Functional MRIđź”—
Head motion is a serious issue in neuroimaging, especially for resting state fMRI, as it creates brain-wide artifactual effects such as inflated short-distance connectivity and attenuated long-distance connectivity (Power et al. 2012). Researchers employ a variety of strategies to mitigate head motion during acquisition and processing (Power et al. 2014; 2015; Satterthwaite et al. 2013; Siegel et al. 2017; Gratton et al. 2020). This includes motion censoring (discarding individual frames that are proximal to motion events within a run) and/or excluding entire runs due to high motion. These strategies may lead to the exclusion of some participants from further analysis due to lack of sufficient data.
Levels of head motion differ according to demographic factors such as sex, race/ethnicity, and SES (Cosgrove et al., 2022). Therefore, mitigation strategies for head motion introduce questions around fairness and differential exclusions across demographic groups. In addition, motion censoring causes sessions to vary by the amount of data remaining. Such variability may continue to inflate findings especially in the presence of conditions that may correlate with the motion artifact like autism or ADHD (Eggebrecht, 2017). The amount of data remaining influences the variation in the connectivity calculations by affecting the degrees of freedom. Therefore, even after motion censoring, issues concerning fairness may persist when examining factors that might be affected by motion like sex, race/ethnicity, SES, and BMI (Cosgrove et al., 2022). One strategy that avoids this confound is to strictly control the degrees of freedom, where functional connectivity measures are calculated with the exact same amount of data. Researchers should assess whether control of artifactual effects of head motion effects can be achieved by alternative means that mitigate this impact. Examples of such strategies could include data augmentation approaches such as sampling from other datasets, data processing strategies like the include use of ICA-based denoising (Pruim et al., 2015a; 2015b), use of bootstrap aggregation (Ramduny et al., 2024), or the creation of “pseudo-rest” by removing task signals from the task data (Fair et al. 2007), or post-hoc approaches like propensity weighting.
Researchers interested in examining brain-behavior associations or multivariate predictions should follow strategies such as those in Eggebrecht 2017 to: 1) assess how missing data impacts dependent, independent variables and covariates, 2) examine the association between the degrees of freedom and non-FC variables, 3) use trimmed FC measures when needed to mitigate artifacts due to data quality.
V02 data (0-1 month old) were processed in Infant fMRIPrep via two separate surface reconstruction workflows, M-CRIB-S (hash-0f306a2f) and Infant FreeSurfer (hash-2afa9081). Expert review and BrainSwipes QC consistently showed higher-quality surfaces from M-CRIB-S. This is expected, as M-CRIB-S uses T2w images, which are much higher contrast than the T1w (on which FreeSurfer relies) in neonates. Goncalves et al., 2025 report optimal performance for M-CRIB-S ≤5 months and Infant FreeSurfer ≥3 months. We therefore recommend only using data processed via the M-CRIB-S worfklow for analyses (nibabies-0f306a2f/xcp_d-0f306a2f+0ef9c88a).
Note: M-CRIB-S outputs can still be affected by poor T1w quality as it uses an externally generated brain segmentation from BIBSNet fed into Infant fMRIPrep as an external derivative. BIBSNet uses both T1w and T2w (when available), and low-quality T1w data may degrade segmentation quality.
A subset of Philips fMRI scans exhibit signal intensity clipping, where voxel intensities >4095 are capped due to a reconstruction scaling error. This produces hyperintense regions that can distort BOLD registration and impact downstream measures such as functional connectivity. The issue was identified during pilot data collection and fixed at most sites before the main study. Residual cases remain at VAN and CCH (patch implemented Oct 2024). Approximately ~20% of scans at these sites show some clipping, with ~6% classified as severe. Severe cases fail raw data QC, so are naturally filtered out from inclusion in downstream processing steps.
Users should practice caution in sensitive analyses and consider including clipping metrics as covariates. The presence and severity of clipping can be estimated based on raw data QC metrics. Potential clipping is indicated if: (1) the fraction of max intensity voxels in the brain mask (brain_fvox_max) is > 0.001 AND (2) the med/max image intensity ratio (brain_median/brain_max) is > 0.5 (> 0.8 indicates severe clipping).
Functional MRI (fMRI) measures brain activity via the blood oxygen level–dependent (BOLD) signal. The HBCD Study includes resting-state fMRI (rs-fMRI), with head motion monitored in real time using FIRMM to estimate usable data during acquisition as quantified by framewise displacement (FD) (Dosenbach et al., 2017).
A minimum of 2 runs are collected (during sleep for infants <30 months old), each lasting 7.5 minutes. A target of 7.5 minutes of usable low-motion data ((FD<0.3 mm)) is acquired across runs per session, with additional runs acquired as needed/possible to reach the target. Each run includes forward and reverse phase-encoding spin-echo EPI images for distortion correction. Data are acquired at 2 mm isotropic resolution with a repetition time (TR) of 1725 ms and multiband (MB) factor of 4.
