Search

Multisite machine-learning neuroimaging studies, such as those conducted by the ENIGMA Consortium, need to remove the differences between sites to avoid effects of the site (EoS) that may prevent or fraudulently help the creation of prediction models, leading to impoverished or inflated prediction accuracy. Unfortunately, we have shown earlier that current Methods Aiming to Remove the EoS (MAREoS, e.g., ComBat) cannot remove complex EoS (e.g., including interactions between regions). And complex EoS may bias the accuracy. To overcome this hurdle, groups worldwide are developing novel MAREoS. However, we cannot assess their effectiveness because EoS may either inflate or shrink the accuracy, and MAREoS may both remove the EoS and degrade the data. In this work, we propose a strategy to measure the effectiveness of a MAREoS in removing different types of EoS. FOR MAREOS DEVELOPERS, we provide two multisite MRI datasets with only simple true effects (i.e., detectable by most machine-learning algorithms) and two with only simple EoS (i.e., removable by most MAREoS). First, they should use these datasets to fit machine-learning algorithms after applying the MAREoS. Second, they should use the formulas we provide to calculate the relative accuracy change associated with the MAREoS in each dataset and derive an EoS-removal effectiveness statistic. We also offer similar datasets and formulas for complex true effects and EoS that include first-order interactions. FOR MACHINE-LEARNING RESEARCHERS, we provide an extendable benchmark website to show: a) the types of EoS they should remove for each given machine-learning algorithm and b) the effectiveness of each MAREoS for removing each type of EoS. Relevantly, a MAREoS only able to remove the simple EoS may suffice for simple machine-learning algorithms, whereas more complex algorithms need a MAREoS that can remove more complex EoS. For instance, ComBat removes all simple EoS as needed for predictions based on simple lasso algorithms, but it leaves residual complex EoS that may bias the predictions based on standard support vector machine algorithms., (Copyright © 2022. Published by Elsevier Inc.)

A common limitation of neuroimaging studies is their small sample sizes. To overcome this hurdle, the Enhancing Neuro Imaging Genetics through Meta-Analysis (ENIGMA) Consortium combines neuroimaging data from many institutions worldwide. However, this introduces heterogeneity due to different scanning devices and sequences. ENIGMA projects commonly address this heterogeneity with random-effects meta-analysis or mixed-effects mega-analysis. Here we tested whether the batch adjustment method, ComBat, can further reduce site-related heterogeneity and thus increase statistical power. We conducted random-effects meta-analyses, mixed-effects mega-analyses and ComBat mega-analyses to compare cortical thickness, surface area and subcortical volumes between 2897 individuals with a diagnosis of schizophrenia and 3141 healthy controls from 33 sites. Specifically, we compared the imaging data between individuals with schizophrenia and healthy controls, covarying for age and sex. The use of ComBat substantially increased the statistical significance of the findings as compared to random-effects meta-analyses. The findings were more similar when comparing ComBat with mixed-effects mega-analysis, although ComBat still slightly increased the statistical significance. ComBat also showed increased statistical power when we repeated the analyses with fewer sites. Results were nearly identical when we applied the ComBat harmonization separately for cortical thickness, cortical surface area and subcortical volumes. Therefore, we recommend applying the ComBat function to attenuate potential effects of site in ENIGMA projects and other multi-site structural imaging work. We provide easy-to-use functions in R that work even if imaging data are partially missing in some brain regions, and they can be trained with one data set and then applied to another (a requirement for some analyses such as machine learning)., Competing Interests: Declaration of competing interest AB has consulting fees from Biogen and lecture fees from Lundbeck, Otsuka and Janssen. AP has served as a consultant for Boehringer Ingelheim. AS: Advisory board (DSP), Research grants (CynK, DSP, MTPC, and Ono). CA has been a consultant to or has received honoraria or grants from Acadia, Angelini, Gedeon Richter, Janssen Cilag, Lundbeck, Otsuka, Roche, Sage, Servier, Shire, Schering Plough, Sumitomo Dainippon Pharma, Sunovion and Takeda. CH is faculty member, Lundbeck Psychiatric Institute. CP is on an advisory board for Lundbeck, Australia Pty Ltd and also received honoraria for talks presented at educational meetings organized by Lundbeck. CSW is on an advisory board for Lundbeck, Australia Pty Ltd and in collaboration with Astellas Pharma Inc., Japan. DJS has received research grants or consultancy honoraria from Lundbeck and Sun. EV has received grants and served as a consultant, advisor or CME speaker for the following entities (work unrelated to the topic of this manuscript): AB-Biotics, Abbott, Allergan, Angelini, Dainippon Sumitomo Pharma, Galenica, Janssen, Lundbeck, Novartis, Otsuka, Sage, Sanofi-Aventis, and Takeda. IN has no conflicts of interest to declare. RTS has received consulting fees from Genentech and Roche. SL has received consulting fees and speaking honoraria from Roche, Novartis, Biogen and Merck., (Copyright © 2020 The Author(s). Published by Elsevier Inc. All rights reserved.)

Spherical deconvolution methods are widely used to estimate the brain's white-matter fiber orientations from diffusion MRI data. In this study, eight spherical deconvolution algorithms were implemented and evaluated. These included two model selection techniques based on the extended Bayesian information criterion (i.e., best subset selection and the least absolute shrinkage and selection operator), iteratively reweighted l 2 - and l 1 -norm approaches to approximate the l 0 -norm, sparse Bayesian learning, Cauchy deconvolution, and two accelerated Richardson-Lucy algorithms. Results from our exhaustive evaluation show that there is no single optimal method for all different fiber configurations, suggesting that further studies should be conducted to find the optimal way of combining solutions from different methods. We found l 0 -norm regularization algorithms to resolve more accurately fiber crossings with small inter-fiber angles. However, in voxels with very dominant fibers, algorithms promoting more sparsity are less accurate in detecting smaller fibers. In most cases, the best algorithm to reconstruct fiber crossings with two fibers did not perform optimally in voxels with one or three fibers. Therefore, simplified validation systems as employed in a number of previous studies, where only two fibers with similar volume fractions were tested, should be avoided as they provide incomplete information. Future studies proposing new reconstruction methods based on high angular resolution diffusion imaging data should validate their results by considering, at least, voxels with one, two, and three fibers, as well as voxels with dominant fibers and different diffusion anisotropies., (Copyright © 2018 Elsevier Inc. All rights reserved.)

There is a widespread interest in applying pattern recognition methods to anatomical neuroimaging data, but so far, there has been relatively little investigation into how best to derive image features in order to make the most accurate predictions. In this work, a Gaussian Process machine learning approach was used for predicting age, gender and body mass index (BMI) of subjects in the IXI dataset, as well as age, gender and diagnostic status using the ABIDE and COBRE datasets. MRI data were segmented and aligned using SPM12, and a variety of feature representations were derived from this preprocessing. We compared classification and regression accuracy using the different sorts of features, and with various degrees of spatial smoothing. Results suggested that feature sets that did not ignore the implicit background tissue class, tended to result in better overall performance, whereas some of the most commonly used feature sets performed relatively poorly., (Copyright © 2018 The Authors. Published by Elsevier Inc. All rights reserved.)

Voxel-based morphometry (VBM) is a widely-used structural neuroimaging technique for comparing meso- and macroscopic regional brain volumes between patients and controls in vivo, but some of its steps, particularly the modulation, lack an experimental validation. The aims of this study were two-fold: a) to assess the effects of modulation to detect mesoscopic (i.e. between microscopic and macroscopic) abnormalities on published, classic VBM; and b) to suggest a set of potentially optimal settings for detecting mesoscopic abnormalities with new, advanced, high-resolution diffeomorphic VBM normalization algorithms. Sensitivity and false positive rate after modulating or not in classic VBM using different software packages and spatial statistics, and after setting a range of different parameters in advanced VBM (ANTS-SyN), were calculated in 10 VBM comparisons of 32 altered vs. 32 unaltered gray matter images from different healthy controls. Simulated brain abnormalities comprised mesoscopic volume differences mainly due to cortical thinning. In classic VBM, modulation was associated with a substantial decrease of the sensitivity to detect mesoscopic abnormalities (p<0.001). Optimal settings for advanced VBM included the omission of modulation, the use of large smoothing kernels, and the application of voxel-based or threshold-free cluster enhancement (TFCE) spatial statistics. The modulation-related decrease in sensitivity was due to an increase in variance, and it was more severe in higher-resolution normalization algorithms. Findings from this study suggest the use of unmodulated VBM to detect mesoscopic abnormalities such as cortical thinning., (© 2013.)

