Non-negative matrix factorization of multimodal MRI, fMRI and phenotypic data reveals differential changes in default mode subnetworks in ADHD

Abstract

In the multimodal neuroimaging framework, data on a single subject are collected from inherently different sources such as functional MRI, structural MRI, behavioral and/or phenotypic information. The information each source provides is not independent; a subset of features from each modality maps to one or more common latent dimensions, which can be interpreted using generative models. These latent dimensions, or “topics,” provide a sparse summary of the generative process behind the features for each individual. Topic modeling, an unsupervised generative model, has been used to map seemingly disparate features to a common domain. We use Non-Negative Matrix Factorization (NMF) to infer the latent structure of multimodal ADHD data containing fMRI, MRI, phenotypic and behavioral measurements. We compare four different NMF algorithms and find the sparsest decomposition is also the most differentiating between ADHD and healthy patients. We identify dimensions that map to interpretable, recognizable dimensions such as motion, default mode network activity, and other such features of the input data. For example, structural and functional graph theory features related to default mode subnetworks clustered with the ADHD inattentive diagnosis. Structural measurements of the default mode network (DMN) regions such as the posterior cingulate, precuneus, and parahippocampal regions were all related to the ADHD-Inattentive diagnosis. Ventral DMN subnetworks may have more functional connections in ADHD-I, while dorsal DMN may have less. We also find that ADHD topics may be dependent upon diagnostic site, raising the possibility of the diagnostic differences across geographic locations. We assess our findings in light of the ADHD-200 classification competition, and contrast our unsupervised, nominated topics with previously published supervised learning methods. Finally, we demonstrate the validity of these latent variables as biomarkers by using them for classification of ADHD in 730 patients. Cumulatively, this manuscript addresses how multi-modal data in ADHD can be interpreted by latent dimensions.

Multivariate classification of autism spectrum disorder using frequency-specific resting-state functional connectivity—A multi-center study

“…Utilizing of LOOCV strategy could get stable weights of each feature and the weights got from the training dataset were more close to the whole dataset (Anderson et al, 2013). However the classification performance might be biased by overfitting (Hawkins, 2004).…”

“…On the expectation that the classification model and weights are closely resemblant in each trial, we applied leave-one-out validation (LOOCV) strategy to evaluate the performance of the classifier as only one of the 240 subjects were left out in each LOOCV trial (Anderson et al, 2013;Schölkopf and Smola, 2002). In brief, suppose there were n samples in total.…”

Fused Estimation of Sparse Connectivity Patterns From Rest fMRI—Application to Comparison of Children and Adult Brains

“…Independent component analysis (ICA) is an interesting data-driven tool that has been applied[15] to identify maximally independent spatial/temporal maps from BOLD time series. Non-negative matrix factorization (NMF) methods[16], [17] are increasingly used to extract multiple potentially overlapping brain networks by decomposing data matrices into linear combinations of non-negative basis functions. Other matrix decompositions methods such as Principal Component Analysis (PCA)[18] have been to used to extract intrinsic structure in the data by maximizing the amount of variance explained.…”

Intrinsic Functional Connectivity in Attention-Deficit/Hyperactivity Disorder: A Science in Development

“…Instead, the challenge provided an initial milestone of progress. Importantly, the ADHD-200 initiative has also supported numerous novel applications of analytic algorithms (46-57). As summarized elsewhere (45), neuroimaging is far from attaining psychiatric clinical utility, but initial progress is being made.…”