Key result
Machine learning for cardiac stress and anxiety detection requires high-quality, diverse data to generalize robustly.
Why the study?
Although machine learning classifiers can detect stress and anxiety from physiological sensor signals, the robustness and generalizability of these models for deployment in the field is rarely addressed.
Machine learning models for stress and anxiety recognition using cardiac signals require high-quality training data with minimal artifacts to generalize effectively outside their original experimental settings.
High-quality cardiac datasets may constrain clinical ML deployment for stress detection; leaves open which curation standards best enable generalization.
Many recent studies have addressed the detection of negative affective states such as stress and anxiety from physiological signals taken from body-worn sensors. Typically, machine learning classifiers are applied to features derived from sensor signals, and several authors have reported high accuracy results from a range of signals including cardiac, skin conductance, and skin temperature. However, the issue of how robust these models are for deployment in the field is rarely addressed. In this article, we use open data from two large experimental studies to evaluate the generalizability of models derived from cardiac signals, focusing on detection of stress and anxiety. We choose the cardiac signal since the commonly used heart rate variability features can be derived from multiple sensor modalities, allowing us to evaluate the robustness of models within, as well as between, experimental settings. We show that consistent classification outside the original experimental setting relies on high-quality training data with minimal artifacts, and that models may often train on proxies within the noise of lower quality data. Our results also underline the importance of including a wide range of emotional states in the training data to minimize erroneous classification from unseen regions of feature space.
No takes yet. Share an insight, caveat, or question.
Henry et al. (2023) studied Stress and anxiety. Machine learning models derived from cardiac signals was evaluated on Generalizability and robustness of models for stress and anxiety detection. Machine learning models for stress and anxiety recognition from cardiac signals require high-quality training data with minimal artifacts and diverse emotional states to generalize robustly.
Synapse has enriched 4 closely related papers on similar clinical questions. Consider them for comparative context: