Objectives: To study working memory function in untreated major depression
using a digit probe identification and matching task.
Methods: We compared behavioural performance and event-related potentials d
uring processing of the Sternberg working memory task in 14 depressed patie
nts and 13 healthy matched control subjects.
Results: patients made more mistakes than controls as the memory load was i
ncreased from one to 5 digits and had significantly slower reaction times a
t all levels of memory load. The patients' event-related potentials (ERPs)
differed significantly from controls. pathological changes were similar for
auditory and visual presentation. Surface negative activity in the 157-210
ms section of the waveform was reduced for all levels of memory load, sugg
esting abnormal sensory/perceptual processing in the modality-specific asso
ciation cortices, possibly due to a failure of selective attention mechanis
ms. In the 375-840 ms epoch, the patients' responses showed large amplitude
sustained negative activity, maximal at CZ and a reduced late positive wav
e. The large prolonged negativity in the patients' ERPs suggests activation
of additional neuronal assemblies than those normally participating in the
task. This could reflect either compensatory mechanism or dysfunction of i
nhibitory systems. These changes were sensitive to memory load, suggesting
that they reflect alterations of memory-related processes.
Conclusions: This study provides objective evidence that major depression s
ignificantly affects working memory. The ERP changes in depression could he
accounted for by dysfunction of the central executive control of working m
emory. (C) 2000 Elsevier Science Ireland Ltd. All rights reserved.