Limitations of the Derived Respiratory Variation Measurements Used In Functional Magnetic Resonance Imaging

Among different physiological sources of noise in blood oxygenation level-dependent functional magnetic resonance imaging (BOLD-fMRI), low-frequency fluctuation in arterial carbon dioxide (CO2) constitutes the strongest modulator of the BOLD signal. In this paper, the performance of respiration variation (RV) and respiratory volume per time (RVT) in identifying abnormal but prominent respiratory patterns are studied. We used Human Connectome Project in the Developmental dataset, as children are a challenging cohort in fMRI studies and have irregular breathing. According to our findings, there is no guarantee that a given respiratory event evident in the abdominal respiratory belt transducer timeseries, such as a deep breath or pause in breathing, will be detectable in both RVT and RV. In addition, RVT and RV do not show similar behavior during some respiratory events, especially when the subject breathes deeply at a low rate, while they use almost similar respiratory response functions.