Artificial Intelligence (AI) refers to a broad field of computer science focused on creating systems capable of performing tasks that typically require human intelligence, such as decision-making, problem-solving, and pattern recognition. Machine Learning (ML) is a subset of AI that focuses specifically on developing algorithms that enable computers to learn from data and improve their performance over time without explicit programming. This distinction is crucial in understanding how AI technologies, including Machine Learning, are transforming various fields, particularly in healthcare and rehabilitation.
Artificial Intelligence (AI) including Machine learning has rapidly become integral to healthcare, with rehabilitation emerging as a field that stands to benefit significantly from its data-driven capabilities. By utilizing algorithms capable of analyzing extensive datasets, AI enables clinicians to identify patterns in patient progress, predict treatment outcomes, and personalize interventions to suit individual needs. This approach aligns well with the shift toward patient-centered care in modern healthcare, where AI acts as a key tool in creating precise, adaptive, and responsive treatment plans.1
Research consistently demonstrates that AI’s ability to process vast amounts of data enhances the accuracy and personalization of rehabilitation.2 For instance, literature on AI in physical therapy and rehabilitation report improvements in patient adherence to treatment and recovery timelines due to tailored, data-informed intervention plans.3
AI’s role in real-time monitoring is also highlighted in recent studies, which indicate that wearable devices and sensor technologies allow for continuous tracking of patient activity, providing rehabilitation professionals with insights that help optimize therapy plans dynamically.4 A systematic review on wearable AI devices in healthcare shows the effectiveness of these tools in capturing essential patient metrics, allowing adjustments to be made in response to real-time data.5
Furthermore, predictive analytics—another key application of AI in rehabilitation—enables clinicians to anticipate potential treatment outcomes and adjust interventions preemptively, reducing the likelihood of setbacks. Systematic reviews emphasize that AI’s predictive capabilities are especially beneficial in managing chronic conditions and complex rehabilitation cases, as it empowers therapists to refine approaches based on individual risk factors and likely responses.6
AI are also instrumental in addressing gaps in rehabilitation services, such as limited access in rural or underserved areas. Tele-rehabilitation platforms powered by AI can facilitate remote monitoring and continuous care, making rehabilitation more accessible.7
Moreover, AI applications in rehabilitation extend beyond physical recovery; cognitive and emotional rehabilitation can be enhanced through tools such as virtual reality (VR) and AI-driven cognitive training systems.8 The intersection of AI, rehabilitation, and patient-centered care is poised to revolutionize the field, offering a more dynamic, efficient, and patient-tailored approach to rehabilitation.
In this review, we will explore the core applications of AI in rehabilitation, highlighting their potential benefits, challenges, and future directions. This exploration aims to provide a comprehensive understanding of how AI is reshaping rehabilitation practices and to encourage further research into the transformative possibilities these technologies offer.
Methodology
Narrative Review Approach
This manuscript is a narrative review that aims to synthesize existing evidence on the role of Artificial Intelligence (AI) in rehabilitation and its implications for patient care. A narrative review approach was selected to provide a comprehensive and descriptive overview of current knowledge in this evolving field.
Databases and Search Strategy
The literature search was conducted across the following databases:
- PubMed
- Scopus
- Web of Science
- Google Scholar
- IEEE Xplore
Search terms included combinations of:
- “Artificial Intelligence (AI)”
- “Machine Learning (ML)”
- “Rehabilitation”
- “Patient care outcomes”
- “Healthcare technology”
- “Robotic therapy”
- “Ethical considerations in AI”
Boolean operators (AND, OR) were used to refine the search queries. For example: (“Artificial Intelligence” OR “Machine Learning”) AND (“Rehabilitation” OR “Patient care”) AND (“Robotic therapy” OR “Ethical challenges”).
Inclusion and Exclusion Criteria
Inclusion criteria:
- Peer-reviewed articles published in English.
- Studies focusing on AI applications in rehabilitation.
- Articles published between 2019 and 2024 to capture recent advancements.
- Narrative, systematic, or scoping reviews, as well as original research articles.
Exclusion criteria:
- Non-English publications.
- Studies not related to healthcare or rehabilitation.
- Articles lacking sufficient focus on AI technologies.
Timeframe
The literature search was conducted between January 2019 and December 2024, focusing on studies published in the past five years to capture the latest advancements in AI and rehabilitation technologies.
Data Extraction and Synthesis
Data from the selected studies were extracted and summarized. Key points from each study, including objectives, methodologies, findings, and limitations, were synthesized to provide a holistic overview of the field.
Rationale for a Narrative Review
This narrative review methodology was chosen due to the diversity and heterogeneity of the available literature on AI in rehabilitation. A narrative approach enables an integrative discussion, emphasizing thematic insights and practical applications rather than a rigid systematic analysis.
The Role of AI and Machine Learning in Rehabilitation: Key Applications
Personalized Treatment Plans
AI algorithms are revolutionizing the personalization of rehabilitation programs. By analyzing patient-specific data, such as medical history, real-time performance metrics, and recovery progress, AI can tailor treatment plans that cater to individual needs. For instance, machine learning models can assess how a patient is responding to therapy and recommend adjustments to their exercises, intensity, or duration to optimize recovery.
