Circadian Rhythmicity Analysis Using Optimized Cosinor Models and Entropy Measures in Real Night-Shift Workers

Introduction: Shift work, particularly night shifts, disrupts circadian rhythms, adversely impacting workers' health and performance. Core body temperature (CBT) is a robust physiological marker of circadian rhythmicity, yet in real world setting it’s underused due to measurement challenges. When available, traditional methods of analysis assume a fixed period (typically 24 hours) to characterize circadian oscillations. However, shift workers may exhibit non-standard or irregular rhythms, highlighting the need for more flexible and rigorous analytical approaches. In this study, we aimed to develop and apply advanced mathematical models to detect rhythmicity, quantify irregularity, and assess circadian disruptions in real-world shift work settings.

Methods: We collected continuous core body temperature data from 27 shift workers engaged in rotating night shifts using CorTemp® sensors. To analyze circadian rhythmicity, we applied the cosinor model but optimized the period parameter rather than fixing it at 24 hours, thus allowing the identification of the most representative periodicity for each individual. For cases where rhythmicity was weak or absent, we employed entropy-based measures and nonlinear dynamic indicators of chaos (e.g., approximate entropy, sample entropy) to quantify irregularity and complexity in temperature signals.

Results: Optimizing the period parameter revealed significant heterogeneity in circadian rhythmicity among shift workers. While many exhibited near-24-hour periodicity, indicative of preserved circadian rhythmicity, a subset of workers displayed significantly altered or irregular periodicities. Entropy and chaos analysis effectively distinguished between stable and disrupted temperature patterns. Higher entropy values correlated with severe disruption of circadian rhythmicity, suggesting a possible physiological signature of chronic circadian misalignment.

Conclusion: Our findings demonstrate that mathematical optimization of cosinor models and the application of entropy-based metrics provide sensitive tools for identifying and quantifying circadian rhythm disruptions in shift workers. This approach offers enhanced capabilities to detect workers at greater risk, guiding personalized interventions to mitigate adverse health effects associated with disrupted circadian rhythms. Support: The authors would like to thank ANID-FONDEF ID22I10053.

Publikationsverlauf

Artikel online veröffentlicht:
08. Oktober 2025