Memory-guided saccades in Parkinson's disease: long delays can improve performance.

A recent study in control subjects suggested the existence of separate pathways for oculomotor spatial working memory tasks depending on whether the delay before movement execution is either short or long (>20 s). The long delay pathway might bypass brain areas commonly affected by Parkinson's disease (PD). This study aimed to assess spatial working memory in Parkinson's disease using short (3 s) and long (30 s) delays in a memory-guided saccade task. Fifteen mild-moderately affected PD subjects off-medication, and 15 age and sex-matched controls were tested (PD mean age 65.3; control 65.9). Subjects were tested in a darkened room using a horizontal LED bar to generate eye movements which were recorded using an infrared limbus tracker. Percentage error in amplitude of the primary saccade was analysed by repeated measures ANOVA. There was a significant interaction between the groups and their response to the short and long delay periods (P<0.02). PD subjects were more strongly impaired in the short delay than the long delay trials when compared with controls. Analysis of the percentage error in amplitude of the final eye position showed the same pattern but only in female subjects. This study provides the first evidence that the proposed parallel spatial memory pathway utilised in longer delay periods is relatively unimpaired in PD. In a broader sense, our results suggest there might be other alternative pathways to overcome deficits in functions impaired by PD.