DARWIN — Handwriting & Alzheimer's Dataset

174 participants · 25 handwriting tasks · 18 measured features per task

What is this? DARWIN (Diagnosis AlzheimeR WIth haNdwriting) is a research dataset exploring whether patterns in handwriting can help detect Alzheimer's disease. Participants used a digital pen to complete 25 tasks — copying words, drawing spirals, connecting dots — while sensors recorded timing, pressure, speed, and movement. The dataset compares 89 healthy volunteers (H) with 85 people diagnosed with Alzheimer's (P).
H Healthy P Alzheimer's patient

Time Components — H vs P

Healthy participants spend proportionally more time with the pen in the air between strokes, suggesting more fluid, gesture-rich writing. Alzheimer's patients spend more time with pen on paper, moving more slowly.

Scatter — Air Time vs Paper Time (per person)

Each dot is one participant, averaged across all 25 tasks. The two groups tend to cluster in opposite corners.

Per-Task Average by Group

Each point is the group average for that task. Task numbering follows the original study protocol.

All Features — H vs P (sorted by gap)