Saltar al contenido
Dispositivos y mediciónRevisión de evidencia3 min read

AI plus wearable motion sensors in medicine: small datasets, mostly neurology, mostly clinics

Colectivo editorial MoveVitals · Study published · Reading updated

Original paper:The Use of Artificial Intelligence and Wearable Inertial Measurement Units in Medicine: Systematic Review. · JMIR mHealth and uHealth · 2025 · Artículo original ↗

Idea principal

In this systematic review, studies pairing AI with wearable inertial measurement units (IMUs) had a median dataset of only 50 participants; 76% used traditional machine learning, 93% used supervised learning and 65% focused on neurological problems. The abstract does not give the number of studies reviewed or how well the models performed, so it describes research habits, not how well the tools work.

Study design

Revisión de evidencia

Reading basis

Published abstract

Reading focus

Design & applicability

Esta explicación editorial está disponible en inglés, chino simplificado y chino tradicional. Esta interfaz muestra la lectura en inglés.

Pregunta de investigación

Wearable motion sensors (accelerometers and gyroscopes, known as IMUs) combined with AI are expected to change patient care, but the research base is not well mapped. The authors set out to chart which medical fields use these tools, which models are applied, and what the datasets look like.

Población y método

Systematic review following PRISMA reporting guidelines, built around three questions: which medical fields research AI with IMU data, which models analyse the data, and what the datasets are like. The abstract does not report the databases searched, the number of included studies, or the dates covered.

Resultados

The authors’ reported results. Read each estimate with its population and uncertainty.

  1. 1
    The median dataset size was 50 participants, which the authors see as a serious limit for AI models that need large datasets to train and generalize.
  2. 2
    Machine learning dominated in 76% of the surveyed studies, mainly traditional models such as linear regression, support vector machine and random forest, with room for more deep learning.
  3. 3
    Supervised learning was used in 93% of studies, so unsupervised methods for finding hidden patterns are little used.
  4. 4
    Most studies took place in clinical settings (77%) rather than in real life, and 65% focused on neurological problems, leaving areas such as musculoskeletal care less explored.

Límites e incertidumbre

  1. 1
    The abstract reports no number of included studies, search dates or quality assessment, so the strength of the evidence base is unclear.
  2. 2
    The review maps what researchers have done; it does not test whether any AI-plus-IMU tool improves patient outcomes.
  3. 3
    The authors themselves name small datasets and the clinic-only setting as limits on practical applicability. This reading is based on the abstract only.

Relevancia práctica

For readers, the main message is that most AI work on wearable motion data is still early: small samples, clinic-based and focused on neurological conditions. Claims that a consumer wearable can diagnose conditions from movement data should be treated with caution until tested in larger, real-life groups.

Tus datos

Phone and watch accelerometer data, such as step counts and activity minutes, are the everyday cousin of the IMU signals studied here. Use them as personal trends on one device, not as a diagnostic of movement disorders.

Cómo interpretar la evidencia

Check which designs were included, how searches and bias assessment were conducted, and the last search date. Publication date alone does not make the underlying evidence current.

Resumen estructurado de los autores English

El resumen original en inglés se reproduce con atribución bajo la licencia indicada. Es una fuente, no una interpretación de MoveVitals. El análisis preparado figura arriba; verifica los resultados importantes en la fuente.

Background

Artificial intelligence (AI) has already revolutionized the analysis of image, text, and tabular data, bringing significant advances across many medical sectors. Now, by combining with wearable inertial measurement units (IMUs), AI could transform health care again by opening new opportunities in patient care and medical research.

Objective

This systematic review aims to evaluate the integration of AI models with wearable IMUs in health care, identifying current applications, challenges, and future opportunities. The focus will be on the types of models used, the characteristics of the datasets, and the potential for expanding and enhancing the use of this technology to improve patient care and advance medical research.

Methods

This study examines this synergy of AI models and IMU data by using a systematic methodology, following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, to explore 3 core questions: (1) Which medical fields are most actively researching AI and IMU data? (2) Which models are being used in the analysis of IMU data within these medical fields? (3) What are the characteristics of the datasets used for in this fields?

