Researchers have developed an artificial intelligence-defined digital biomarker that might outperform clinical scales in a Friedreich ataxia (FA) clinical trial setting.

“This work promises to shorten clinical trials that would otherwise be prohibitively long or reduce the number of patients required to measure the deterioration of the neurological state,” the researchers wrote in Nature Medicine.

The data-driven approach allowed for longitudinal predictions of personal Scale for the Assessment and Rating of Ataxia (SARA) and Spinocerebellar Ataxia Functional Index (SCAFI) scores 9 months into the future. The predictions were 1.7 and 4 times more precise than those obtained using only SARA and SCAFI scores, respectively.

Moreover, the data-driven technique, but not SARA or SCAFI, accurately predicted the expression level of the FXN gene, which encodes for frataxin, in all FA patients included in the study.

Read more about FA clinical trials

The researchers used a full-body motion capture approach to analyze 2 subassessments of the SCAFI scale: the 8-meter walk (8-MW) test, focusing on the behavioral changes of the full-body kinematics, and the 9-hole peg test (9-HPT), focusing on the upper body kinematics. Participants used a motion capture suit to have their entire body movement recorded. Then, the researchers used a machine learning approach to generate a series of markers of participants’ performance.

“The fact that the kinematic features of each of the subtasks (8-MW and 9-HPT) can predict the overall clinical scores with good accuracy suggests that our data-driven approach can predict the overall performance of a patient with a minimal number of tasks in the clinical assessment although they will have to have the suit fitted, which takes about 10 [minutes],” the researchers explained.

Such data-derived wearable biomarkers may offer additional advantages over clinical scales. For instance, measurements are driven by data-rich, digitally accurate data rather than subjective, by-eye assessments. Furthermore, they provide a more realistic picture of the observations made by clinicians in their verbal characterization of FA patients.

The study included 9 patients diagnosed with FA with genetically confirmed GAA-repeat expansion on both alleles of the FXN gene and 9 age- and sex-matched controls. The researchers collected motion capture suit data 4 times from the FA patients and once from the controls during the study.


Kadirvelu B, Gavriel C, Nageshwaran S, et al. A wearable motion capture suit and machine learning predict disease progression in Friedreich’s ataxia. Nat Med. 2023;29(1):86-94. doi:10.1038/s41591-022-02159-6