Answers are grounded in this paper and cite their sources.
This research derives theoretical limits on surface measurements for estimating internal states, highlighting feasibility implications.
Background. Wearable physiological sensing has attracted enormous investment, and every generation of devices claims to recover more internal state from the skin surface: heart rate, blood oxygen, and increasingly blood glucose and other deep quantities. Yet a prior question has gone unasked. For a given internal state and a given surface modality, how much information about that state is physically present in the surface signal at all, before any algorithm touches it? Methods. We treat the body as a communication channel between a latent internal physiological state and a surface measurement, characterised by the conditional probability relationship imposed by tissue physics. We derive an upper bound on the mutual information between the two from Shannon channel capacity, parameterising the channel with published tissue optical and electrical properties. We pair this with a rate-distortion requirement giving the minimum information needed to estimate the internal state to a clinical tolerance, and we compare the two. Results. The feasibility test, information bound greater than or equal to clinical requirement, reproduces the known clinical record across five state-modality pairs. Blood oxygen from optical sensing and heart rate from photoplethysmography sit comfortably above their clinical requirements, matching two decades of clinical success. Non-invasive optical glucose and non-invasive intracranial pressure fall far below theirs, matching two decades of clinical failure. The separation is not marginal; it is a property of the channel, not of the receiver. Conclusions. A surface measurement that fails this test cannot be rescued by a better estimator, only by a better channel. The framework supplies regulators, procurement bodies, and research programmes with a falsifiable screen that separates a hard physical limit from an unsolved engineering problem.
No takes yet. Share an insight, caveat, or question.
Amit Walinjkar (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: