Exploration of scanning effects in multi-site structural MRI studies

Abstract

Background Pooling of multi-site MRI data is often necessary when a large cohort is desired. However, different scanning platforms can introduce systematic differences which confound true effects of interest. One may reduce multi-site bias by calibrating pivotal scanning parameters, or include them as covariates to improve the data integrity. New method In the present study we use a source-based morphometry (SBM) model to explore scanning effects in multi-site sMRI studies and develop a data-driven correction. Specifically, independent components are extracted from the data and investigated for associations with scanning parameters to assess the influence. The identified scanning-related components can be eliminated from the original data for correction. Results A small set of SBM components captured most of the variance associated with the scanning differences. In a dataset of 1460 healthy subjects, pronounced and independent scanning effects were observed in brainstem and thalamus, associated with magnetic field strength-inversion time and RF-receiving coil. A second study with 110 schizophrenia patients and 124 healthy controls demonstrated that scanning effects can be effectively corrected with the SBM approach. Comparison with existing method(s) Both SBM and GLM correction appeared to effectively eliminate the scanning effects. Meanwhile, the SBM-corrected data yielded a more significant patient versus control group difference and less questionable findings. Conclusions It is important to calibrate scanning settings and completely examine individual parameters for the control of confounding effects in multi-site sMRI studies. Both GLM and SBM correction can reduce scanning effects, though SBM’s data-driven nature provides additional flexibility and is better able to handle collinear effects.

Multisite reliability of MR-based functional connectivity

“…Others have investigated reliability of MRI across multiple sites in the domains of resting-state brain network overlap (Jann et al, 2015), anatomical measurements (Cannon et al, 2014; Chen et al, 2014), and task-related activations (Brown et al, 2011; Forsyth et al, 2014; Friedman et al, 2008; Gee et al, 2015). In general, test-retest reliability of functional connectivity is an ongoing field of study within both healthy and clinical populations (Keator et al, 2008; Orban et al, 2015; Van Essen et al, 2013; Zuo et al, 2014).…”

Hjärnan formas av psykologisk behandling

“…Another approach is to make data freely available to facilitate analyses with greater power as well as study replications (Nilsonne, 2014;Poldrack & Gorgolewski, 2014). The usefulness of this strategy is also supported by analytic approaches enabling correction for scanner effects in multi-site MRI studies (Chen et al, 2014;Friedman et al, 2008).…”

Combining Multi-Site/Multi-Study MRI Data: Linked-ICA Denoising for Removing Scanner and Site Variability from Multimodal MRI Data

“…Approaches to minimize the effects of multiple scanners/sites when combining MRI data have been proposed for each MRI modality, i.e., structural MRI data (Chen et al 2014;Fennema-Notestine et al, 2007;Keihaninejad et al, 2010;Pardoe et al, 2015;Takao et al, 2011), DTI data (Pohl et al 2016;Venkatraman er al., 2015;Fortin et al, 2017) and fMRI data (Feis et al, 2015;Glover et al, 2012;Griffanti et al, 2014;Salimi-Khorshidi et al, 2014). For statistical analyses, two techniques have emerged as the most common approaches to combining MRI data across studies/sites.…”

“…For statistical analyses, two techniques have emerged as the most common approaches to combining MRI data across studies/sites. The first, and perhaps most common approach, is the inclusion of site/study covariates in the higherlevel general linear model (GLM) implemented to assess group differences (Glover et al, 2012;Chen et al, 2014;Fennema-Notestine et al, 2007;Venkatraman et al, 2015).…”

“…ICA is a data-driven technique that attempts to decompose a multivariate signal into independent non-Gaussian signals (Bell and Sejnowski, 1995), and identify spatiotemporal patterns that reflect signals of interest versus processes that are related to motion and other artifacts. Chen et al (2014) applied ICA to multi-site structural MRI data and investigated the associations of the resulting spatiotemporal patterns with various scanning parameters to assess the influence of site on the data. Components that were related to scanner effects were then eliminated from the original data to "denoise" the data.…”