Why watching you walk matters more than you might think

How do you know whether your knee is genuinely deteriorating — or just having a difficult week?

An MRI can show cartilage wear; an X-ray can reveal joint space narrowing. What neither captures is how your body actually moves through an ordinary morning: the slight trunk sway compensating for a weaker hip, the extra momentum needed to push out of a chair, the asymmetric push-off that loads one joint more than the other. These movement patterns can signal meaningful joint stress well before pain crosses clinical thresholds — or before anything shows clearly on a scan.

That gap between structural imaging and lived movement is where objective gait analysis sits. MAI Motion® — a UKCA/MHRA-registered, AI-powered markerless motion capture platform — converts a short video of natural movement into reproducible biomechanical data, without wearables, markers, or laboratory equipment. The numbers it generates are structured through a five-dimension framework called C.R.A.F.T., introduced by Professor Paul Lee in Practical Regeneration (published March 2026), which translates raw motion data into a readable picture of how a joint is actually functioning.

That picture, updated over successive scans, is what can change a surgical conversation.

The five C.R.A.F.T. dimensions — and what each one is measuring

Each short video generates scores across five distinct dimensions, and the logic behind each one connects directly to how joints degrade over time.

C — Control refers to how precisely the nervous system moderates joint movement under load. A well-controlled knee, for instance, tracks smoothly and predictably through a sit-to-stand repetition. When control is reduced — often through fatigue, muscle inhibition, or early degeneration — the joint compensates with small wobbles and load shifts that accumulate over thousands of daily repetitions.

R — Repetition measures consistency across multiple movement cycles. Healthy joints repeat the same pattern reliably; a joint under stress tends to drift — each step or squat looking slightly different from the last. That variability is a signal worth capturing, because it often appears before pain does.

A — Asymmetry tracks side-to-side differences in load, range of motion, and timing. It is the dimension patients most readily recognise in everyday life: uneven shoe wear, always pushing off the same leg to rise from a chair, or favouring one side on stairs. Even small asymmetries, sustained across every movement of the day, concentrate stress on one joint surface.

The fourth dimension captures spinal and pelvic alignment — how the body stacks and balances itself in both static posture and dynamic movement. Pelvic tilt and spinal compensation patterns frequently drive abnormal knee loading from above, so this dimension extends the picture beyond the joint itself.

The fifth focuses on velocity and power: how quickly and forcefully movement is generated. A drop in power output — needing a run-up to rise from a low chair, for example — can reflect both muscle capacity and the nervous system's protective down-regulation around a painful joint.

Together, these five measures produce what Professor Lee calls a movement fingerprint: a personal signature that is reproducible across scans, comparable against age-matched population norms, and — crucially — trackable over time.

How MAI Motion captures a movement fingerprint without a lab

You walk, squat, or push up from a chair while a standard camera films you. Nothing is attached to your body — no reflective markers, no laboratory suit, no specialist hardware of any kind. That is the patient's entire contribution to the capture process.

Underneath, the system is doing something considerably more involved. MAI Motion® tracks 15 anatomical keypoints — shoulders, hips, knees, ankles, and key spinal landmarks — at 120 frames per second, then converts that 2D footage into a 3D interactive skeletal model representing how your joints moved through every millisecond of the task. The resulting data is rich enough to distinguish a compensation pattern from genuine strength: not just whether you rose from the chair, but how, with what asymmetry, at what speed, and how consistently from one repetition to the next.

Those same population norms described above underpin the system's headline output: a single number called Motion Age — a biological score derived entirely from how you move, benchmarked against people of the same age. The sit-to-stand metrics that feed this score — specifically the smoothness and cumulative acceleration of knee and elbow flexion — are among the measures with the strongest published evidence, showing statistically significant differences (p<0.05) pre- and post-injection in peer-reviewed work; broader validation of the full system continues alongside clinical use.

The initial assessment takes around 30 minutes with a trained clinician. Subsequent re-scans can be completed at home via the MAI Motion app, with every result plotted automatically against the original baseline in a Regen OS dashboard — giving both patient and clinician a shared view of trajectory over time, rather than a series of isolated snapshots.

