Forecasting the transition to sleep through HRV analysis: insights from ARIMA analysis and the concept of critical slowing down

ABSTRACT

Given the divergent findings regarding heart rate variability during drowsiness in studies, we propose utilizing the concept of critical slowing down (CSD) to detect early warning signals (EWS) of impending sleep onset based on heart rate variability as measured by standard deviation of the interbeat interval between normal peaks (SDNN). To expedite the detection of such EWS, we suggest employing ARIMA models. The study involved 25 healthy individuals (10 males) aged 20–35 years. Heart rate was recorded in natural home-based conditions from 19:40 to 06:00. The moment of sleep onset was recorded using the continuous button pressure paradigm. The results indicated that SDNN nonlinearly changed approaching sleep onset and exhibited spikes interpretable as EWS. SDNN was successfully forecasted using ARIMA analysis within a 10 minute window. An algorithm was developed to determine EWS signaling the impending transition to sleep based on ARIMA analysis of SDNN within the CSD concept. The algorithm allows for the detection of such EWS in 92% of cases.

KEYWORDS:

• Sleep onset
• critical slowing down
• early warning signal
• ARIMA
• heart rate

Acknowledgments

The authors would like to thank all the interns of the Cyberpsychology Lab who communicated with the participants and took part in data collection.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Russian Science Foundation under Grant No. 22-2820509.