The Wearable Says It Tracks Your Nervous System. The Evidence Base Is 18 Studies.
- Scoping review from the Universidad Politecnica de Madrid and Institut Guttmann (Barcelona), following JBI methodology and PRISMA-ScR; it mapped five years of research on using smartphones and commercial wearables to assess or predict interoception, chronic stress and self-regulation in adults.
- From 850 retrieved records, only 18 studies met inclusion criteria; 11 addressed chronic stress or stress reactivity, 5 self-regulation, and just 2 addressed interoception directly.
- Thirteen studies used wearables, three used smartphones, two combined both; ecological momentary assessment via smartphone appeared in eight; heart rate variability was by far the most common physiological measure (14 studies), followed by electrodermal activity and heart rate (4 each).
- Six studies applied machine learning, but only three reported classification accuracy, ranging from 56.8% to 79%; the authors describe the field as early-stage, wearable-dependent, and analytically modest.
Patients now arrive with data. A ring, a watch or an app tells them their "stress," their "recovery," their "nervous-system balance," and increasingly their interoception, and they want to know what you make of it. This scoping review is a useful map of what that industry can actually stand on. The short version: far less than the marketing implies.
What the data shows
Out of 850 records, only 18 studies met criteria for digital phenotyping of these body-mind domains in adults over the last five years. The distribution is telling. Chronic stress dominates because it is the easiest to instrument, self-regulation trails, and interoception, the construct most invoked in consumer copy about "listening to your body," was the direct focus of just two studies. The physiological substrate is narrow: heart rate variability does most of the work, with electrodermal activity and heart rate a distant second. Analytics are thin, six studies used machine learning, three reported accuracy at all, and that accuracy topped out at 79% and dipped to barely-better-than-chance. Integration of long-term monitoring and multiple sensing streams, the thing that would make continuous interoceptive assessment meaningful, is largely absent.
The authors are measured rather than dismissive: smartphones and wearables are genuinely promising as scalable, in-the-wild measurement, but the field is at the beginning, not maturity.
For your practice
This gives you a calibrated way to talk to patients about their devices. The wearable is measuring a real signal, mostly heart rate variability, but the leap from that to "your interoception" or "your stress state" is a marketing claim, not a validated readout, and only two studies in five years even attempted the interoception link. That matters clinically in both directions. For the health-anxious patient, device data is often another surface to scan and catastrophise, and naming the weak evidence base can be part of reducing that vigilance. For the disengaged patient, the same data can be a low-friction entry point into noticing bodily state, provided it is framed as a rough prompt, not a verdict. Either way, treat the numbers as conversation-starters about the patient's relationship to their body, not as measurements of it.
The wearable is reading heart rate variability; the claim that it reads your interoception is a marketing step the science has taken only twice in five years.
As a scoping review this maps the literature rather than pooling effects, and it depends on the small, heterogeneous set of studies it found; the fast-moving consumer market may be running ahead of the peer-reviewed record it summarises.