Abstract
Resting state functional magnetic resonance imaging (rs-fMRI) provides a powerful tool to examine large-scale neural networks in the human brain and their disturbances in neuropsychiatric disorders. Thanks to its low demand and high tolerance, resting state paradigms can be easily acquired from clinical population. However, due to the unconstrained nature, resting state paradigm is associated with excessive head movement and proneness to sleep. Consequently, the test-retest reliability of rs-fMRI measures is moderate at best, falling short of widespread use in the clinic. Here, we characterized the effect of sleep on the test-retest reliability of rs-fMRI. Using measures of heart rate variability (HRV) derived from simultaneous electrocardiogram (ECG) recording, we identified portions of fMRI data when subjects were more alert or sleepy, and examined their effects on the test-retest reliability of functional connectivity measures. When volumes of sleep were excluded, the reliability of rs-fMRI is significantly improved, and the improvement appears to be general across brain networks. The amount of improvement is robust with the removal of as much as 60% volumes of sleepiness. Therefore, test-retest reliability of rs-fMRI is affected by sleep and could be improved by excluding volumes of sleepiness as indexed by HRV. Our results suggest a novel and practical method to improve test-retest reliability of rs-fMRI measures.
Introduction
The resting-state fMRI (rs-fMRI) paradigm is widely used in neuroimaging because it requires minimal effort from participants, making it ideal for clinical populations. However, the lack of task constraints often leads to excessive head motion and an increased likelihood of participants falling asleep during the scan. While motion correction techniques are well-established, the impact of sleep on the reliability of functional connectivity measures has been largely overlooked. Sleep significantly alters brain connectivity, such as reducing thalamocortical interactions, which introduces non-biological variance and reduces the test-retest reliability—measured by the Intraclass Correlation Coefficient (ICC)—of rs-fMRI biomarkers.
Methods
The study utilized a test-retest dataset of 17 healthy adults who completed two scanning sessions spaced three months apart. Each session included an 8-minute resting-state scan with simultaneous electrocardiogram (ECG) recording. Heart Rate Variability (HRV), specifically the root mean square of successive differences (RMSSD), was calculated using a 16-second sliding window to serve as a continuous proxy for alertness.
Researchers compared the reliability of functional connectivity measures using the top 50% most "alert" volumes against the 50% "sleepiest" volumes. Reliability was assessed using ICC across the two sessions to determine how consistent brain network maps remained over time.
Results
Excluding "sleepy" volumes significantly enhanced the test-retest reliability of rs-fMRI measures. Specifically, scan-wise functional connectivity reliability increased by 33.4% compared to using the entire dataset. This improvement was consistent across all seven major brain networks, including the Default Mode Network (DMN), visual, and somatomotor systems.
Furthermore, the reliability of graph theory metrics, such as clustering coefficients and degree centrality, improved by approximately 30%. The results were robust; even when up to 60% of the most "sleepy" volumes were removed, the reliability remained higher than the baseline, indicating that data quality is more critical than total data volume. Importantly, sleepiness levels did not correlate with head motion, confirming that sleep is an independent confounding factor affecting reliability.
Discussion and Conclusion
This study demonstrates that sleep acts as a significant "nuisance" variable in resting-state research, similar to head motion. By using HRV as a practical, non-invasive proxy for alertness, researchers can "scrub" sleepy data points to significantly boost the reliability of brain connectivity measures. This approach is more feasible for clinical settings than EEG-based monitoring. Future work should investigate whether pulse oximetry, a standard feature in most MRI suites, can provide similar results. Ultimately, these findings suggest that accounting for sleep state is a crucial step toward making rs-fMRI a viable tool for clinical diagnostics and longitudinal monitoring of brain disorders.
Summary updated