Wavelet-based methods have been developed for statistical analysis of functional MRI and PET data, where the wavelet transformation is employed as a tool for efficient signal representation. A number of studies using these approaches have reported better estimation capabilities, in terms of increased sensitivity and specificity, than the standard statistical analyses in the spatial domain. In line with these previous studies, the present report proposes a statistical analysis in the wavelet domain for the estimation of inter-group differences from structural MRI data. The procedure, called wavelet-based morphometry (WBM), was implemented under a voxel-based morphometry (VBM) style analysis. It was evaluated by comparing the gray-matter images of a group of 32 healthy subjects whose images were artificially altered to induce thinning of the cortex, with a different group of 32 healthy subjects whose images were unaltered. In order to quantify the performance of the reconstruction from a practical perspective, the same comparison was also conducted with standard VBM using SPM's Gaussian random fields and FSL's cluster-based statistics, family-wise error corrected, for datasets spatially-normalized via two different registration methods (i.e., SyN and FNIRT). The effect of using different amounts of smoothing, Battle-Lemarié filters and resolution levels in the wavelet transform was also investigated. Results support the proposed approach as a different and promising methodology to assess the structural morphometric differences between different populations of subjects., (Copyright © 2013 Elsevier Inc. All rights reserved.)

The Val158Met polymorphism in the COMT gene has been found to be associated with differences in brain activation in both healthy subjects and patients with schizophrenia. The predominant finding has been increased prefrontal activation associated with the Val allele; however, genotype-related de-activations have not been studied. In this study 42 schizophrenia patients and 31 controls underwent fMRI while performing the n-back task. Brain differences related to presence/absence of disease and presence/absence of the Val/Val genotype were examined. Both disease and Val/Val genotype were associated with failure of de-activation in a cluster centred in the medial prefrontal cortex. There was no interaction between disease and genotype at this location, but clusters where there were significant interactions emerged in the right prefrontal cortex and left temporal/parietal cortex. These findings suggest that Val158Met polymorphism influences task-related de-activations in the default mode network in both healthy subjects and schizophrenia patients to an equivalent extent. However the Val158Met polymorphism also has disease-specific effects on DLPFC activation in schizophrenia., (Copyright © 2010. Published by Elsevier Inc.)

Here we develop a measure of functional connectivity describing the degree of covariability between a brain region and the rest of the brain. This measure is based on previous formulas for the mutual information (MI) between clusters of regions in the frequency domain. Under the current scenario, the MI can be given as a simple monotonous function of the multiple coherence and it leads to an easy visual representation of connectivity patterns. Computationally efficient formulas, adequate for short time series, are presented and applied to functional magnetic resonance imaging (fMRI) data measured in subjects (N=34) performing a working memory task or being at rest. While resting state coherence in high (0.17-0.25 Hz) and middle (0.08-0.17 Hz) frequency intervals is bilaterally salient in several limbic and temporal areas including the insula, the amygdala, and the primary auditory cortex, low frequencies (<0.08 Hz) have greatest connectivity in frontal structures. Results from the comparison between resting and N-back conditions show enhanced low frequency coherence in many of the areas previously reported in standard fMRI activation studies of working memory, but task related reductions in high frequency connectivity are also found in regions of the default mode network. Finally, potentially confounding effects of head movement and regional volume on MI are identified and addressed.

Mutual information tools have been recently applied to quantify the connectivity between brain regions in functional magnetic resonance imaging (fMRI). Here we develop measures of mutual information between clusters of brain regions in the frequency domain. The properties and limitations of the method are exemplified through a single resting state fMRI dataset, and with a comparison involving frontostriatal connections in schizophrenic patients and healthy controls.