Umanit
Daily-life gait impairment detection.
Advanced statistical analyses at the service of gait-affected individuals.
COMPANY
UmaniT
RESEARCH CENTRE
LMJL, Laboratoirede Mathématiques Jean Leray, Univ. Nantes & CHU de Nantes France
PRODUCTIVE SECTOR
Biomedecine and Health Care
Problem description
Gait impairment comes with a major loss of autonomy, ultimately affecting the quality of life. Can we detect gait changes before symptoms appears so as to propose therapies to avoid or delay their occurrence?
Challenges and goals
- Measuring gait with a lightweight device tha individuals forget they are wearing to limit biases.
- Designing a biomarker called individual Gait Pattern, sensitive to changes in gait.
Mathematical and computational methods
The main idea is to provide a so-called individual Gait Pattern (IGP) which describes the average rotation of the hip over time during a typical gait cycle. This is achieved using a tiny motion sensor, clipped on the right side of the belt, that records data to a secure server through a dedicated smartphone. The data comes in the form of a time series of unit quaternions. These are complex mathematical objects that motivated the development of specific statistical methods such as finding clusters of patients with similar gait impairment of designing risk indicators.
Results and Benefits
We analysed and clustered the IGP of 27 patients diagnosed with multiple sclerosis, Cluster 1 regroups patients with no clinical evidence of gait impairment only, while Cluster 2 regroups patients with mild walking deficiency. This suggests that the IGP detects clinically-invisible gait differences.
