Respiration recording for fMRI: breathing belt versus spine coil sensor
Poster No:
2613
Submission Type:
Abstract Submission
Authors:
Marilena Wilding1, Anja Ischebeck1, Natalia Zaretskaya1
Institutions:
1Karl-Franzens-University, Graz, Styria
First Author:
Co-Author(s):
Introduction:
Physiological signals such as pulse and respiration contribute strongly to non-neuronal signal change of the blood oxygenation level-dependent (BOLD) contrast in fMRI. This has been observed not only during task-based but also during resting-state fMRI measurements. Both approaches are susceptible to effects of cardiac and respiratory signals, either through the introduction of spurious correlations or through possible systemic confounds due to time-locking of noise to the stimulus/task, respectively (Birn et al., 2009; Caballero-Gaudes & Reynolds, 2017; Murphy et al., 2013). Over the last decades, a variety of techniques evolved, aiming at detecting and removing physiological artifacts in fMRI time series. These follow either a data-driven or a reference-based approach, i.e., they rely on externally recorded physiological data. To record cardiac and respiratory signals, typically pulse oximetry or electrocardiography (ECG) and a respiration belt are used, respectively. New technologies allow to capture respiratory signal directly with a sensor placed within the spine coil in the patient table, eliminating the need of a respiration belt. However, little is known about the effectiveness of these new technologies and how they compare to the standard respiration belt recording.
Methods:
In the current study, we evaluated these two different respiratory activity recording devices in their effectiveness for eliminating physiological noise in task-based and resting-state fMRI. We recorded respiration using a breathing belt that is placed around a person's diaphragm and detects changes in abdominal circumference due to inhalation and expiration (belt). Simultaneously, respiratory activity was captured by a body coil sensor below the spine, embedded in the scanner table (spine), which generates a local magnetic field that varies with moving body tissue during breathing (Runge et al., 2019).
To investigate the performance of physiological noise correction based on either belt or spine in both data types, we recorded physiological signals with both devices during a perception task (n= 24; 12 female, mean age= 23.61) as well as during a resting-state session (n= 56; 35 female, mean age= 24.22). Physiological signals were processed using the RETROICOR (retrospective image correction) algorithm (Glover et al., 2000), where the periodic effects of pulsation- and respiration-related motion are modelled as a Fourier expansion of the cardiac and respiratory phase, using the PhysIO toolbox (Kasper et al., 2017). This ultimately resulted in 18 physiological regressors per device, which were then added to the different analyses. For the resting-state data, we compared ROI-to-ROI functional connectivity (FC) between denoised signals derived with the two devices. For the task data, we examined potential differences in stimulus-related activity between the denoised time series. In addition, the remaining noise after the application of the belt-derived or spine-derived physiological noise correction was compared in both datasets by analyzing the standard deviation (SD) of the general linear model (GLM) residuals.
To investigate the performance of physiological noise correction based on either belt or spine in both data types, we recorded physiological signals with both devices during a perception task (n= 24; 12 female, mean age= 23.61) as well as during a resting-state session (n= 56; 35 female, mean age= 24.22). Physiological signals were processed using the RETROICOR (retrospective image correction) algorithm (Glover et al., 2000), where the periodic effects of pulsation- and respiration-related motion are modelled as a Fourier expansion of the cardiac and respiratory phase, using the PhysIO toolbox (Kasper et al., 2017). This ultimately resulted in 18 physiological regressors per device, which were then added to the different analyses. For the resting-state data, we compared ROI-to-ROI functional connectivity (FC) between denoised signals derived with the two devices. For the task data, we examined potential differences in stimulus-related activity between the denoised time series. In addition, the remaining noise after the application of the belt-derived or spine-derived physiological noise correction was compared in both datasets by analyzing the standard deviation (SD) of the general linear model (GLM) residuals.
Results:
We observed no significant differences in ROI-to-ROI resting-state FC or stimulus-related activity between functional data denoised with either belt or spine regressors. However, we found that signal corrected with belt-derived regressors showed more residual noise than spine-corrected signal in task-induced activity (Fig. 1). Apart from this, belt recording contained artifacts like belt detachment and signal clipping, which were not present in the spine recording. Consequently, some subjects had to be removed due to technical issues during the belt recording.
Conclusions:
Our results suggest that body coil sensors are at least as well suited for physiological noise removal as the established breathing belt. It additionally offers an advantage in terms of signal quality, artifact proneness and acquisition comfort.
Modeling and Analysis Methods:
Activation (eg. BOLD task-fMRI)
Connectivity (eg. functional, effective, structural)
Exploratory Modeling and Artifact Removal
Task-Independent and Resting-State Analysis 2
Physiology, Metabolism and Neurotransmission :
Neurophysiology of Imaging Signals 1
Keywords:
FUNCTIONAL MRI
Other - physiological noise correction, respiratory recording, RETROICOR, functional connectivity, task-based fMRI, resting-state fMRI, reference-based physiology correction
1|2Indicates the priority used for review
Provide references using author date format
Birn, R.M., 2009. fMRI in the presence of task-correlated breathing variations. NeuroImage 47, 1092–1104. https://doi.org/10.1016/j.neuroimage.2009.05.030
Caballero-Gaudes, C., 2017. Methods for cleaning the BOLD fMRI signal. NeuroImage 154, 128–149. https://doi.org/10.1016/j.neuroimage.2016.12.018
Glover, G.H., 2000. Image-based method for retrospective correction of physiological motion effects in fMRI: RETROICOR. Magn. Reson. Med. 44, 162–167. https://doi.org/10.1002/1522-2594(200007)44:1<162::AID-MRM23>3.0.CO;2-E
Kasper, L., 2017. The PhysIO Toolbox for Modeling Physiological Noise in fMRI Data. J. Neurosci. Methods 276, 56–72. https://doi.org/10.1016/j.jneumeth.2016.10.019
Murphy, K., 2013. Resting-state FMRI confounds and cleanup. NeuroImage 349–359. https://doi.org/doi:10.1016/j.neuroimage.2013.04.001
Runge, V.M., 2019. Motion in Magnetic Resonance: New Paradigms for Improved Clinical Diagnosis. Invest. Radiol. 54, 383–395. https://doi.org/10.1097/RLI.0000000000000566