Time-Series & Sensors

Forecasting Loss of Ambulation in Friedreich Ataxia

Predicting Loss of Ambulation in Friedreich’s Ataxia Using IMU-Based Accelerated Failure Time Models

Forecasting when someone with Friedreich ataxia may lose the ability to walk, so care can be planned ahead.

The problem

Timely intervention needs an estimate of when ambulatory decline is likely, not just a current severity score.

The approach

Accelerated failure-time survival models with a log-logistic distribution, combining cup-sensor signals with genetic and clinical covariates.

The outcome

Predicted survival probabilities tracked observed disease stages closely.

In detail

Built a disease-progression forecasting approach that combines movement data from the AIM-C instrumented cup with genetic and clinical variables. Accelerated Failure Time modelling is used to estimate the likelihood and timing of transitions between ambulatory states, with the aim of supporting earlier intervention and more personalised monitoring of Friedreich ataxia progression.

Built with

• Survival Analysis
• Wearable Sensors
• Clinical Forecasting

More in Time-Series & Sensors

• Probabilistic Deep Ensembles for FRDA Severity Assessment
• Cross-Device Domain Adaptation for FRDA Assessment
• AIM-M: Multiview IMU Assessment of Friedreich Ataxia

Next project: Reliable FRDA Assessment with Isolation Forest Anomaly Detection

By Kanishka Ranaweera, PhD, Machine Learning Engineer in Melbourne, Victoria, Australia. hello@kanishkabuilt.me