NeuroGait

Landing page
Dashboard
Stride length
Normal stride length and swing
Real time gait analysis and confidence score with abnormal stride lenght and swing
Options to schedule booking with family doctor and download progress report

💡 Inspiration

With a focus on healthcare, we aimed to address one of the most financially burdensome diseases—neurodegenerative disorders like Alzheimer’s and Parkinson's that significantly affect patients’ lives and lack a direct cure. Instead of relying solely on biological treatments, our approach emphasizes data-driven monitoring to help slow disease progression. We developed a method for early detection based on gait analysis, moving beyond traditional memory-based tests. By making diagnosis more affordable and accessible, we hope to enable earlier intervention and reduce long-term impacts on both patients and the healthcare system.

⚙️ What it does

NeuroGait is a data-driven application designed to analyze gait metrics for the early detection of cognitive and motor decline associated with Alzheimer’s and Parkinson’s diseases. The application focuses on identifying anomalies in walking patterns by examining key parameters such as:

Stride length: NeuroGait monitors whether the average stride length falls within the 87–89 cm range, a benchmark established by scientific studies for elderly individuals at risk of these conditions.
Swing time: The application evaluates the gait cycle by measuring the ratio of half the stride length to the swing time, ensuring it aligns with the typical elderly gait speed range of 1.1 to 1.5 meters per second.

🏗️ How we built it

The front-end was developed using React.js with Next.js to manage routing and rendering. We used React Bits to improve UI responsiveness and interactivity, while TypeScript provided static type checking to catch errors early and maintain consistent data structures across the application. For real-time functionality, we integrated the front-end with a Python-based back-end using WebSockets. This enabled continuous data transmission for gait analysis. FastAPI was used to manage asynchronous WebSocket connections, and the computer vision components were powered by OpenCV and MediaPipe to perform real-time gait detection and skeletal tracking.

🔐 Privacy (PIPEDA)

To ensure compliance with PIPEDA (Personal Information Protection and Electronic Documents Act), our system was designed to prioritize user consent, data minimization, and local processing. All video and gait data are analyzed locally on the user's device, meaning no raw footage or personally identifiable information is transmitted or stored without explicit consent. By adhering to PIPEDA's principles of transparency, accountability, and data protection, we ensure that patient data remains private, secure, and fully under their control. This helps to safeguard users dignity and privacy.

🚧 Challenges we ran into

Throughout the development process, we encountered several technical challenges, particularly in implementing the real-time chart display. One of the primary issues involved ensuring continuous data updates on the chart. Initially, the chart failed to render dynamic updates due to redundant logic and poorly defined conditional statements within our update algorithm. Through systematic debugging and code optimization, we corrected these issues and enabled the chart to reflect live data streams accurately.

Additionally, we faced challenges in integrating the back-end data stream with the React-based front-end. The graph was not accurately representing user motion captured by the camera. Upon investigation, we identified that the visualization was based on averaged values rather than raw sensor data, leading to a loss of fidelity and a disconnect between actual motion and displayed results. By modifying the data processing pipeline to transmit raw, time sensitive values, we restored the accuracy of the real time graph and ensured it responded correctly to movement.

🏆 Accomplishments that we're proud of

As part of developing NeuroGait’s real-time gait analysis and neural risk assessment capabilities, we expanded our backend technology stack to include FastAPI and WebSockets for low-latency, bi-directional communication. This allowed for continuous data streaming and real-time plotting between the server and client, which is essential for responsive and interactive gait monitoring. We implemented WebSocket endpoints using FastAPI’s asynchronous capabilities.

🔮 What's next for NeuroGait

To enhance the impact of NeuroGait, we have identified three key areas for improvement.

First, we plan to integrate advanced sensors such as LiDAR to accurately measure distances in real world units. This offers significantly greater precision than traditional webcam based systems that rely on pixel based measurements, enabling more reliable gait analysis and improved clinical insights. This will allow us to get more tangible physical measurements rather than relying on

Second, we aim to personalize the user experience. At present, NeuroGait compares a user’s gait data against generalized statistics for elderly populations. Moving forward, we plan to collect personalized baseline data through a brief onboarding survey, including details like height, mobility history, and previous diagnoses. This will allow us to tailor the analysis to each individual, such as calculating gait ratios based on step length relative to height, resulting in more accurate and meaningful assessments.

Finally, we intend to broaden NeuroGait’s reach through both hardware accessibility and strategic partnerships. Our goal is to release an affordable hardware kit that could be subsidized by healthcare programs, much like publicly funded colon cancer screening kits. Additionally, collaborating with institutions such as nursing homes and elder care centers will help us deliver NeuroGait to a larger population of seniors who could benefit from early detection of neurological conditions.