DigiSpan: Development and Evaluation of a Computer-Based, Adaptive Test of Short-Term Memory and Working Memory
Issued Date
2024-08-05
Resource Type
eISSN
15589102
Scopus ID
2-s2.0-85200712362
Pubmed ID
39052433
Journal Title
Journal of speech, language, and hearing research : JSLHR
Volume
67
Issue
8
Start Page
2729
End Page
2742
Rights Holder(s)
SCOPUS
Bibliographic Citation
Journal of speech, language, and hearing research : JSLHR Vol.67 No.8 (2024) , 2729-2742
Suggested Citation
Dillon H., Boyle C., Gaikwad S., Luengtaweekul P., Cameron S. DigiSpan: Development and Evaluation of a Computer-Based, Adaptive Test of Short-Term Memory and Working Memory. Journal of speech, language, and hearing research : JSLHR Vol.67 No.8 (2024) , 2729-2742. 2742. doi:10.1044/2024_JSLHR-23-00466 Retrieved from: https://repository.li.mahidol.ac.th/handle/20.500.14594/100488
Title
DigiSpan: Development and Evaluation of a Computer-Based, Adaptive Test of Short-Term Memory and Working Memory
Author(s)
Corresponding Author(s)
Other Contributor(s)
Abstract
PURPOSE: This article describes DigiSpan, a new computer-controlled auditory test of forward and reverse digit span, designed to be administered by clinicians, and presents normative and test-retest reliability data for adults. METHOD: DigiSpan mimics conventional live-voice tests in that it commences with trials that ascend in length until a stopping criterion is met, giving rise to a conventional scaled score. It then administers five additional adaptive trials, the length of which depends on the correctness of the response to the previous trial. Each of these two segments of the measurement gives rise to a scaled score. The ascending and adaptive scores are averaged to give an overall score and subtracted to produce an internal measure of consistency, and hence reliability. Young adults with an Mage of 25 years (N = 163) were tested, of whom 65 were retested on a separate day. RESULTS: The scaled scores from the conventional ascending trials were highly consistent with existing normative data based on live-voice tests. Combination of the conventional scaled score with a scaled score based on the adaptive trials led to 44% reduction in error variance for forward memory span and 20% reduction for reverse memory span. The average of these (32%) is similar to but (insignificantly) less than the 42% reduction in error variance that can be predicted based on adding the five adaptive trials. CONCLUSIONS: Replacing live-voice production of digits by a clinician with recorded, computer-controlled production has not affected the difficulty of the test. Adding five additional trials around the sequence length that a test participant can just remember has produced a decrease in measurement error. In addition, the availability of separate scaled scores for the ascending and adaptive phases enables the reliability of the combined score to be checked, for both forward and reverse measurements. The combination of standardized delivery, increased accuracy, internal reliability check, and fast automated scoring makes the test highly suitable for clinical use.