A multidisciplinary team of researchers has developed a method to monitor the progression of movement disorders using motion capture technology and artificial intelligence.
In two groundbreaking studies, published in the journal Nature Medicine, a multidisciplinary team of AI and clinical researchers shows that by combining human movement data collected from wearable technology with powerful new medical AI technology, they can identify apparent movement patterns, anticipate the evolution of future disease and dramatically increased clinical trial efficiency in two very different rare disorders, Duchenne muscular dystrophy (DMD) and Friedreich’s ataxia (FA).
DMD and FA are rare degenerative genetic diseases that affect movement and eventually lead to paralysis. There are currently no cures for either disease, but the researchers hope that these findings will greatly speed up the search for new treatments.
The progression of FA and DMD is usually tracked through extensive testing in a clinical setting. These sheets provide a significantly more accurate assessment which also increases the accuracy and objectivity of the data collected.
The researchers estimate that using these disease markers means that far fewer patients are needed to develop a new drug than with current methods. This is particularly important for rare diseases where it is difficult to identify suitable patients.
Scientists hope that in addition to using the technology to monitor patients in clinical trials, it could also one day be used to monitor or diagnose a range of common diseases affecting movement behavior such as dementia, stroke, and orthopedic conditions.
Lead author and corresponding author of both papers, Professor Aldo Faisal, from the Departments of Bioengineering and Computing at Imperial College London, is also Director of the UKRI Doctoral Training Center in Artificial Intelligence for Healthcare, Head of the Department of Digital Health at the University of Bayreuth (Germany), and holder of a UKRI Turing AI Fellowship: “Our approach collects massive amounts of data from a person’s whole body movement – more than any neurologist will have the precision or time to monitor in a patient. Our AI technology builds the patient’s digital twin and allows us to make unprecedented, accurate predictions about how an individual patient’s disease will progress. We believe that the same AI technology working in two very different diseases, is showing how effective it can be applied to many diseases and helping us develop treatments for many diseases faster, cheaper and more precisely.”
The two papers highlight the work of a significant collaboration of researchers and expertise across AI technology, engineering, genetics, and clinical disciplines. These include researchers at Imperials Department of Bioengineering and Department of Computing, MRC London Institute of Medical Sciences (MRC LMS), UKRI’s Center for AI Healthcare, UCL Great Ormond Street Institute of Child Health (UCL GOS ICH), and NIHR Great Hospital Biomedical Research Centre. Ormond Street (NIHR GOSH BRC), Imperial College London, Ataxia Center in Queen Square Institute of Neurology, Great Ormond Street Hospital, National Hospital for Neurology and Neurosurgery, National Hospital for Neurology and Neurosurgery (UCLH and UCL/UCL BRC), and University of Bayreuth in Germany and the Gemelli Hospital in Rome, Italy.
Motion Imprints – Trials in Detail
In the study focused on DMD, researchers and clinicians at Imperial College London, Great Ormond Street Hospital and University College London trialed the body-worn sensor prosthesis in 21 children with musculoskeletal dystrophy and 17 healthy age-control subjects. Children wore the sensors while performing standard clinical assessments (such as the 6-minute walk test) as well as going about their daily activities such as eating lunch or playing.
In the FA study, teams at Imperial College London, the Ataxia Center and the Queen Square Institute of Neurology worked with patients to identify key movement patterns and predict genetic markers of the disease. FA is the most common type of hereditary ataxia and is caused by an unusually large triplet duplication of DNA, which turns the FA gene off. With this new AI technique, the team was able to use movement data to accurately predict when the FA gene is “off”, and measure how active it is without having to take any biological samples from patients.
The team was able to administer a rating scale to determine the level of SARA ataxia disability and functional assessments such as gait and hand/arm movements (SCAFI) in 9 FA patients and their matched controls. The results of these validated clinical evaluations were then compared with those obtained from the use of the new technology on the same patients and controls. The latter shows greater sensitivity in predicting disease progression.
In both studies, all data from the sensors was collected and fed into AI technology to create individual avatars and analyze movements. This massive dataset and powerful computing tool allowed the researchers to identify key movement signatures seen in children with DMD as well as adults with a muscle tear, which were different in the control group. Many AI-based movement patterns have not been previously described clinically in DMD or FA.
The scientists also discovered that the new AI technology can also significantly improve predictions of how an individual disease will progress for patients over a six-month period compared to current gold-standard assessments. Such accurate prediction allows clinical trials to run more efficiently so patients can access new treatments faster, as well as help with drug dosing more accurately.
Lower numbers for future clinical trials
This new method of analyzing whole-body movement measurements provides clinical teams with clear signs of disease and predictors of progression. These are invaluable tools during clinical trials to measure the benefits of new treatments.
The new technology could help researchers conduct clinical trials of conditions that affect movement faster and more accurately. In the DMD study, researchers show that this new technology can reduce the number of children needed to discover if a new treatment will work to a quarter of those needed with current methods.
Similarly, in the FA study, the researchers showed that they could achieve the same accuracy with 10 patients instead of more than 160. This AI technique is especially powerful when studying rare diseases, when there are fewer patients. In addition, this technique allows patients to be studied across life-changing pathological events such as loss of ambulation while current clinical trials target ambulatory or non-ambulatory patient populations.
Professor Thomas Voight, Director of the Biomedical Research Center at the National Institute of Human Rights (NIHR GOSH BRC) and Professor of Developmental Neurosciences at UCL GOS ICH, said, “These studies show how innovative technology can dramatically improve the way we study diseases day in and day out. The impact of this, combined with expert clinical knowledge, will not only improve the efficiency of clinical trials but has the potential to translate across a wide range of conditions affecting movement.This is thanks to collaborations across research institutes, hospitals, clinical specialties and with patients and families. Dedicated people can begin to solve the difficult problems facing rare disease research.”
Joint first author on both studies, Dr. Balasundram Kadervelo, Postdoctoral Researcher in the Departments of Computing and Bioengineering at Imperial College London, said, “We were surprised to see how our AI algorithm was able to discover some new ways to analyze human movements. We call them behavioral fingerprints because Fingerprints on your hand allow us to identify a person, these digital fingerprints accurately characterize the disease, regardless of whether the patient is in a wheelchair or walking, in the clinic doing an evaluation or having lunch in a cafe.”
Co-first author on the DMD study and co-author on the FA study, Dr Valeria Ricotti, Honorary Clinical Lecturer at UCL GOS ICH, said: “Searching for rare conditions can be more expensive and logistically difficult, meaning patients miss out on potential new treatments. Increasing the efficiency of clinical trials gives us hope that we will be able to successfully test many more treatments.”
Co-author Professor Paola Giunti, Head of UCL Ataxia Centre, Queen Square Institute of Neurology, and Honorary Consultant at the National Hospital for Neurology and Neurosurgery, UCLH, said: “We are delighted with the results of this project which have shown how close artificial intelligence is certainly superior in capturing disease progression. In a rare disease such as Friedreich’s ataxia. With this new approach, we can revolutionize the design of clinical trials for new drugs and monitor the effects of already existing drugs with a precision not known with previous methods.”
“The large number of clinically and genetically well-characterized FA patients at the Ataxia Center UCL Queen Square Institute of Neurology as well as our critical input into the clinical protocol made the project possible. We are also grateful to all of our patients who participated in this project.”
Professor Richard Feistenstein, co-author of both studies, from the MRC London Institute of Medical Sciences and the Department of Brain Sciences at Imperial College London, said: “Patients and families often want to know how their disease is progressing, and motion capture technology combined with artificial intelligence can help.” In providing this information. We hope that this research will have the potential to transform clinical trials in rare movement disorders, as well as improve diagnosis and monitoring for patients above levels of human functioning.”
The research was funded by a UKRI Turing AI Fellowship by Professor Faisal, NIHR Imperial College Biomedical Research Center (BRC), MRC London Institute of Medical Sciences, the Duchenne Research Fund, NIHR Great Ormond Street Hospital (GOSH) BRC, the UCL/UCLH BRC UK Medical Research Council.
Caderfelo, b. et al. (2023) A wearable motion capture prosthesis and machine learning predict disease progression in Friedreich’s ataxia. Nature medicine. doi.org/10.1038/s41591-022-02159-6.