Multimodal Modeling of Ultradian Rhythms Using the Hankel Alternative View of Koopman (HAVOK) Analysis

Area of Science:

• Physiology
• Chronobiology
• Complex Systems

Background:

• Ultradian rhythms are biological fluctuations shorter than 24 hours, crucial for sleep and hormone regulation.
• Their underlying mechanisms are not well understood, hindering therapeutic interventions.
• Disruptions in ultradian rhythms are linked to neurological and psychiatric disorders.

Purpose of the Study:

• To apply the Hankel Alternative View of Koopman (HAVOK) framework to model and understand ultradian rhythm dynamics.
• To investigate the potential role of sex hormones in regulating ultradian rhythms.
• To explore the clinical relevance of ultradian rhythm dynamics for therapeutic strategies.

Main Methods:

• Continuous physiological monitoring using wearable sensors in 16 healthy participants over two weeks.
• Systematic removal of circadian and longer-scale rhythms to isolate ultradian dynamics.
• Application of the HAVOK framework, a data-driven method based on Koopman operator theory, to model the isolated ultradian fluctuations.

Main Results:

• Ultradian rhythms were characterized as intermittently forced linear systems by the HAVOK model.
• The model demonstrated high forecasting accuracy for ultradian dynamics (RMSE ~0.02-0.03).
• A significant sex difference in model rank suggests a role for sex hormones in ultradian rhythm regulation.

Conclusions:

• The HAVOK framework provides a robust method for characterizing ultradian rhythms as intermittently forced linear systems.
• Intermittent forcing is a consistent feature across participants' ultradian dynamics.
• Findings suggest sex hormones influence ultradian rhythm dynamics, offering potential targets for chronotherapy and biomedical interventions.