0145 Breathing Pattern Estimated By Breathing Rate Variability Correlated With Amplitude Of Low Frequency Resting State MRI Fluctuations May Predict Sleep Quality

Abstract

Abstract Introduction We recently showed that people with obstructive sleep apnea (OSA) have higher breathing rate (BR) variability (BRV) in population (Pal et al. 2021)1. Simultaneously, Lynch et al. 20202 had assessed impaired breathing in the resting state healthy Human Connectome Dataset, and identified 21 people in each of the three groups of breathing patterns: eupnea (clean), sighs (deep breath) and suspected apnea breathing pattern (burst). During resting state functional MRI recordings, increased amplitude of low frequency fluctuations (ALFF) were found within areas related to hyperarousal such as the midbrain and bilateral extra-nucleus, whereas decreased ALFF were found within areas associated with memory and attention involving the parietal and occipital lobule and others; furthermore, the altered ALFF was associated with sleep efficiency (Ran et al. 20173). Could altered breathing pattern reflect an altered spontaneous neuronal activity state measured by fractional ALFF (fALFF) and associated psycho-physiological markers of sleep quality? Our goal was to characterize BRV and fALFF from the resting state HCP data in the three breathing pattern groups identified in Lynch et al. 2020 along with their psychophysiological states including sleep quality measured by Pittsburg Sleep Quality Index (PSQI). Methods In the three groups (n=21 X 3), clean group (6 males) age (mean±std. dev) 29±4 years, deep breath group (5 males) 29±4 years, burst group (14 males) 30±4 years - we calculated BR, absolute BRV (Interquartile range of BR) and relative BRV% from 15 minutes of resting state respiration belt data; and, fALFF from the minimally preprocessed resting state fMRI data filtered at 0.1-0.08 Hz and smoothed. We correlated fALFF with BR, BRV. Additionally, in the three groups, we report the mean±stdev of PSQI, BMI, systolic and diastolic blood pressure (BP), self-reported anxiety, attention problem, aggression scores along with their correlations with the absolute BRV. Results Absolute BRV was lower in deep breath group 3±3 breaths per minute (bpm) compared to clean (4±2 bpm) and burst (4±3 bpm) groups. BR was also lower in deep breath 14±7 bpm and correlated with BRV at Pearson R = 0.57 (p<0.05), compared to 18±2 bpm, R = -0.51 (p<0.05) in clean group and 18±3 bpm, R = 0.2 in burst group. In the deep breath group, the relative change in BRV 24±14%, correlated less with absolute BRV R = 0.78 (p<0.05) compared to the clean 24±13%, R = 0.98 (p<0.05) and burst groups 24±18%, R = 0.97 (p<0.05) indicating some voluntary sighs in the deep breath group as also validated by visual data inspection. The sleep quality (lower PSQI better sleep quality) was best in the clean 4±2 points, compared to both deep breath 5±3 and burst groups 5±3. Improved psychophysiological state in the clean and deep breath group compared to burst group was indicated by the systolic BP (121±12, 121±13 and 128±14 mmHg), BMI (25±5, 26±5, 28±6 kg/m2), anxiety (5±5, 5±5, 7±7)/attention problems (5±3, 5±4, 7±4)/aggression (3±2, 3±3, 5±2) scores in clean, deep breath and burst groups respectively. Only in the burst group, higher BRV correlated with higher BMI (R = 0.5, p<0.05). fALFF correlated with BRV (p<0.05, FWE corrected), not BR in all three groups at the cerebrospinal fluid (CSF) ventricles. With BRV as co-variate, burst group showed higher fALFF activity (p<0.001, uncorrected) compared to both clean and deep breath groups at the visual and somatosensory regions. Additionally, fALFF at the central executive network (CEN) was higher (p<0.001, uncorrected) for both clean and deep breath groups compared to burst. Interestingly, the clean group as well as the burst group had higher right somatosensory fALFF activity compared to deep breath group, that corresponded to the lower BRV in deep breath group. Conclusion Higher fALFF activity of burst compared to clean and deep breath groups in the visual and somatosensory regions were associated with sleep deprivation states (Dai et al., 20124; Wang et al., 20165 , Ran et al. 20173). Higher CEN fALFF activity indicating better sleep and physiological states (Zeighami et al., 20216, Ran et al. 20173) were found in the clean and deep breath groups compared to burst. Higher right somatosensory fALFF activity in the clean and burst groups compared to deep breath indicating higher breathing related movements in the groups having higher BR and BRV (clean and burst) compared to the deep breath group. The fALFF results are consistent with the indication of breathing coupled hemodynamic and CSF low-frequency oscillations that indicate sleep/wakefulness states during resting state (Fultz 20197). Overall, our study supports that BRV could be a potential indicator of psychophysiology, and taking sighs or deep breaths could potentially improve some psycho-neuro-physiological states but not necessarily sleep quality. As indicated earlier by Lynch et al. 20202, males report more burst breathing pattern while females report more deep breathing pattern indicating the ability to take deep breaths may counteract potential sleep disordered breathing problems. Support (If Any) HL135562