .39** .29* .31* -.12
(N = 72)
DSB .45'"' .62''* .34*'' .35''* -.16

* p < .05; ** p < .01.

As noted from the table, performance on the Wechsler Memory Scale

was significantly related to both digit span forward and backward for vir­
tually all samples, with the exception of digit span forward in the right
18 F. W'. Black and R. L. Strub

frontal lesion sample. WAIS Full Scale I.Q. was significantly related to
digit span forward in all samples and with digit span backward in the right
frontal and posterior lesion samples. Significant correlations among digit
repetition and other cognitive measures were much less consistent, tending
to predominate in the right posterior lesion sample.

DISCUSSION

This data stimulates considerable question regarding the ability of the

digit repetition test to effectively dichotomize brain-damaged from normal
patients. Although the digit span forward scores of the left frontal, right
frontal, and left posterior lesion samples differed significantly from those of
normal controls, this score did not differ appreciably between normals and
the right posterior lesion sample. Similarly, digit span backward scores showed
no significant differences between normal controls and any of the four qua­
drant lesion samples. In contrast to the findings of Costa (1975), the results
of this study suggest that digit span backward, in fact, is less sensitive to
the effects of focal brain damage than is digit span forward. The magnitude
of those few significant differences that were obtained was sufficiently slight
that this task would seem to have little clinical utility in patients with focal
brain lesions.
Only 57% of all brain-damaged patients showed significantly impaired
digit span forward, while only 5% had significant impairment on digit span
backward. As no control patient had a significant impairment of digit repeti­
tion either forward or backward, the clinical finding of a significant decrease
in digit span performance does suggest brain dysfunction. Conversely, a nor­
mal digit span performance does not rule out the possibility of brain da­
mage, as 43% of this sample did have normal digit repetition (false nega­
tives).
Our data would not support the contention of De Renzi and Nichelli
(1975) that digit repetition is uniquely a left hemisphere function. Whereas
digit repetition forward and backward did differ from that of controls in
both the pooled right and left hemisphere lesion samples, there were no
significant difference in either score between the right and left hemisphere
samples. Only six of our patients had clinical evidence of aphasia; their
performance on both digit forward and backward was, however, significant­
ly more impaired than that of the other left hemisphere lesion subjects.
Thusly, performance on this task, while impaired in some patients with
lesions in any of the brain quadrants, is most dramatically impaired in
patients with evidence of aphasia.
We had postulated that digit span backward would be more vulnerable
 Digit repetition performance 19

to brain damage in general and to right hemisphere damage specifically. Unex­

pectedly, digit span backward showed less effect from focal brain damage
than did digit span forward. In fact, none of the quadrant samples showed
significant differences in performance from controls on the digit span back­
ward measure. Increasing the N by pooling hemispheric samples resulted
in slight, but significant (p < .05), differences between controls and brain­
damaged samples. In general, the data from this study strongly suggest
that performance on digit span forward is far more sensitive to the effects
of focal brain damage than is digit span backward. As previously mentioned,
there was no significant difference in mean digit span backward scores by
right and left hemisphere samples. Accordingly, performance on this task
does not appear to be uniquely a right hemisphere function.
Our investigation of the hypothesis that impaired ability to repeat digits
backward is associated with defective visual spatial functions resulted in the
finding that there was no appreciable difference in performance by subjects
with visual constructional deficits (as measured by Bender Gestalt errors)
or subjects without such deficits. Performance on the Bender showed vir­
tually no correlation with digit backward scores in any sample. Similarly,
performance on another visual constructional measure, the WAIS Block
Designs subtest, correlated significantly with digit span backward only in
the right posterior lesion sample. The correlations between digit span back­
ward and Block Designs scores in the right hemisphere lesion samples (.56
and .61) were both consistently and appreciably higher than the correspond·
ing correlations within the left hemisphere samples (- .01 and .04 ). This
difference would seem to indicate that to some degree, visual constructive
ability (as measured by the Block Designs subtest) is associated with the
ability to repeat digits backward. Examination of individual cases, however,
indicated that those patients with the poorest performance on this visual
constructive task did not necessarily have the lowest scores on the digit span
backward test. Perhaps there is some other neuropsychological mechanism
subserved by the right hemisphere- which effects both block designs and
digit span backward.
Correlational data between WAIS Full Scale I.Q. and digit repetition
suggests that the ability to repeat digits both forward and backward is rather
closely associated with both general intelligence and general memory per­
formance. Defective digit repetition performance may well be a reflection
of reduced intelligence or other general cognitive impairment. This is in
agreement with one of the alternate hypotheses posed by Costa (1975).
In contrast to the findings of Rudel and Denckla (1974), no significant
differences in the incidence of large digit forward-backward discrepancies
were noted between our right' and left hemisphere lesion samples. In the
current sample the low incidence of such discrepancies is accounted for by
the minimal impairment in digit span backward. Since performance on digit
20 F. \V. Black and R. L. Strub