Quality Control Summary Statisticsđź”—
We evaluated the impact of data quality on functional connectivity. Average functional connectivity matrices were computed using the Gordon-parcellated time series available in the V02 XCP-D derivatives. Data were included based on varying thresholds of BrainSwipes QC scores. Functional connectivity patterns were not substantially altered with the inclusion of lower-quality data, indicating robustness to mild quality variation
Connectivity matrices as data quality improves (left -> right) based on QC thresholds of 0.1, 0.5, and 0.9:
Cosgrove KT, McDermott TJ, White EJ, Mosconi MW, Thompson WK, Paulus MP, Cardenas-Iniguez C, Aupperle RL. Limits to the generalizability of resting-state functional magnetic resonance imaging studies of youth: An examination of ABCD Study® baseline data. Brain Imaging Behav 16, 1919-1925, 2022. doi: 10.1007/s11682-022-00665-2
Dean III, D. C., Tisdall, M. D., Wisnowski, J. L., Feczko, E., Gagoski, B., Alexander, A. L., ... & HBCD MRI Working Group. (2024). Quantifying brain development in the HEALthy Brain and Child Development (HBCD) Study: The magnetic resonance imaging and spectroscopy protocol. Developmental Cognitive Neuroscience, 70, 101452. 10.1016/j.dcn.2024.101452
Dosenbach, N. U. F., Koller, J. M., Earl, E. A., Miranda-Dominguez, O., Klein, R. L., Van, A. N., Snyder, A. Z., Nagel, B. J., Nigg, J. T., Nguyen, A. L., Wesevich, V., Greene, D. J., & Fair, D. A. (2017). Real-time motion analytics during brain MRI improve data quality and reduce costs. NeuroImage, 161, 80-93. https://doi.org/10.1016/j.neuroimage.2017.08.025
Eggebrecht, A. T., Elison, J. T., Feczko, E., Todorov, A., Wolff, J. J., Kandala, S., Adams, C. M., Snyder, A. Z., Lewis, J. D., Estes, A. M., Zwaigenbaum, L., Botteron, K. N., McKinstry, R. C., Constantino, J. N., Evans, A., Hazlett, H. C., Dager, S., Paterson, S. J., Schultz, R. T., … Pruett, J. R., Jr. (2017). Joint attention and brain functional connectivity in infants and toddlers. Cerebral Cortex (New York, N.Y.: 1991), 27(3), 1709–1720. doi: 10.1093/cercor/bhw403
Fair, D. A., Schlaggar, B. L., Cohen, A. L., Miezin, F. M., Dosenbach, N. U. F., Wenger, K. K., Fox, M. D., Snyder, A. Z., Raichle, M. E., & Petersen, S. E. (2007). A method for using blocked and event-related fMRI data to study “resting state” functional connectivity. NeuroImage, 35(1), 396–405. doi: 10.1016/j.neuroimage.2006.11.051
Gratton, C., Dworetsky, A., Coalson, R. S., Adeyemo, B., Laumann, T. O., Wig, G. S., Kong, T. S., Gratton, G., Fabiani, M., Barch, D. M., Tranel, D., Miranda-Dominguez, O., Fair, D. A., Dosenbach, N. U. F., Snyder, A. Z., Perlmutter, J. S., Petersen, S. E., & Campbell, M. C. (2020). Removal of high frequency contamination from motion estimates in single-band fMRI saves data without biasing functional connectivity. NeuroImage, 217(116866), 116866. doi: 10.1016/j.neuroimage.2020.116866
Power, J. D., Barnes, K. A., Snyder, A. Z., Schlaggar, B. L., & Petersen, S. E. (2012). Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. NeuroImage, 59(3), 2142–2154. doi: 10.1016/j.neuroimage.2011.10.018
Power, J. D., Mitra, A., Laumann, T. O., Snyder, A. Z., Schlaggar, B. L., & Petersen, S. E. (2014). Methods to detect, characterize, and remove motion artifact in resting state fMRI. NeuroImage, 84, 320–341. doi: 10.1016/j.neuroimage.2013.08.048
Power, J. D., Schlaggar, B. L., & Petersen, S. E. (2015). Recent progress and outstanding issues in motion correction in resting state fMRI. NeuroImage, 105, 536–551. doi: 10.1016/j.neuroimage.2014.10.044
Pruim RHR, Mennes M, van Rooij D, Llera A, Buitelaar JK, Beckmann CF. ICA-AROMA: A robust ICA-based strategy for removing motion artifacts from fMRI data. Neuroimage 112, 267-277, 2015a. doi: 10.1016/j.neuroimage.2015.02.064
Pruim RHR, Mennes M, Buitelaar JK, Beckmann CF. Evaluation of ICA-AROMA and alternative strategies for motion artifact removal in resting state fMRI. Neuroimage 112, 278-287, 2015b. doi: 10.1016/j.neuroimage.2015.02.063
Ramduny, J., Uddin, L. Q., Vanderwal, T., Feczko, E., Fair, D. A., Kelly, C., & Baskin-Sommers, A. (2024). Increasing the representation of minoritized youth for inclusive and reproducible brain-behavior associations. bioRxiv. doi: 10.1101/2024.06.22.600221
Siegel JS, Mitra A, Laumann TO, Seitzman BA, Raichle M, Corbetta M, Snyder AZ. Data Quality Influences Observed Links Between Functional Connectivity and Behavior. Cereb Cortex 27, 4492-4502, 2017. doi: 10.1093/cercor/bhw253