What Motion Age tells you — and how it differs from your real age

Think of Motion Age the way you might think of a metabolic age score from a fitness tracker — except grounded in clinical movement measurement rather than a consumer device. If your joints move the way a 58-year-old's typically do and you are 45, that 13-year gap becomes the starting point for intervention, not an abstract number. Equally, a Motion Age below your chronological age suggests your joints are loading and recovering efficiently for someone younger than you — a meaningful reassurance in its own right.

The benchmark is deliberately realistic: population norms drawn from age-matched individuals, not elite athletes. The question is not 'how close are you to peak performance?' but 'how do you compare with others your age?'

Some of the patterns that widen that gap are ones patients already sense before they name them:

• Uneven shoe wear on the inner or outer heel
• Recurring tightness that is reliably worse on one side
• Needing a small momentum shift — a forward lean or arm push — to rise from a low chair
• Noticeably slower or less stable balance on one leg during routine tasks such as putting on shoes

Most people on a structured, evidence-based programme see their Motion Age fall below their chronological age within 16 weeks — that is an observed trend within such programmes, not a clinical guarantee for any individual.

The practical shift for patients is straightforward: instead of 'my knee feels a bit better this week', the Regen OS dashboard shows whether the underlying movement pattern is genuinely improving — or holding steady in a way that may warrant a different clinical decision.

When movement data changes a surgical decision

Consider what that trajectory data means when surgery is on the table.

In Practical Regeneration, Professor Paul Lee describes the case of a patient named David, whose outcome illustrates precisely how C.R.A.F.T. scores feed into a clinical decision rather than simply confirming one. After an initial intervention, the team repeated the MAI Motion® assessment at six weeks and again at twelve weeks, tracking three specific outputs: stance symmetry, the shape of the flexion curve through each movement cycle, and the timing of rotational patterns. Those three variables formed the basis of the next decision.

The curves moved in the right direction. Surgery was deferred.

Had they not — had the data shown a flat or declining trajectory — the plan was already in place: prompt escalation to a personalised knee replacement using the Twis-TKR approach, with the same movement data used to individualise the implant design. As Lee writes, without objective tracking 'David's path would have been guesswork… with MAI-Motion, we had evidence, timelines and options.'

That transparency is the patient value proposition Lee is most explicit about. Traditional follow-up appointments involve a clinician observing the patient walk, bend, and describe their symptoms — then forming an impression. 'You're basically at the mercy of someone's Tuesday mood,' he notes. C.R.A.F.T. replaces that impression with a number, and a trend, that any clinician reading the dashboard will interpret the same way on any day of the week.

The biomarker validity that makes this decision logic clinically defensible rests on peer-reviewed motion-capture research — including Armstrong et al. (2022) and the work in Musculoskeletal Regeneration Medicine discussed earlier — which established that markerless video analysis can produce biomechanically meaningful data without a laboratory setting.

For patients, the practical implication is straightforward: the threshold for escalation is defined in advance, not improvised at the point where something feels wrong.

Getting a C.R.A.F.T. assessment — what the process looks like

For anyone who has spent months describing the same knee symptom to different clinicians and left each appointment uncertain whether anything had actually changed, an objective movement baseline is itself a form of progress — the start of a record that can be interrogated rather than recalled.

No GP letter is needed, and no specialist equipment. The initial assessment — conducted by a trained MSK Doctors clinician using MAI Motion® — takes around 30 minutes and requires nothing more than moving naturally in front of a standard camera. Patients attending the Regeneration Hub in Sleaford (NG34) or the Grantham diagnostics centre (NG31) can book directly, with consultant-led review available without an NHS-style triage delay. London-based patients can access the same assessment through the London Cartilage Clinic at Harley Street.

Once the baseline is established, subsequent re-scans can be completed at home via the MAI Motion® app, so the trend that guides clinical decisions continues to build between appointments — without the need to attend clinic each time.

The assessment itself is one appointment. What it produces is an ongoing, objective record: the kind of data that, as the earlier sections describe, has the capacity to defer unnecessary surgery or — where a flat trajectory demands it — accelerate a more definitive plan before the window for one closes.

To book without a referral, visit mskdoctors.com.