AI in Healthcare

Mumbai's Jaslok Hospital Is Using AI and a Walking Video to Predict Parkinson's Disease Before It Cripples — India-France Collaboration, 25 Years of Data

By Mustafa Shariff, Editor in Chief · April 15 , 2026 · 5 min read

Mustafa Shariff, Editor in Chief — Mumbai's Jaslok Hospital Is Using AI and a Walking Video to Predict Parkinson's Disease Before It Cripples — India-France Collaboration, 25 Years of Data

Summary

  • Jaslok Hospital's AI Parkinson's Study: How a Walking Video Could Predict Neurological Disaster Before It Strikes
  • Jaslok Hospital Mumbai has launched a pioneering AI study to predict Freezing of Gait in Parkinson's patients using walking videos. Here is a complete guide to the science, the India-France collaboration, and what it means for AI in Indian healthcare.

There is a moment in advanced Parkinson's disease that patients and their families describe with striking consistency: you are walking, and then suddenly you are not. Your feet feel as if they are glued to the floor. Your body's instruction to move simply does not reach your legs. This condition — Freezing of Gait, or FOG — affects up to 80% of patients with advanced Parkinson's disease. It causes frequent falls, progressive loss of independence, and an enormous caregiver burden. And despite decades of research, there has been no reliable way to predict when it will occur or who is most at risk.

Jaslok Hospital and Research Centre in Mumbai is now attempting to change that — through a clinical study that combines 25 years of longitudinal patient data, artificial intelligence, and a technology so accessible it requires nothing more than a standard video camera.

The Study Design

The clinical study is led by Dr Paresh Doshi, Director of Neurosurgery at Jaslok Hospital, in collaboration with Dr Carine Karachi, one of the world's leading FOG researchers from the Paris Brain Institute, with CEFIPRA funding currently under review. The project has a specific and unprecedented advantage: Jaslok Hospital has longitudinal clinical and video data spanning 25 years for more than 750 patients who underwent Deep Brain Stimulation surgery at the facility. This dataset — tracking the same patients across two and a half decades of neurological progression — is, according to Dr Karachi, unique worldwide.

The AI system is designed around a simple framework: it analyses routine video recordings of a patient walking to identify subtle changes in gait and body coordination that may signal early vulnerability to FOG. Computer vision and machine learning techniques detect changes in gait pattern that are imperceptible to the human eye — changes that precede the actual onset of FOG by weeks or months, creating a window for preventive intervention.

The study is structured in two phases. Phase 1 involves retrospective training of the AI model on data and walking videos from more than 150 patients who subsequently developed FOG. Phase 2 is a prospective validation study involving 337 patients tracked over up to three years — testing whether the trained model can accurately predict FOG onset in patients it has not previously seen.

Why This Matters for Indian Healthcare

Parkinson's disease is India's second most common neurodegenerative disorder after Alzheimer's, and its incidence is rising as India's population ages. The burden of Parkinson's care in India falls heavily on families — specialist neurological care is concentrated in a small number of major urban centres, leaving patients in smaller cities and rural areas without adequate monitoring or treatment adjustment as their disease progresses.

The Jaslok AI system is specifically designed for deployment beyond specialist centres. Because it uses routine video recordings rather than specialised wearable devices or expensive diagnostic infrastructure, it is intended for use across hospitals, outpatient clinics, and telemedicine platforms, including resource-limited settings. The long-term goal is a user-friendly, open-access application that any clinician can use to flag at-risk patients early and predict the timeline of FOG onset — enabling early interventions like structured motor training that may delay or reduce the severity of freezing episodes.

Jitendra Haryan, CEO of Jaslok Hospital, has articulated the hospital's commitment to practical AI for clinical use — the emphasis on 'practical' is critical. Much of India's AI in healthcare conversation has been concentrated in diagnostics imaging and radiology, where the commercial case is clearest. The Jaslok Parkinson's study demonstrates that AI can address conditions that are harder to monetise but carry an enormous burden of disability and caregiver cost.

The India-France Research Partnership

The collaboration between Jaslok Hospital and the Paris Brain Institute is part of a growing pattern of India-France research partnerships in neuroscience and healthcare, supported by CEFIPRA — the Centre Franco-Indien pour la Promotion de la Recherche Avancée — which funds collaborative research between Indian and French scientific institutions. For Indian clinical research institutions, international collaborations of this kind are exactly what the Biopharma SHAKTI mission and the ICMR India-centric research directive are trying to catalyse at scale.

BHC Perspective

At Bengaluru Health Community, the Jaslok Parkinson's study is exactly the kind of clinical AI application that the BHC ecosystem should amplify. It demonstrates that Indian hospitals with longitudinal patient data assets can build AI research programmes that are globally competitive. It shows that AI in Indian healthcare can address neurological and chronic disease burdens — not just radiology and diagnostics. And it provides a model for how Indian clinical research institutions can partner internationally to access funding and expertise that accelerates their work.

Frequently Asked Questions

Q: What is the Jaslok Hospital Parkinson's AI study?
A: Jaslok Hospital Mumbai has launched a clinical study using AI and computer vision to analyse routine walking videos and predict Freezing of Gait — a debilitating symptom of advanced Parkinson's — before it occurs. The study is led by Dr Paresh Doshi in collaboration with Dr Carine Karachi of the Paris Brain Institute.

Q: What makes this study scientifically unique?
A: The study draws on 25 years of longitudinal data and video recordings from over 750 Parkinson's patients who underwent Deep Brain Stimulation at Jaslok Hospital — described by the Paris Brain Institute collaborator as a dataset unique worldwide.

Q: Why is this approach important for India?
A: The system requires only routine video recordings — no specialised wearables or expensive diagnostics — making it deployable across all hospital settings including resource-limited ones. The long-term goal is an open-access clinical app for early FOG prediction across India.