In recent months, much has been made about the sheer power of the artificial intelligence (AI) assistant ChapGPT and what it means for the future of work. Its ability to mimic human thinking in accomplishing a staggering amount of tasks has sent an unmistakable signal across the globe: AI has arrived, and it is here to stay. 

Talks about AI often prompt serious philosophical reflections: what is sentience; what is free will? In medicine, AI triggers additional questions on how well AI can be integrated into clinical practice without disrupting the physician-patient relationship—a decidedly human interaction centered on trust. 

Read more about DMD treatment

Continue Reading

Meanwhile, the global population continues to rise, healthcare needs are increasing, and diseases of old age are becoming more prominent as people live longer. There is no way that the infrastructures of medicine today can meet the demands of tomorrow. We desperately need a means to make quick, informed decisions on diagnosis and treatment. Could AI be the answer? 

Ricotti and colleagues think so. Writing in Nature Medicine, they express regret that medicine is so slow to incorporate AI into its machinery: “Artificial intelligence has the potential to revolutionize healthcare, yet clinical trials in neurological diseases continue to rely on subjective, semiquantitative and motivation-dependent endpoints for drug development.” 

The world of neurological clinical research is very much dependent on observation-based outcomes; for example, in Parkinson’s disease, endpoints such as gait, balance, and coordination are all subjectively measured. This is a sticking point, because human movement is complex and highly variable; observer bias is impossible to eliminate. 

Ricotti and colleagues advocate for what they term “digital biomarkers.” Digital biomarkers are quantifiable data obtained via the use of digital devices; they have the potential to make objective what has only been measured subjectively in the past. However, to make full use of digital biomarkers, scientists need to incorporate them into existing AI processes, allowing patterns and trends to be more quickly discerned. 

Wearable Tech and AI 

Duchenne muscular dystrophy (DMD) is characterized by the progressive loss of muscle function. Because this loss happens gradually over time, physicians are unable to track minute function deterioration. Instead, physicians are only able to note major changes in a patient’s muscle function, such as when a patient loses the ability to walk independently. 

Is there a better way to track disease progression in DMD? Ricotti and colleagues proposed the use of wearable sensor technology and AI machine learning methods to “identify barely perceptible complex patterns in patient movement behavior.” This means that common muscle function assessments, such as the 6-minute walk test, can be done away with, given that they have little relevance in real life (as patients do not only walk in 6-minute spurts). 

Read more about DMD etiology 

They developed a system that can seamlessly capture motor performance using data-driven, whole-body movement behavior. This resulted in a 17-sensor bodysuit which can be used to measure small changes in muscle function. The research team recruited 21 male individuals with DMD and conducted follow-up meetings at baseline, 6 months, and 12 months. During these visits, patients were measured using traditional means of assessing muscle function, as well as the newly constructed bodysuit. 

The research team defined new movement behavioral fingerprints that were useful in distinguishing patients with DMD from controls. They then fed that information into machine learning algorithms that can predict cross-sectional and longitudinal disease course with the goal of outperforming predictions derived from current clinical standards. 

Using Bayesian optimization, the research team was able to construct a behavioral biomarker, derived from daily-life movement behavior, that reliably measures disease progression. This allows clinicians to monitor a patient’s disease state in ways beyond the reach of the 5 senses, which can be useful in a clinical setting and also in assessing the merits of proposed new therapies. 

Into the Future 

As humanity wrestles with the role of AI in the world, it is clear, at least in the medical space, that AI can enhance the work of clinicians by helping to speed up processes that manually take up much time and interpret data with greater swiftness and objectivity. There is potential for AI to be integrated into clinical processes from disease surveillance to treatment to medical education—the possibilities are endless. 

“There have been two industrial revolutions in human society, steam revolution and electrical revolution, which both profoundly changed the way of human life and promoted human civilization,” Liu and colleagues wrote in Current Medical Science. “Now the scientific and technical revolution, including AI technology, has already shown an irresistible trend that has grown vigorously.” 

It is only a matter of time before medical students learn about AI and how it has the potential to elevate clinical practice. A world with AI-powered medicine can open up global access to healthcare, increase our understanding of disease processes, and streamline diagnostic and treatment protocols. 


Liu PR, Lu L, Zhang JY, Huo TT, Liu SX, Ye ZW. Application of artificial intelligence in medicine: an overviewCurr Med Sci. 2021;41(6):1105-1115. Published online December 6, 2021. doi:10.1007/s11596-021-2474-3

Ricotti V, Kadirvelu B, Selby V, et al. Wearable full-body motion tracking of activities of daily living predicts disease trajectory in Duchenne muscular dystrophyNat Med. 2023;29(1):95-103. Published online January 19, 2023. doi:10.1038/s41591-022-02045-1