Reproducibility study: Evaluation of PREPAIR and PESTICA using publicly shared data.
Poster No:
1539
Submission Type:
Abstract Submission
Authors:
Wanyong Shin1, Mark Lowe1
Institutions:
1The Cleveland Clinic, Cleveland, OH
First Author:
Co-Author:
Introduction:
Physiologic noise correction is the essential pre-process step of fMRI analysis. The simultaneously measuring of the respiratory pattern and cardiac pulsation signal during fMRI scan provides the physiologic noise signal modeling and its correction in fMRI dataset (Glover et al., 2000). However, the equipment and its setting during fMRI scan are not always feasible on the sites. Alternatively, the various methods to estimate the physiologic noise signal from fMRI dataset have been proposed (Beall & Lowe, 2007; Behzadi et al., 2007; De Martino et al., 2007; Perlbarg et al., 2007; Salimi-Khorshidi et al., 2014). Recently, the unsupervised physiological noise correction using phase and magnitude information (PREPAIR) has been proposed and its performance was compared to PESTICA (Bancelin et al., 2023). In this study, we reproduce the comparison of PREPAIR and PESTICA result with 3T datasets which was used PREPAIR study.
Methods:
Data & software: Forty 3T fMRI datasets with 4 different TR values (700, 1020, 1520 and 2000ms) were downloaded in the Harvard Dataverse (https://dataverse.harvard.edu/dataverse/prepair). The tested PREPAIR software was found at https://github.com/daveB1978/PREPAIR. The latest version of PESTICA (PESTICA_20240630) was found in https://github.com/wanyongshinccf/PESTICA and used in the analysis (Shin et al., 2022).
Analysis: Both PREPAIR and PESTICA estimate single cardiac and respiratory signals from fMRI dataset. Then PREPAIR uses the 2nd order harmonic Fourier expansion of each with total degree of freedom (DOF) = 8 while PESTICA employes the second order harmonic Fourier expansion of the cardiac signal and single respiratory signal as regressors with DOF=5. For the fair comparison, the original PESTICA is referred as PESTICA5, and the modified PESTICA to use the second order Fourier expansion of each as PESTICA8 (DOF=8).
Gray matter (GM) ROI were defined in EPI space using SPM with the probability > 0.9. The relative reduction of sum of squared root (relSoS) and tSNR in a GM ROI were calculated after PESTICA5, PESTICA8 and PREPAIR with the polynomial detrending based on the polynomial detrending only.
Analysis: Both PREPAIR and PESTICA estimate single cardiac and respiratory signals from fMRI dataset. Then PREPAIR uses the 2nd order harmonic Fourier expansion of each with total degree of freedom (DOF) = 8 while PESTICA employes the second order harmonic Fourier expansion of the cardiac signal and single respiratory signal as regressors with DOF=5. For the fair comparison, the original PESTICA is referred as PESTICA5, and the modified PESTICA to use the second order Fourier expansion of each as PESTICA8 (DOF=8).
Gray matter (GM) ROI were defined in EPI space using SPM with the probability > 0.9. The relative reduction of sum of squared root (relSoS) and tSNR in a GM ROI were calculated after PESTICA5, PESTICA8 and PREPAIR with the polynomial detrending based on the polynomial detrending only.
Results:
We do not find any significant difference of relSoS and tSNR in GM ROI between PESTICA8 and PREPAIR while the statistically different relSoS and tSNR values were observed between PESTICA5 and PREPAIR, indicating that the source of the difference is the discrepancy of DOF (See Fig1).
Figure 2 shows the example of the 4th slice in S1 TR=1520ms fMRI dataset. PREPAIR reduced the residual signal by the cardiac noise correction in the middle cerebral artery (MCA) while PESTICA8 does not (see the yellow arrow). When checking EPI image of PREPAIR data and the aligned PESTICA cardiac template, we observe the opposite directional EPI distortion in EPI image and PESTICA cardiac template because PREPAIR dataset was acquired with posterior to anterior (P to A) phase encoding (PE) direction while PESTICA template was generated with A to P PE direction. Note that the respiratory noise is corrected effectively in both PREPAIR and PESTICA.
Figure 2 shows the example of the 4th slice in S1 TR=1520ms fMRI dataset. PREPAIR reduced the residual signal by the cardiac noise correction in the middle cerebral artery (MCA) while PESTICA8 does not (see the yellow arrow). When checking EPI image of PREPAIR data and the aligned PESTICA cardiac template, we observe the opposite directional EPI distortion in EPI image and PESTICA cardiac template because PREPAIR dataset was acquired with posterior to anterior (P to A) phase encoding (PE) direction while PESTICA template was generated with A to P PE direction. Note that the respiratory noise is corrected effectively in both PREPAIR and PESTICA.