Results

The median dataset size is of 50 participants, which poses significant limitations for AI models given their dependency on large datasets for effective training and generalization. Furthermore, our analysis reveals the current dominance of machine learning models in 76% on the surveyed studies, suggesting a preference for traditional models like linear regression, support vector machine, and random forest, but also indicating significant growth potential for deep learning models in this area. Impressively, 93% of the studies used supervised learning, revealing an underuse of unsupervised learning, and indicating an important area for future exploration on discovering hidden patterns and insights without predefined labels or outcomes. In addition, there was a preference for conducting studies in clinical settings (77%), rather than in real-life scenarios, a choice that, along with the underapplication of the full potential of wearable IMUs, is recognized as a limitation in terms of practical applicability. Furthermore, the focus of 65% of the studies on neurological issues suggests an opportunity to broaden research scope to other clinical areas such as musculoskeletal applications, where AI could have significant impacts.

Conclusions

In conclusion, the review calls for a collaborative effort to address the highlighted challenges, including improvements in data collection, increasing dataset sizes, a move that inherently pushes the field toward the adoption of more complex deep learning models, and the expansion of the application of AI models on IMU data methodologies across various medical fields. This approach aims to enhance the reliability, generalizability, and clinical applicability of research findings, ultimately improving patient outcomes and advancing medical research.

Fuentes y atribución

The Use of Artificial Intelligence and Wearable Inertial Measurement Units in Medicine: Systematic Review.

Smits Serena R, Hinterwimmer F, Burgkart R, von Eisenhart-Rothe R, Rueckert D.

JMIR mHealth and uHealth · 2025-01-29

DOI: 10.2196/60521 ↗PubMed · 39880389 ↗Europe PMC · Licencia registrada: cc by ↗

Abstract © the original authors, reproduced under the recorded open licence. Translations support reading; use the English record to check terminology and numerical results. Our methodological interpretation is separate from the authors’ conclusions.

Interpretación editorial de MoveVitals · contenido educativo · sin revisión clínica individual registrada

Fuentes verificadas 2026-11-01

Guías prácticas

Leer en contexto

·Estudio de cohorte

Activity–stroke and heart attack curves differ by accelerometer processing method

Among 90,237 UK Biobank adults free of cardiovascular disease at baseline, followed until December 2022 for 1,298 strokes and 2,031 myocardial infarctions, the shape of the link between moderate-vigorous physical activity (MVPA) and each outcome depended on how the accelerometer data were processed. For stroke, hazard fell in a straight line only with the machine-learning metric; for myocardial infarction, two metrics gave a curved decline and activity counts a straight one. The abstract gives no hazard ratios, so it shows that methods disagree, not which one is right, and the findings are observational associations.

Preventive Medicine

Explicaciones detalladasLeer estudio →
·Estudio/revisión de validación

Smartwatch data as forensic evidence: heart rate is best validated, the rest uneven

A systematic review and meta-analysis of studies on smartwatches from four major brands, framed around digital forensics, found that heart rate was both the most studied and the most accurate measure. Step count, distance, energy expenditure, activity classification and sleep were less consistently validated, and results varied by brand, protocol and reporting. The abstract gives no pooled numbers, and most studies were run in healthy people in laboratories, so it cannot say how a particular watch performs for a particular person in daily life.

Forensic Science International

Explicaciones detalladasLeer estudio →
·Revisión de revisiones

Activity trackers: about 1,800 more steps a day, but small effects on other outcomes

An umbrella review of 39 systematic reviews and meta-analyses, covering 163,992 participants, found that activity trackers were associated with more physical activity (SMD 0.3–0.6), equating to about 1,800 extra steps and 40 minutes more walking per day, and with a bodyweight reduction of about 1 kg. Effects on blood pressure, cholesterol, glycosylated haemoglobin, quality of life and pain were typically small and often non-significant. The abstract does not say what else came with the trackers in these programs or how much the reviews overlap, so the gains cannot be credited to the device alone.

The Lancet Digital Health

Explicaciones detalladasLeer estudio →
·Estudio/revisión de validación

Seven sleep trackers: good at spotting sleep, weak on wake and sleep stages

In a sleep laboratory, 34 healthy young adults wore a subset of seven consumer sleep trackers for three nights. Against polysomnography every device detected sleep very well (sensitivity 0.93 or higher) but caught wake poorly (specificity 0.18–0.54), and five of the seven matched or beat research actigraphy (specificity 0.39). Sleep-stage estimates were inconsistent, and the participants were healthy adults aged 18 to 35 in a fixed 8-hour bed window, with each device tested in only 11 to 20 of them, so the results say little about people with sleep problems.

Sleep

Explicaciones detalladasLeer estudio →
Dispositivos y medición →