Uncovering Gait Clusters in Incomplete SCI via Unsupervised Algorithms
Although gait function is a key factor for community re-integration of ambulatory individuals with a spinal cord injury, prediction of the recoverable gait function after rehabilitation remains elusive.
Our research combines 3-dimensional gait analysis data with an extensive feature set on SCI attributes to explore and characterize clusters in the heterogeneous SCI population. A clear prediction of the recovery potential can improve personalized rehabilitation and manage the patient’s expectations from an early stage.
Data collected at key stages during rehabilitation and follow-up visits allows us to paint a comprehensive picture of gait evolution in the early stages of an SCI injury.
This project will retrospectively analyze 3-dimensional gait analysis data collected at the Swiss Paraplegic Centre between 2010 and 2021. We will categorize the gait patterns observed in our dataset, focusing on identifying distinct groups based on spatiotemporal, kinematic, and kinetic parameters. This categorization will help us explore potential relationships between specific gait patterns and SCI attributes. The extensive list of SCI attributes we consider is informed from literature and refined through expert discussion. It consists of lesion level and completeness, spasticity, depth sensitivity characterized by proprioception and vibration, mobility characterized by SCIM (Spinal Cord Independence Measure) scores, and the muscle and joint state of the lower extremity.
Our analysis will also include a longitudinal component, examining changes in gait patterns over time. We will analyze gait data collected at three key stages: during early rehabilitation, towards the end of rehabilitation, and during follow-up visits after patients have returned home. Here, we will track the evolution of gait characteristics in ambulatory persons with SCI.
We expect to observe significant differences in gait patterns depending on SCI characteristics and to identify changes in these patterns over time. An early prediction of the recoverable gait function through rehabilitation will facilitate personalized rehabilitation strategies and treatment decisions.