span forward was impaired in many patients, there was a general narrowing
of the digit forward-backward discrepancy.
Our findings may in part be explained by the nature of the particular
sample studied. The patients examined were young, healthy prior to injury,
and without histories of neurological or psychiatric problems. They received
rather discrete brain lesions from small metallic fragments, and in general
suffered limited cognitive and neurological deficits. The cognitive effects seen
by others researching this topic are probably due to the more extensive lesions
and longstanding neurological disease suffered by their patients.

SuMMARY

Digit span forward and backward was investigated in well-matched samples

of patients with discrete quadrant brain lesions. The incidence of significantly
impaired digit repetition performance and the incidence of large forward and
backward digit span discrepancies were also studied. Correlational data of digit
span performance and various intellectual, memory, and constructional measures
was examined.
Approximately 60% of all brain-damaged patients showed an impairment
of digit span forward, while only 5% showed a similar impairment on the digit
span backward task. These data indicate that digit span forward is a more
sensitive measure of brain dysfunction from focal brain lesions. No difference
was found in the performance of patients with right or left hemisphere lesions;
however, the low incidence of aphasia (8%) in this sample may account in part
for the relatively adequate performance by the left hemisphere patients. Although
the current data regarding visual constructive deficits and impaired ability to
repeat digits backward is inconclusive, there did not appear to be a strong rela­
tionship between these two functions for these patients. Digit repetition perform­
ance does appear to be related to both general intellectual ability and to per­
formance on the Wechsler Memory Scale.

REFERENCES

BENSON, D. F., ,and BARTON, M; (1970) Disturbances in constructional ability, "Cortex," 6, 19-46.
CosTA, L. D. (1975) The relation of visuospatial dysfunction to digit span performance on patients
with cerebral lesions, "Cortex," 11, 31-36.
DE RENZI, E., and NICHELLI, P. (1975) Verbal and non-verbal short-term memory impairment
following hemispheric damage, "Cortex," 11, 341-354.
DRACHMAN, D. A., and ABIT, J. (1966) Memory and the hippocampal complex, II, "Arch.
Neurol.," 15, 52-61.
GooDGLASS, H., GLEASON, ]. B., and HYDE, M. R. (1970) Some dimensions of auditory language
comprehension in aphasics, "J. Speech Hear. Res.," 13, 595-606.
HEILBRUN, A. B. (1958) The digit span test and the prediction of cerebral pathology, "Arch.
Neurol. Psychiat.," 80, 228-231. ·
KoPPITZ, E. M. (1964) The Bander Gestalt Test for Young Children, Grune and Stratton, New
York.
McFIE, J. (1969) The diagnostic significance of disorders of higher nervous activity, in Handbook
of Clinical Neurology, Vol. 4, ed. by P. J. Vinken and G. V. Bruyn, North Holland
Publishing Company, Amsterdam.
 Digit repetition performance 21

MILNER, B. (1962) Laterality effects in audition, in Interhemispheric Relations and Cerebral

Dominance, ed. by V. B. Mountcastle, Johns Hopkins Press, Baltimore.
NEWCOMBE, F. (1969) Missile Wounds of the Brain: A Study of Psychological Deficts, Oxford
University Press, London.
RuDEL, R. G., and DENCKLA, M. B. (1974) Relation of forward and backward digit repetition to
neurological impairment in children with learning disabilities, "Neuropsychologia," 12,
109-118.
RuESCH, J., and MooRE, B. E. (1943) Measurement of the intellectual functions in the acute
state of head iniury, "Arch. Neurol. Psychiat.," 50, 165-170.
STERNE, D. M. (1969) The Benton, Porteus and WAIS digit span tests with normals and brain
in;ured sub;ects, "]. Clin. Psycho!.," 25, 173-175.
WEINBERG,]., DILLER, L., GERSTMAN, L., and ScHuLMAN, P. (1972) Digit span in right and left
hemiplegics, "]. Clin. Psycho!.," 28, 361.

F. William Black, Ph.D. & Richard L. Strub, M.D., Department of Neurology, LSU Medical Center, 1542
Tulane Avenue, New Orleans, Louisiana 70112.