Japanese walking: what the study actually showed and how to test it yourself
Updated: 2026-09-22
"Japanese walking" is the viral name for interval walking training (IWT), a protocol designed at Shinshu University back in 2007. Social media sells it as a way to lose weight in a couple of weeks without dieting. The original study everyone cites measured nothing of the sort — it looked at leg strength, aerobic capacity and blood pressure in people 60+ over 5 months. The gap between what's proven and what's promised is what this piece is about. Then: how to test the effect on yourself in 4 weeks using your own ring or watch data, and the Health OS protocol built for exactly that.
What it is and where it comes from
Interval walking training (IWT) is a protocol from Hiroshi Nose and colleagues at Shinshu University: 3 minutes of fast walking alternating with 3 minutes of slow walking, at least 5 sets in a row (30 minutes), at least 4 days a week. It was designed in the mid-2000s as a more achievable alternative to continuous aerobic training for older adults.
In 2025–2026 the protocol went viral under the name "Japanese walking" on TikTok and Instagram: search interest rose 2968% year over year. Most videos frame it as a fast way to lose weight, without mentioning the original study or who it was actually tested on and for how long.
What the study actually showed
Nemoto et al. (2007, Mayo Clinic Proceedings) was a randomized controlled trial, 246 middle-aged and older participants (63±6 years old), 5 months. Three groups were compared: interval walking, continuous walking at a moderate pace, and a no-intervention control.
In the IWT group: isometric knee extension strength rose 13%, knee flexion 17%, peak aerobic capacity for walking 9%. Resting systolic blood pressure dropped more than in the continuous-walking group (P=.01). The trial doesn't report weight or fat-mass loss separately — the viral promise of losing weight "in weeks" has nothing to do with these numbers.
How to do it
Instead of a percentage of VO2max, use the talk test. Fast phase: a pace at which you can't hold a conversation easily (roughly 70% of peak aerobic capacity in the original protocol). Slow phase: a pace at which talking is easy (around 40%).
3 minutes fast, then 3 minutes slow, repeated 5 times = 30 minutes, at least 4 days a week. A 3-minute phone timer is the most reliable way to keep count of sets — you don't need a separate VO2 test.
What to watch in your own data over 4 weeks, and what is noise
The original trial didn't measure HRV, resting heart rate or readiness — you'll get those numbers from your own ring or watch, not from the paper. What to look at: your resting heart rate trend over 4 weeks of training (the steadier signal), HRV (noisier), and readiness/sleep score as a secondary marker.
One night tells you nothing: across our 701 Oura nights, consecutive nights differ in HRV by 17% on average (see the "HRV night to night" guide), while resting heart rate is several times steadier. Look at the 7-day rolling mean of resting heart rate before and after you start, not at individual days. Cardio adaptation in untrained people is usually visible within the first 2–4 weeks, but one good day after a session proves nothing on its own.
Who should be careful
This is general information, not medical advice. On blood-pressure medication or beta-blockers, check the fast-phase pace with a doctor — heart rate may not reflect real effort. With joint injuries, shorten the fast phase or choose softer ground. Stop and see a doctor for chest pain, shortness of breath disproportionate to the effort, or dizziness.
FAQ
Is it really a way to lose weight in a couple of weeks?
No — that's a viral promise, not a study result. Nemoto et al. (2007) measured leg strength, peak aerobic capacity and blood pressure in people 60+ over 5 months, not weight or fat mass over weeks. If the goal is weight loss, this isn't the protocol whose data supports that expectation.
How is it different from just walking fast the whole way at the same pace?
In the trial, the IWT group showed a greater drop in resting systolic blood pressure than the continuous-walking group at a moderate pace (P=.01). Alternating intensity is what was tested; there's no evidence the same result comes from steady walking of the same total duration.
How do I judge the effect if HRV bounces around from night to night?
Not from single nights. Look at the 7-day rolling mean of resting heart rate (several times steadier than HRV) and, if you want HRV too, its weekly mean. See the "HRV night to night" guide for how much HRV swings between nights.