Studies indicate that AI-based treatment personalization improves patient engagement and accelerates recovery.3 AI can predict the likely progression of rehabilitation for different types of patients, including those with chronic conditions or neurological impairments, enabling therapists to make informed, data-driven decisions. A review highlighted that machine learning models can enhance patient outcomes by adapting rehabilitation plans to specific patient profiles.3, 9
Predictive Analytics
Predictive analytics is transforming rehabilitation by enabling practitioners to make more accurate forecasts regarding a patient’s recovery journey and potential complications. Through machine learning algorithms, predictive models analyze patient-specific data (such as demographics, health history, and prior recovery outcomes) alongside vast datasets from previous cases to make personalized predictions.
This information helps therapists to anticipate recovery milestones, adjust treatment plans proactively, and even identify risks of complications, leading to a more effective, customized care approach.10
Research highlights that AI-driven predictive analytics enhance care by allowing early detection of risks, enabling interventions that can prevent complications and improve recovery outcomes. In orthopedics, for example, these predictive tools have been shown to support more precise planning for patient recovery by identifying risk factors in real time and helping tailor therapeutic interventions accordingly. Overall, predictive analytics contributes to a customized, data-informed approach to rehabilitation that maximizes patient engagement and optimizes outcomes, reinforcing its role as a transformative tool in modern healthcare.11
Remote Monitoring
In neurorehabilitation, AI including machine learning (ML) techniques are increasingly employed to support personalized interventions and remote patient tracking, especially for conditions like stroke and spinal cord injuries. These technologies enable continuous monitoring of motor recovery, and their predictive analytics capabilities allow clinicians to tailor treatments to each patient’s specific progress and needs.
AI-driven models also enhance the accuracy of diagnosing neurological conditions and tracking functional improvements through devices that measure biometrics like movement and heart rate, which are central to rehabilitation outcomes.12
For cardiac rehabilitation, wearable devices combined with interpretable ML models allow for the tracking of functional capacity over time, using multi-parameter sensors such as ECG and accelerometers. Studies have shown that these devices, when integrated with ML algorithms, can predict outcomes like walking distance—a key indicator of functional improvement in cardiac rehab—by analyzing various data points such as chronotropic response. These advancements make it feasible to monitor patients remotely, supporting home-based rehabilitation and increasing accessibility for patients who may otherwise face barriers to traditional rehab programs.13
Rehabilitation Robotics
Rehabilitation robotics has emerged as a transformative approach in Rehabilitation, particularly for patients with mobility impairments due to neurological conditions, such as stroke or spinal cord injury. These robotic devices support task-specific, repetitive movement training—key elements for motor skill recovery in neurorehabilitation.
While robotic therapy offers promising advancements, evidence suggests it performs comparably to conventional therapy in certain areas Rehabilitation robots range from devices assisting limb movements to wearable exoskeletons, enhancing patient participation and optimizing rehabilitation outcomes through consistent practice.
Research indicates that robotic-assisted therapies improve motor function and patient engagement compared to conventional approaches.14 For example, upper-limb robotic devices, designed to facilitate arm and shoulder movements, allow patients to practice fine motor tasks with greater accuracy.15
Rehabilitation robots are classified into various types, each designed to support different aspects of movement recovery:
- Exoskeletons: Wearable devices providing structural support, enabling patients with lower-limb paralysis to perform actions such as walking.
- End-effector robots: Support specific limb segments like hands or feet for isolated movement exercises. Examples include Lokomat for gait training and ArmeoSpring for arm rehabilitation.
- Wearable robotics: Compact, sensor-driven devices such as smart braces and gloves that monitor movements and provide resistance for strengthening muscles.
Evidence suggests that exoskeletons and end-effectors play a crucial role in enhancing motor function and engagement in stroke and spinal cord injury recovery, making them valuable tools in personalized rehabilitation plans.14, 16
Benefits of AI and Machine Learning in Rehabilitation
AI and ML are revolutionizing rehabilitation medicine by significantly enhancing motor recovery.17 Their ability to assess, personalize, monitor, and optimize treatment plans underscores their growing importance and transformative potential in this field.
By incorporating AI which includes wearable sensors, virtual and augmented reality, and robotic devices, these technologies enable accurate movement analysis and adaptive neurorehabilitation methods. Incorporating AI and ML into motor recovery can significantly boost rehabilitation results, increase patient involvement, and make better use of resources.9
AI is making significant strides in neurologic physical therapy. Narrow AI is being developed for use in robotic-assisted therapy,18, 19 and to evaluate motor function, gait, functional status, upper extremity recovery, and movement.20 AI holds significant potential in analyzing functional task movements21 and prescribing personalized interventions. It can predict outcomes and create tailored care plans for individuals with neurological disorders.22, 23 Neurologic physical therapy professionals should collaborate with engineers, data scientists, and computer scientists to create AI tools that enhance clinical decision-making and provide patient-centered care.24
