Augmented Reality (AR) is increasingly explored as a low-burden alternative to pencil-and-paper cognitive tests for dementia and Parkinson’s Disease. Our objective with this review is to synthesize ten years (2014-2024) of empirical evidence on AR-based cognitive screening, estimate pooled diagnostic accuracy, and distil user-experience (UX) guidelines for people with neurodegenerative disorders. We searched Scopus with the string “aug-mented reality” AND cognitive AND (dementia OR Parkinson), screened 399 records, and retained 38 primary studies. Two reviewers independently extracted sample, task, hardware, and accuracy metrics. Optical see-through AR improved test sensitivity over matched non-immersive tests, while projection-based AR offered the largest UX gains. Hardware cost and eye-tracker drift were the main precision bottlenecks. AR can raise both diagnostic sensitivity and patient engagement, but only four studies used clinical-stage participants. Future work should couple low-cost hand-held AR with cloud inference to widen accessibility.
