Building & SalesOct 2026–nowProjectBuilding now
Digital biomarkers
Phone-based touch study
- Tasks
- 5
- Per session
- 6 min
Context
A small study to build a framework for using digital biomarkers (DBMs) as predictors of neurodegenerative diseases (NDDs). Subtle changes in how someone types, taps or traces a line can track these conditions, and every phone can measure them. The first step is knowing what normal looks like.
My role
Built it solo.
What I did
- 01Built five touch tasks that run in a phone browser: a typing sprint, two-finger tapping, spiral tracing, reaction targets and a trail-making test.
- 02Extracted measures from each: keystroke hold and flight times, tapping rhythm and fatigue, tremor frequency, reaction time and left/right asymmetry.
- 03Collected data anonymously: participant codes, no names or emails, and a UK-hosted database the public page can only write to.
Outcome
Collecting a baseline from adults without neurological conditions.
Try it
The five tasks, running live. This is the practice version: nothing is sent anywhere, and results stay on your device.
Start the tasks ↗Best on a phone. Scan to open it there.
What it stores
| Participant | Age band | Hand | Key hold (ms) | Key gap (ms) | Taps / 10 s | Tap rhythm (CV) | Tremor (Hz) | Reaction (ms) | Trails B (s) |
|---|---|---|---|---|---|---|---|---|---|
| 8RA-GZD | 25–29 | Right | 387 | 408 | 42 | 0.09 | 7.0 | 351 | 18.5 |
| LMU-SZC | 25–29 | Right | 342 | 297 | 42 | 0.17 | 5.3 | 206 | 32.7 |
| SPR-WLG | 45–49 | Right | 261 | 355 | 44 | 0.16 | 6.6 | 366 | 44.8 |
| RF4-LD7 | 35–39 | Right | 334 | 243 | 64 | 0.12 | 5.0 | 373 | 43.9 |
| 8MM-SK7 | 25–29 | Left | 241 | 348 | 60 | 0.09 | 8.4 | 273 | 14.9 |
| XMT-DR9 | 30–34 | Right | 158 | 376 | 39 | 0.20 | 8.9 | 250 | 14.5 |
| PER-NDW | 25–29 | Right | 136 | 236 | 33 | 0.14 | 4.7 | 291 | 26.5 |
| D9V-GL5 | 35–39 | Right | 241 | 414 | 60 | 0.20 | 6.4 | 360 | 37.0 |
Tools
- JavaScript
- Signal processing
- Supabase
- Postgres