Publication Details

Category Text Publication
Reference Category Journals
DOI 10.3389/fdgth.2026.1952373
Licence creative commons licence
Title (Primary) Sensor-wide association studies: a methodological architecture for developing and evaluating functional human traits
Author Lausch, A. ORCID logo ; Kirsten, T.; Selsam, P.; Bumberger, J. ORCID logo
Source Titel Frontiers in Digital Health
Year 2026
Department CLE; MET
Volume 8
Page From art. 1952373
Language englisch
Topic T5 Future Landscapes
Keywords digital biomarkers, explainable artificial intelligence, functional human traits, multimodal sensing, remote sensing, semantic data integration, sensor-wide association studies, wearable sensing
Abstract

Background:

Wearable and other digital sensors enable longitudinal observation of physiology and behaviour. However, reproducible associations and predictive performance do not establish which physiological construct a sensor-derived representation measures. An explicit framework is needed to distinguish statistical representation, construct interpretation and mechanistic explanation.

Objective:

This Hypothesis and Theory article proposes Functional Human Traits (FHTs) and Sensor-Wide Association Studies (SWAS) as a conceptual and methodological architecture for developing and evaluating individual-level representations of physiological response capacities.

Conceptual framework:

An FHT is defined as a quantitative representation of a specified physiological response capacity, supported by temporal repeatability over a justified interval and independent biological evidence under stated conditions. SWAS links documented sensor features, exploratory analysis, candidate construction and independent evaluation. Outcome-guided screening estimates feature–outcome associations, whereas unsupervised analysis identifies statistical structure; neither directly identifies a physiological construct. Candidate construction requires an explicit measurement model, individual-level estimation with uncertainty and tests addressing competing interpretations. Evaluation distinguishes statistical replication, technical robustness, temporal repeatability and biological support. A hypothetical cardiovascular heat-response example illustrates candidate construction using a longitudinal model. Environmental sensing and semantic integration are optional supporting components where relevant to the research question.

Implications:

SWAS integrates established association-study and sensor-based measurement principles around explicit FHT identification requirements. Biological support does not establish a causal mechanism or clinical usefulness; digital biomarker applications require separate validation for their intended use. The feasibility and added value of FHT/SWAS remain to be demonstrated empirically.

Lausch, A., Kirsten, T., Selsam, P., Bumberger, J. (2026):
Sensor-wide association studies: a methodological architecture for developing and evaluating functional human traits
Front. Digit. Health 8 , art. 1952373
10.3389/fdgth.2026.1952373