The gap in your movement health data

Your wristband says 9,000 steps. Your knee has been quietly aching for three weeks. Neither number tells the other's story.

Consumer wearables are exceptionally good at one thing: counting. Steps, active minutes, heart rate — the arithmetic of movement. What they cannot see is the quality beneath those numbers: whether your left hip drops fractionally with every stride, whether you are offloading a stiff ankle by rotating your pelvis, whether a compensation pattern is slowly accumulating load in a joint that will, eventually, make itself known as pain. That distinction — how much you move versus how well you move — is not a minor technical footnote. It is, as the evidence now makes clear, the difference between a signal that predicts risk and one that merely counts activity.

Clinician assessment can see gait quality, but it arrives rarely, costs time, and is shaped by whoever happens to be in the room that day. For most people, the picture of their own movement mechanics exists nowhere in their personal health data.

This is the movement data gap: a whole category of biomechanical signal that is visible in every step you take yet almost entirely unmeasured. Professor Paul Lee's MAI Motion® — a markerless AI platform requiring nothing more than a standard camera — is designed specifically to close it.

Why gait speed predicts lifespan

The numbers from a 2011 pooled analysis by Studenski and colleagues are difficult to argue with. Drawing on nine cohort studies and 34,485 community-dwelling adults aged 65 and over, the research found that for every 0.1 metres per second increase in walking speed, all-cause mortality risk fell by 12%. At age 75, predicted ten-year survival in men ranged from 19% to 87% — a range determined almost entirely by how fast those men walked. The benchmarks that emerged are now widely cited: a pace of 0.6 m/s or below signals significantly elevated mortality risk; 0.8 m/s tracks with median life expectancy for age and sex; 1.2 m/s or above is associated with exceptional longevity.

What makes this more than an orthopaedic curiosity is the reach of that signal. A 2025 BBC Future investigation highlighted research showing that slower walkers tend to have smaller brain volumes and structural differences in key brain regions. Walking pace, it appears, reflects the integrated state of the nervous system, cardiovascular capacity, muscle quality, and joint function simultaneously — a single observable behaviour that encodes the condition of many systems at once.

This is not a niche rehabilitation metric. It is arguably the most accessible whole-body health readout available to anyone with a stretch of pavement. The question the science raises is a practical one: given how much information gait carries, almost nobody is tracking it with any consistency or precision — and that is exactly the gap the next generation of movement tools is built to fill.

How markerless capture makes clinical measurement accessible

Traditional gait laboratories solve the measurement problem expensively. Reflective markers must be attached to the body at precise anatomical landmarks, multiple calibrated cameras ring the walkway, and a trained operator is needed before a single stride is recorded. The setup is clinically rigorous — and entirely impractical as a routine personal monitoring tool for anyone not referred into a specialist centre.

Markerless capture changes that equation. Rather than instrumenting the body, deep-learning pose estimation models read standard video — from a smartphone or tablet — and extract skeletal keypoints in real time. No markers, no calibration rigs, no specialist setup. The body moves naturally; the system infers structure from motion.

The critical question is whether the results are credible. A 2022 feasibility study by McGuirk and colleagues, published in Frontiers in Human Neuroscience, assessed 166 individuals aged 9 to 87 across six community settings. All 12 spatiotemporal parameters — including cadence, step length, and stride time — showed good-to-excellent agreement with a gold-standard pressure walkway. That is community-setting validation, not a controlled laboratory finding. Separately, Carvalho et al. (2025) established that as few as seven strides are sufficient for reliable kinematic data in older adults, which makes the brief at-home re-scan a practically achievable proposition rather than a theoretical convenience.

MAI Motion operates within this validated technology class, tracking 15 skeletal keypoints at 120 frames per second with no wearables or calibration required. Known limitations of markerless systems apply here too: ankle-joint accuracy is reduced compared with full marker-based analysis, and accuracy narrows at extreme movement speeds. For the movement quality assessment that makes up routine monitoring, however, these constraints sit well outside everyday walking, squatting, and sit-to-stand tasks.

What C.R.A.F.T. scoring actually measures

Five dimensions sit underneath every MAI Motion assessment, grouped under the C.R.A.F.T. framework: Control, Repetition, Asymmetry, Flow, and Twist. Together they map how the body actually distributes load, compensates for weakness, and organises movement through space — information that a step count never touches.

Control examines how well a joint is governed through its range at a given speed and load. Repetition asks whether that quality holds across multiple cycles, or degrades with fatigue. Asymmetry compares left and right sides: a hip that drops three degrees more on one stride than the other is flagged even when it produces no pain. Flow captures the smoothness and efficiency of movement — the difference between a stride that absorbs energy well and one that leaks it through compensatory micro-adjustments. Twist, the rotational dimension, looks at how the pelvis, thorax, and shoulders sequence through a step, since rotational stiffness is a common early marker of functional decline.

The design motivation is interception before symptoms arrive. Professor Paul Lee built the platform around the principle that objective, frame-by-frame data should replace the inherent inconsistency of purely visual clinical assessment — a judgement call that varies with the assessor's experience, fatigue, and angle of view. The case of Raj in Practical Regeneration illustrates what that shift produces: a short video surfaced foot flare on the left, hip drop on the right, and absent glute engagement. A targeted retraining plan produced significant improvement within six weeks, without surgery.

All of this feeds into a single output: Motion Age, a biological age score derived by comparing an individual's movement signature against age-matched population norms. It is a proprietary Regen PhD metric and has not yet been independently peer-reviewed — it should be read as a consistent internal benchmark rather than a clinically validated universal standard. That said, a repeatable internal benchmark is itself useful: it removes the noise of inter-assessor variation and makes progress trackable across months of training.

From one-off scan to ongoing habit

The pathway begins in person. The initial MAI Motion session — a supervised 30-minute appointment at the Harley Street clinic — produces a full baseline report from which everything else is built. It runs alongside a 32-marker blood panel covering six biological systems: inflammation, metabolic, hormonal, cellular energy, cardiovascular, and liver and renal function. Together these form the dual Physics and Chemistry baseline that anchors the Regen PhD Scan layer — the first stage in the Scan → Optimise → Learn framework.

From that baseline, re-scans can happen at home via the MAI Motion app, using the same capture pipeline. The brevity that makes markerless assessment feasible in community settings applies equally here — a session takes only minutes. Every result feeds into Regen OS, where delta tracking turns individual data points into a trend: a personal movement record that accumulates meaning with each visit rather than sitting as a single historical snapshot.

The re-scan cadence is where habit formation enters. In Practical Regeneration, Professor Paul Lee sets out a clear behavioural architecture: six days to ignite a new behaviour, six weeks for it to embed itself into daily life. The re-scan schedule is designed with that window in mind — a consistent, low-friction check-in intended to shift gait monitoring from a one-off clinical event into an ongoing personal practice. It is worth noting that evidence directly linking repeated scanning to durable long-term behaviour change remains limited; what regular re-scanning does provide is a structured prompt and a visible record of progress — both of which the habit-formation literature considers meaningful conditions for sustaining any new routine. Movement monitoring, under this model, becomes less a diagnostic intervention and more a pillar of the Physics dimension in everyday life.

MAI Motion as the Physics pillar in practice

Professor Paul Lee organises health around four interdependent pillars: Physics, Chemistry, Biology, and Time. MAI Motion sits squarely in the Physics column — but a compensatory gait pattern does not stay there.

When one hip drops on every stride, the joints above and below absorb load unevenly. Sustained asymmetrical loading may drive a low-grade inflammatory response — a Chemistry signal. That same chronic micro-stress can disrupt tissue repair and recovery signalling, which belong to the Biology dimension. This is why the Regen PhD Scan layer pairs the MAI Motion movement assessment with a blood panel: Physics and Chemistry are read together by design, because the body does not process them in separate lanes.

The Time pillar is the most practically urgent dimension here. Early biomechanical drift — a gradual shortening of stride, a creeping hip drop, reduced thoracic rotation — is detectable in frame-by-frame movement data before it registers as pain or limits function. Waiting for a symptom means waiting for the easiest correction window to close. The re-scan cadence exists for precisely this reason: a consistent C.R.A.F.T. record across months makes any regression visible at the point where it is still straightforward to address, rather than after compensation has become load-bearing habit.

The practical entry point is understanding where your compensations currently sit and what they are costing the rest of the system. Raj's foot flare and absent glute engagement were invisible without objective data; once identified, they became a directed six-week retraining plan with measurable results. That movement from unnamed compensation to specific, actionable target is where the Physics pillar stops being theoretical and starts shaping how the other three pillars perform.