Early Parkinson’s Insights from Hand-Drawn Patterns
NeuroSketch is a digital diagnostic tool that detects early signs of Parkinson’s through simple spiral and sine wave drawings. We worked with the client to design and engineer the solution end-to-end, from an accessible drawing interface for older adults to an AI/ML pipeline that transforms sketches into reliable neuromotor insights. The result is a non-invasive, scalable, and clinician-ready screening tool.
70%
faster screenings compared to conventional neurology visits
85%
model accuracy in detecting early Parkinsonian signs
3X
higher accessibility with at-home, self-administered tests
How We Engineered a Digital Pathway for Preventive Neurology
The client, a digital health startup, faced a critical challenge: early detection of Parkinson’s is vital, but current diagnostic tools are invasive, costly, and inaccessible for regular screening. We partnered with them to design and engineer NeuroSketch, a domain-informed, AI-powered diagnostic platform. By combining expertise in AI/ML, mobile UX for older adults, and secure health data management, we built a tool that empowers both patients and clinicians with real-time, actionable insights.
Use Cases
Sketch Capture Guidance
At-home spiral/sine tests on mobile or tablet with stroke smoothing for reliability.
HOG-based signal extraction
Measures jitter, angle, amplitude to surface tremor patterns.
Dual-pattern protocol
Standardizes tasks for consistent results across sessions/devices.
Instant, explainable scoring
RF/XGBoost probabilities with heatmaps & velocity maps.
Clinician access & sharing
Secure portal, session tracking, consented result exchange.
Longitudinal tracking
History, trendlines, stability index for remote monitoring.
Tech Stack
- React Native
- React
- react-native-canvas
- Go API Gateway
- Flask microservices
- JWT + Firebase Auth
- OpenCV (HOG)
- scikit-learn (RF, XGBoost)
- AWS ECS
- Docker
- S3
- Helm
- GitHub Actions
- HIPAA-compliant data handling
Explore More Case Studies.
NutriBot
Developed an AI-powered nutrition assistant that personalized meal planning and automated patient engagement workflows.
PrideLedger
Built a blockchain-based platform that secured financial transactions, automated reconciliations, and improved auditability.
Get in touch to discuss your software vision with industry experts
Prefer to write directly? info@pinnacloid.com or +1 (312) 780-8825
What stood out the most was how easy it was to communicate with their team. We always knew where things stood, and there were no surprises.
CEO, DigitArtisan