·Fig1. Averaged relSoS & tSNR in GM ROI after PESTICA5 (PES5), PESTICA8 (PES8) and PREPAIR (PREP) in blue & orange error bars, respectively (n=10 each). *: p < 0.01; ***: p < 0.001
·Fig2. The example of relSoS map after PESTICA8 (A) and PREPAIR (B). EPI image of PREPAIR dataset (C) and the aligned PESTICA (PES) cardiac mixing matrix template (D).
Conclusions:
We find that PREPAIR and PESTICA reduces the physiologic noise effectively and generate the similar range of tSNR in GM ROI when the identical DOF is used. However, PESTICA underperforms to detect the cardiac pulsatile signal in MCA regions as shown in Fig2. The source of this underperformance is the different PE direction of tested PREPAIR data and PESTICA template, resulting in the poor alignment of MCA. It is recommended that the different PE directional PESTICA templates are provided or EPI distortion correction step is applied prior to PESTICA pipeline.
If the phase image is available in addition to magnitude images in fMRI dataset and the pre-phase offset of RF coil using ASPIRE scan (Eckstein et al., 2018) is measured prior to fMRI study, PREPAIR could be an alternative tool to estimate/correct the physiologic noise when the external physiologic measures are not available.
If the phase image is available in addition to magnitude images in fMRI dataset and the pre-phase offset of RF coil using ASPIRE scan (Eckstein et al., 2018) is measured prior to fMRI study, PREPAIR could be an alternative tool to estimate/correct the physiologic noise when the external physiologic measures are not available.
Modeling and Analysis Methods:
Methods Development 1
Motion Correction and Preprocessing 2
Keywords:
Data analysis
FUNCTIONAL MRI
Open-Source Software
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Which processing packages did you use for your study?
Provide references using APA citation style.
Beall, E. B., & Lowe, M. J. (2007). Isolating physiologic noise sources with independently determined spatial measures. Neuroimage, 37(4), 1286-1300. https://doi.org/S1053-8119(07)00585-X [pii]
10.1016/j.neuroimage.2007.07.004
Behzadi, Y., Restom, K., Liau, J., & Liu, T. T. (2007). A component based noise correction method (CompCor) for BOLD and perfusion based fMRI. Neuroimage, 37(1), 90-101. https://doi.org/10.1016/j.neuroimage.2007.04.042
De Martino, F., Gentile, F., Esposito, F., Balsi, M., Di Salle, F., Goebel, R., & Formisano, E. (2007). Classification of fMRI independent components using IC-fingerprints and support vector machine classifiers. Neuroimage, 34(1), 177-194. https://doi.org/10.1016/j.neuroimage.2006.08.041
Eckstein, K., Dymerska, B., Bachrata, B., Bogner, W., Poljanc, K., Trattnig, S., & Robinson, S. D. (2018). Computationally Efficient Combination of Multi-channel Phase Data From Multi-echo Acquisitions (ASPIRE). Magn Reson Med, 79(6), 2996-3006. https://doi.org/10.1002/mrm.26963
Glover, G. H., Li, T. Q., & Ress, D. (2000). Image-based method for retrospective correction of physiological motion effects in fMRI: RETROICOR. Magn Reson Med, 44(1), 162-167. https://doi.org/10.1002/1522-2594(200007)44:1<162::AID-MRM23>3.0.CO;2-E [pii]
Perlbarg, V., Bellec, P., Anton, J. L., Pelegrini-Issac, M., Doyon, J., & Benali, H. (2007). CORSICA: correction of structured noise in fMRI by automatic identification of ICA components. Magn Reson Imaging, 25(1), 35-46. https://doi.org/S0730-725X(06)00342-0 [pii]
10.1016/j.mri.2006.09.042
Salimi-Khorshidi, G., Douaud, G., Beckmann, C. F., Glasser, M. F., Griffanti, L., & Smith, S. M. (2014). Automatic denoising of functional MRI data: combining independent component analysis and hierarchical fusion of classifiers. Neuroimage, 90, 449-468. https://doi.org/10.1016/j.neuroimage.2013.11.046
Shin, W., Koenig, K. A., & Lowe, M. J. (2022). A comprehensive investigation of physiologic noise modeling in resting state fMRI; time shifted cardiac noise in EPI and its removal without external physiologic signal measures. Neuroimage, 254, 119136.