The connected body: Why your wearable is only seeing half the picture
A scientific perspective on why true athlete monitoring demands a full-body systems approach and what we're building at KULG to get there.
Something has bothered me for years in sports tech: we treat the human body like a collection of separate dashboards.
Heart rate over here. Sleep quality over there. GPS traces somewhere else. Stress scores in yet another app. Each metric lives in its own silo, reported as if it operates independently.
But here's what we know from physiology and what every athlete intuitively feels: Everything is connected. Every cell, every mitochondrion, every nerve.
The fatigue in your legs isn't just about running volume. It's about the quality of your sleep, which is affected by your stress levels, which are driven by your autonomic nervous system, which we can measure through heart rate variability. But only if we look at the whole picture simultaneously.
What the science actually shows
Recent research in wearable-based health monitoring confirms what many of us have suspected: psychological states, physiological markers, and behavioral patterns form an interconnected web, not isolated data points.
Consider what the evidence maps out:
Anxiety and stress don't just show up as "feeling tense." They manifest as measurable changes in heart rate variability, electrodermal activity (sweat response), respiratory rate shifts, and altered breathing patterns. Your nervous system broadcasts its state through multiple channels simultaneously.
Depression-related symptoms go even deeper. Research links them not only to physiological markers like reduced HRV, but strongly to behavioral signals: changes in circadian movement patterns, GPS location semantics (where you go and when), phone usage patterns, and disrupted daily rhythms. The body tells the story before the mind articulates it.
Cognitive performance - i.e. focus, attention, decision-making connects to heart rate reserve, phone interaction patterns, and typing behavior. Your brain's performance is visible in your fingers and your pulse.
Relaxation, the state every recovery protocol aims for, correlates with end-tidal CO₂ levels, fingertip temperature changes, oxygen saturation shifts, and skin conductance patterns.
This isn't theoretical. These are measurable, reproducible connections documented across hundreds of studies.
The problem with how we monitor today
Most wearable platforms pick one or two of these signals and build a product around them.
Whoop focuses on recovery scores based primarily on HRV and sleep
Oura tracks sleep and readiness through temperature and movement
Garmin gives you training load based on heart rate dynamics
Apple Watch monitors heart rhythm irregularities
Each captures a slice of reality. None captures the whole. It's like having six doctors, each looking at a different organ, none talking to each other and then wondering why the diagnosis is incomplete.
The interconnected system
Here's the framework that changes how we should think about athlete monitoring:
Layer 1: Psychological states
The internal experience, such as stress, anxiety, motivation, cognitive load, emotional valence. These are the upstream drivers.
Layer 2: Physiological markers
The body's measurable responses, such as heart rate, HRV, respiration, skin temperature, electrodermal activity, blood oxygen, CO₂ levels. These are the middleware between mind and behavior.
Layer 3: Behavioral signals
The outward patterns, such as movement rhythms, circadian regularity, sleep-wake cycles, interaction patterns, GPS semantics. These are the downstream expression.
The critical insight is that you cannot fully understand any one layer without the other two.
A drop in HRV might mean overtraining. Or it might mean poor sleep due to work stress. Or it might mean early signs of illness. Without behavioral context and psychological awareness, the physiological data is ambiguous.
What this means for athletes and coaches
For both recreational runners and elite athletes, the difference between a breakthrough performance and a DNF often comes down to reading these signals correctly, not just collecting them.
The question isn't "what was your resting heart rate today?" It's:
How does your HRV trend relate to your sleep this week?
Does your circadian movement pattern suggest recovery or suppressed motivation?
Are your respiratory patterns during rest consistent with parasympathetic dominance or sympathetic stress?
Is your training load sustainable given the full constellation of signals?
This is the difference between monitoring and understanding.
Why explainability matters
At KULG, our research pipeline processes over 100 scientific papers per week across sports science, AI coaching, physiological monitoring, and wearable technology. One finding keeps recurring: the AI models that athletes and coaches actually trust are the ones they can understand.
Research on explainable AI for wearable health monitoring (Kuschel et al., 2026) demonstrates that it's possible to build interpretable systems without sacrificing performance, using what the authors call Inherently Interpretable Components. This matters enormously because a coach who can see why the system recommends reducing training load ("your HRV is suppressed 15% below baseline, your circadian movement irregularity score is elevated, and your sleep onset variability has increased for three consecutive nights") will make better decisions than the one who just sees "Recovery Score: 42."
The explanation is the value. Not just the number.
The full-body monitoring thesis
We believe the next generation of athlete monitoring won't come from better individual sensors. It will come from connecting the signals that already exist into a coherent, interpretable, whole-athlete picture.
Every cell, every mitochondrion, every nerve is connected. Your mitochondria produce ATP in response to training stimuli, regulated by hormonal signals influenced by sleep quality, affected by stress hormones, driven by psychological state, reflected in heart rate variability, and expressed in running performance.
You don't need to measure every cell. But you need to measure enough of the right signals, in the right layers, with the right interpretive framework.
That's what we're building at KULG: not another dashboard, but a connected understanding of the whole athlete.
References and further reading:
Kuschel, M., Vieluf, S., Reinsberger, C., Loddenkemper, C., & Hasija, T. (2026). Explainable AI Using Inherently Interpretable Components for Wearable-based Health Monitoring. arXiv:2603.12880
Bangsbo, J. et al. (2024). Physiological basis of elite endurance performance. Scandinavian Journal of Medicine & Science in Sports
KULG Research Pipeline (2026). Continuous autonomous synthesis of sports science literature. Internal technical documentation.
Picard, R. W. (2020). The Circle of Emotion. MIT Media Lab — on measuring what matters in affective computing.
