Recognition memory impairment in Parkinson's disease: Power and meta-analyses.

Abstract

Contrary findings notwithstanding, the prevailing notion is that recognition memory is little affected by Parkinson's disease (PD). Both a power analysis and a meta-analysis were conducted to help clarify the degree of recognition memory deficit associated with PD. The power analysis confirmed that, in general, memory studies of PD participants have been underpowered. This analysis indicated the need to pool study results in a subsequent meta-analysis, the main finding of which was that recognition memory deficits do occur with PD. The largest deficit occurs in PD participants with dementia. Nevertheless, deficits also occur in PD participants without dementia on medication, but nondopaminergic central nervous system abnormalities are more likely to underlie this deficit than PD medication itself. Future development of a theory of cognitive dysfunction in PD should take into account these recognition memory deficits, which may increase with disease progression.