Caffeine and attentional control: improved and impaired performance in healthy older adults and Parkinson’s disease according to task demands

INTRODUCTION: Caffeine is frequently consumed to boost goal-directed attention. These procognitive effects may occur due to the adenosine-mediated enhancement of monoamines, such as dopamine, after caffeine administration. As such, caffeine’s beneficial effects may be altered in conditions such as Parkinson’s disease (PD). However, whether caffeine improves cognition, and at what cost, has not been experimentally established in patients with neurodegenerative disease. METHODS: Single-dose trials to probe cognitive effects of caffeine are often confounded by short-term caffeine abstinence which conflates caffeine’s effects with treatment of withdrawal. Using a placebo controlled, blinded, randomised trial design, we assessed the effect of 100 mg of caffeine across well-established tasks (Choice reaction time, Stroop Task and Rapid Serial Visual Presentation Task; RSVP) that probe different aspects of attention in PD patients (n = 24) and controls (n = 44). Critically, participants withdrew from caffeine for a week prior to testing to eliminate the possibility that withdrawal reversal explained any cognitive benefit. RESULTS: Caffeine administration was found to reduce the overall number of errors in patients and controls on the Stroop (p = .018, η(2)(p) = .086) and Choice reaction time (p < . 0001, η(2)(p) = .588) tasks, but there was no specific effect of caffeine on ignoring irrelevant information in the Stroop task. On the RSVP task, caffeine improved dual item accuracy (p = .037) but impaired single item accuracy (p = .044). Across all tasks, there was little evidence that caffeine has different effects in PD participants and controls. CONCLUSION: When removing withdrawal effects as a factor, we demonstrate caffeine has beneficial effects on selective attention but is a double-edge sword for visual temporal attention and would need careful targeting to be clinically useful. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00213-021-06054-9.