RESUMEN
RATIONALE: Parkinson's disease (PD) impairs working memory (WM)-the ability to maintain items in memory for short periods of time and manipulate them. There is conflicting evidence on the nature of the deficits caused by the disease, and the potential beneficial and detrimental effects of dopaminergic medication on different WM processes. OBJECTIVES: We hypothesised that PD impairs both maintenance and manipulation of items in WM and dopaminergic medications improve this in PD patients but impair it in healthy older adults. METHODS: We tested 68 PD patients ON and OFF their dopaminergic medication, 83 healthy age-matched controls, and 30 healthy older adults after placebo and levodopa administration. We used the digit span, a WM test with three components (forwards, backwards, and sequence recall) that differ in the amount of manipulation required. We analysed the maximum spans and the percentage of lists correctly recalled, which probe capacity of WM and the accuracy of the memory processes within this capacity, respectively. RESULTS: PD patients had lower WM capacity across all three digit span components, but only showed reduced percentage accuracy on the components requiring manipulation (backwards and sequence spans). Dopaminergic medication did not affect performance in PD patients. In healthy older adults, levodopa did not affect capacity, but did impair accuracy on one of the manipulation components (sequence), without affecting the other (backwards). CONCLUSIONS: This suggests that the deficit of maintenance capacity and manipulation accuracy in PD patients is not primarily a dopaminergic one and supports a potential "overdosing" of intact manipulation mechanisms in healthy older adults by levodopa.
