The review, published in Discover Artificial Intelligence and conducted according to PRISMA guidelines with prospective registration in PROSPERO, systematically searched PubMed, Scopus, Web of Science, and Google Scholar up to July 27, 2026. From 2,182 records identified, the researchers led by Parsa Samaei and colleagues at the University of Tehran ultimately included thirteen studies published between 2022 and 2026. Each study applied some form of artificial intelligence, from conventional machine learning to deep neural networks, to the diagnosis, classification, or subtype identification of ankle instability. The team then appraised the methodological quality and risk of bias of every included study using PROBAST + AI, a specialized assessment tool designed for prediction models in medicine.

The headline result is striking. Across the five studies that evaluated diagnostic performance, reported area under the receiver operating characteristic curve values ranged from 0.892 to 0.999, figures that approach the theoretical maximum for a diagnostic test. The strongest results came from imaging. Transfer learning models built on ultrasound images of the ankle, using architectures such as EfficientNet, ResNet-50, and MobileNetV2, achieved AUC values between 0.995 and 0.999 for diagnosing chronic lateral ankle instability. A transformer-based multilabel network called AnkleNet detected lateral and medial ligament injuries on magnetic resonance imaging with AUC values of 0.910 and 0.892 respectively, and notably improved the diagnostic performance of the clinicians who used it. A radiomics-based approach even detected subtle architectural changes in cartilage and subchondral bone that are invisible to conventional image reading, reaching an AUC of 0.921.

Perhaps most intriguing for everyday medicine, the review found that advanced imaging is not the only path to accurate AI diagnosis. One study equipped shoes with integrated sensors and used an XGBoost model trained on plantar pressure and inertial sensor data to identify individuals with chronic ankle instability, achieving 93.39 percent accuracy and an AUC of 0.959, while also predicting rehabilitation probability in ways that correlated with clinical recovery. Another deep learning system automatically assessed weight-bearing ankle radiographs, correlating strongly with expert clinician measurements of talar tilt and anterior talar translation. These results suggest that AI could eventually serve as a quantitative second reader, supporting rather than replacing clinical expertise.

Classification studies, which used biomechanical, neuromuscular, and clinical data rather than imaging, reported equally eye-catching numbers. A ConvLSTM model trained on gait data augmented with a dual generative adversarial network achieved 100 percent accuracy, albeit in a small laboratory sample of just thirteen participants. A graph neural network combined with attention reinforcement learning, dubbed GaitNet, reached 96 percent accuracy and an AUC of 1.00 on three-dimensional gait kinematics. A hybrid convolutional neural network and random forest model identified neuromuscular deficits from surface electromyography during unanticipated landing tasks with 96 percent accuracy and an F1-score of 0.95. Even conventional algorithms performed well: random forest reached an AUC of 0.967 using heel-rise kinematic variables, and support vector machines achieved AUC values above 0.80 using anthropometric and functional measurements in delivery workers.

One study pushed beyond diagnosis altogether. Using unsupervised k-means clustering, researchers identified five clinically meaningful subtypes of chronic ankle instability, hinting that what clinicians currently treat as a single condition may in fact be several distinct disorders with different mechanical and functional signatures. If confirmed, this could open the door to genuinely personalized rehabilitation, targeting interventions to the specific impairments that define each patient’s subtype rather than applying one-size-fits-all protocols.

But the review’s authors are careful to pour cold water on the temptation to declare victory. When they applied the PROBAST + AI tool, seven studies showed low overall risk of bias, five were judged high risk, and one was unclear. The concerns clustered in the analysis domain, covering model development, validation, sample size, and generalizability. Several studies achieved extremely high accuracy using small or highly selected datasets, a classic warning sign of overfitting, where a model memorizes the quirks of its training data rather than learning generalizable patterns. External validation, in which a model is tested on data from an entirely different institution or population, was rare. The authors emphasize that even a low risk-of-bias judgment simply means no major concerns were identified, not that a model is proven robust.

