Case Study: Parkinson's Voice Initiative achieves scalable, accurate Parkinson’s detection from phone calls with Twilio

Parkinson's Voice Initiative Detects Parkinson's One Call at a Time

Parkinson’s Voice Initiative, led by mathematician Max Little, developed a noninvasive way to detect Parkinson’s symptoms using voice recordings to address the high cost, slow pace, and limited access of traditional diagnostic tests. The challenge was to create a low‑barrier, scalable method that could reliably capture subtle vocal tremors and fluctuations associated with the disease from a wide and often elderly population.

The team used Twilio to collect 20–30 second vowel recordings via local phone numbers, routing audio into a PHP platform where proprietary algorithms analyze and compare samples against healthy and affected cohorts; data is augmented by self‑reports from PatientsLikeMe. The system has delivered a 100% capture rate, gathered over 600 calls in its first month, supports multiple languages/accents, and is being prepared for open scientific sharing—offering clinicians a faster, cheaper way to run trials and monitor symptom severity.

Parkinson's Voice Initiative

Max Little

Project Director