The Paced Auditory Serial Addition Test (PASAT)
is a commonly used procedure that combines elements of
both a working memory task and a test of information processing
speed. Patients with multiple sclerosis (MS) have consistently
been found to be impaired on this test and it has been
recommended as a core outcome measure in clinical trials.
The standard score for this task is the number of correct
responses at each stimulus presentation rate but a concern
has been raised that subjects may ignore some test items
in order to chunk the information into manageable portions
and avoid performing several cognitive tasks simultaneously.
To account for this strategy, one can examine the proportion
of correct responses that are consecutive (termed dyads),
since such responses require that the task be performed
according to the instructions. We compared a group of 35
mildly to moderately disabled MS patients and matched healthy
controls on the PASAT. The MS patients made significantly
fewer correct responses at the 2 slowest presentation rates
(2.4, 2.0 s/digit) while their scores at faster rates (1.6,
1.2 s/digit) did not discriminate them from controls as
well. Nevertheless, the MS patients' percentage of
dyads was significantly lower than that of the control
sample across all stimulus presentation rates. While our
study supports the use of the PASAT as a test that distinguishes
MS patients from healthy individuals, our results also
illustrate problems that lie in the interpretation of this
difference in performance. It appears that a chunking strategy
may be common in the PASAT, particularly as task demands
increase, and that this may mask actual performance differences.
If so, the total correct response score alone is limited
as a measure of working memory and information processing
speed. More detailed analyses of PASAT performance, coupled
with other measures of information processing, may help
clarify the underlying cognitive deficits of MS patients.
(JINS, 2001, 7, 363–372.)