Heterogeneity compounds the problem. The thirteen studies differed in nearly every dimension imaginable: participant populations ranged from athletes to parcel delivery workers to retrospective imaging cohorts of 4,000 patients; input data spanned MRI, ultrasound, radiographs, gait kinematics, electromyography, wearable sensors, questionnaires, and anthropometric measurements; and the AI architectures ranged from logistic regression and naive Bayes to transformers, graph neural networks, and generative adversarial networks. Because each study reported different performance metrics on different tasks with different reference standards, the numbers cannot be directly compared, and no single algorithm can yet be crowned superior. The review also notes that its own search excluded IEEE Xplore, EMBASE, and the Cochrane Library, leaving open the possibility that relevant studies were missed.

What emerges is a field moving at remarkable speed but still in its adolescence. The authors conclude that AI-based approaches show substantial potential for supporting the diagnosis and classification of ankle instability across imaging, biomechanical, wearable-sensor, and clinical modalities, but that reported performance should be interpreted as preliminary rather than as evidence of established clinical effectiveness. Their prescription for the road ahead is clear: larger, multicenter studies with standardized reference standards, transparent reporting, rigorous external validation, and prospective clinical evaluation. Until then, the dazzling accuracy figures should be read as a promise of what AI might deliver for the millions of unstable ankles worldwide, not yet as a diagnosis you can expect to receive at your next appointment.

Subject of Research: Artificial intelligence for the diagnosis and classification of chronic ankle instability

Article Title: The role of artificial intelligence for the diagnosis and classification of ankle instability: a systematic review

Article References: Samaei, P., Ebrahimi, E., Molavi, M. R., & Ardakani, M. K. (2026). The role of artificial intelligence for the diagnosis and classification of ankle instability: a systematic review. Discover Artificial Intelligence, 6(1), Article 1306. https://doi.org/10.1007/s44163-026-02412-8

Image Credits: AI Generated

DOI: 10.1007/s44163-026-02412-8

Keywords: artificial intelligence, chronic ankle instability, machine learning, deep learning, diagnosis, classification, ankle sprain, systematic review, wearable sensors, medical imaging, biomechanics, risk of bias

Gait initiation is a deceptively complex act. The nervous system must shift the body’s center of mass forward and toward the swing limb, unweight the trailing leg, and generate carefully sequenced bursts of muscle activity, all within a few hundred milliseconds. During this window, sensory receptors embedded in muscles, tendons, joints, and skin fire continuously, informing the central nervous system about joint position, movement velocity, and load. Researchers have long measured how quickly muscles respond to perturbations in people with chronic ankle instability, but far less attention has been paid to the afferent side of the circuit — the incoming stream of proprioceptive signals that arrives before any muscle can react. Because direct recordings of these nerve signals in humans are ethically and technically impractical, the research team turned to a computational approach.

The study combined detailed biomechanical analysis of gait initiation with a neurophysiological model of proprioceptor behavior. Participants, including individuals with chronic ankle instability and healthy controls, walked in a motion capture laboratory equipped with force plates that recorded the ground reaction forces beneath each foot. From the movement trajectories and forces, the researchers reconstructed the mechanical events inside and around the ankle joint: changes in joint angle, angular velocity, muscle fascicle length, and the deformation of skin and joint capsules. These mechanical variables were then fed into mathematical models describing how different classes of proprioceptive afferents translate mechanical deformation into trains of neural impulses.

The models captured the behavior of several key sensory populations. Muscle spindles, the primary detectors of muscle length and stretch velocity, were modeled with their characteristic dynamic and static response components. Golgi tendon organs, which signal tensile load within tendons, were modeled as force-sensitive elements. Cutaneous and joint mechanoreceptors, which fire in response to skin stretch and joint capsule deformation, were included as well. Each model produced an estimate of instantaneous firing rate — effectively, a simulated neurogram of what the ankle’s sensory apparatus was telling the spinal cord and brain at every millisecond of the gait initiation sequence.

The central finding is striking: in individuals with chronic ankle instability, the estimated timing of proprioceptive afferent activity was systematically altered during the anticipatory postural phase of gait initiation. Whereas healthy participants showed well-organized, temporally structured afferent volleys that preceded and accompanied the postural adjustments needed to step forward, the unstable ankles displayed shifts and distortions in these signal patterns. In practical terms, the sensory information reaching the nervous system about the ankle’s position and loading arrived in a different temporal pattern than in uninjured joints. Because the nervous system depends on precisely timed sensory feedback to calibrate motor commands, even modest shifts in afferent timing could degrade the coordination of the postural adjustments that keep the body balanced during the transition to stepping.

The researchers emphasize that the disruption is not simply a matter of weaker signals. The models suggested changes in the relative timing between different receptor populations, meaning that the brain may receive internally inconsistent information about what the ankle is doing. Muscle spindle signals indicating one joint position may arrive in a different relationship to cutaneous signals than the nervous system has learned to expect. This form of sensory reweighting or desynchronization is thought to be a hallmark of the maladaptations that follow ligament injury, when damaged mechanoreceptors in the torn ligament are lost and remaining receptors must compensate — often imperfectly — for the missing input.

What makes this approach particularly powerful is that it opens a window onto a process that has been essentially invisible in human experiments. Traditional assessments of proprioception in chronic ankle instability, such as joint position matching tasks or thresholds for detecting passive movement, measure perception at one static moment in time. They cannot capture the rapid, continuous stream of sensory traffic that flows during movement. By deriving afferent firing patterns from movement data through validated receptor models, the study provides a dynamic, millisecond-by-millisecond estimate of sensory signaling during a functional task. This is a fundamentally different lens from reactive measures such as reaction times or postural sway scores, which reflect the end result of sensory-motor integration rather than its incoming raw material.

The clinical implications are potentially significant. Rehabilitation programs for chronic ankle instability currently emphasize balance training, strengthening, and perturbation-based exercises, which are effective for many but not all patients. If the root of the problem lies in altered afferent timing during the anticipatory phase of movement, then therapies could be designed specifically to recalibrate sensory signaling. Tasks that demand precise ankle positioning under time pressure, or training environments that manipulate the relationship between movement and sensory feedback, might encourage the nervous system to relearn the temporal patterns it lost after injury. The modeling framework also suggests new outcome measures: instead of relying solely on questionnaires and postural tests, clinicians could one day track afferent timing metrics derived from inexpensive motion capture to quantify sensory recovery objectively.

The study also fits into a broader scientific movement that treats sensory timing as a central currency of motor control. Research across domains, from upper limb reaching to locomotion in older adults, has shown that the nervous system is exquisitely sensitive to delays and variability in sensory feedback. Even artificial delays of tens of milliseconds can disrupt motor learning and adaptation. In this light, chronic ankle instability can be understood not merely as a mechanical joint problem or a strength deficit, but as a disorder of sensorimotor timing — a failure of the dialogue between the ankle and the spinal cord to keep the conversation synchronized during the most demanding phases of movement.

There are, of course, important caveats. The afferent patterns in this study are model-derived estimates, not direct neural recordings, and their accuracy depends on the fidelity of the underlying neurophysiological models and the biomechanical reconstructions. Individual variability in anatomy, injury history, and compensation strategies means that the group-level patterns reported here may not apply uniformly to every patient. The researchers also note that gait initiation, while clinically relevant, is only one of many tasks in which altered afferent timing could play a role; landing, cutting, and uneven-terrain walking may each impose distinct sensory demands. Future work combining these modeling techniques with other modalities, such as electroencephalography or high-density electromyography, could trace how altered afferent timing propagates through the central nervous system to produce the characteristic motor deficits of the condition.

Nevertheless, the study marks an important step toward a mechanistic, quantitative understanding of one of sports medicine’s most stubborn problems. By converting observable movement into estimated neural signaling, it transforms chronic ankle instability from a fuzzy clinical label into a measurable pattern of sensory disruption with specific temporal signatures. If subsequent studies confirm and extend these findings, the humble ankle sprain — so often dismissed as a trivial injury — may finally be understood at the level of the nerve impulses it deranges, and treated accordingly.

Subject of Research: Model-derived proprioceptive afferent timing during gait initiation in chronic ankle instability

Article Title: Model-derived proprioceptive afferent timing during gait initiation in chronic ankle instability

Article References: Bijari, K., Sadeghi, H., Yousefi, M., & Yiou, E. (2026). Model-derived proprioceptive afferent timing during gait initiation in chronic ankle instability. Scientific Reports. https://doi.org/10.1038/s41598-026-71074-4

Image Credits: AI Generated

DOI: 10.1038/s41598-026-71074-4

Keywords: chronic ankle instability, proprioception, gait initiation, afferent timing, muscle spindles, Golgi tendon organs, sensorimotor control, computational modeling, biomechanics, ankle sprain, postural control, neurogram