Classification of schizophrenia from functional MRI using large-scale extended Granger causality

Abstract

The literature manifests that schizophrenia is associated with alterations in brain network connectivity. We investigate whether large-scale Extended Granger Causality (lsXGC) can capture such alterations using restingstate fMRI data. Our method utilizes dimension reduction combined with the augmentation of source time-series in a predictive time-series model for estimating directed causal relationships among fMRI time-series. The lsXGC is a multivariate approach since it identifies the relationship of the underlying dynamic system in the presence of all other time-series. Here lsXGC serves as a biomarker for classifying schizophrenia patients from typical controls using a subset of 62 subjects from the Centers of Biomedical Research Excellence (COBRE) data repository. We use brain connections estimated by lsXGC as features for classification. After feature extraction, we perform feature selection by Kendall's tau rank correlation coefficient followed by classification using a support vector machine. As a reference method, we compare our results with cross-correlation, typically used in the literature as a standard measure of functional connectivity. We cross-validate 100 different training/test (90%/10%) data split to obtain mean accuracy and a mean Area Under the receiver operating characteristic Curve (AUC) across all tested numbers of features for lsXGC. Our results demonstrate a mean accuracy range of [0.767, 0.940] and a mean AUC range of [0.861, 0.983] for lsXGC. The result of lsXGC is significantly higher than the results obtained with the cross-correlation, namely mean accuracy of [0.721, 0.751] and mean AUC of [0.744, 0.860]. Our results suggest the applicability of lsXGC as a potential biomarker for schizophrenia.

Investigation of large-scale extended Granger causality (lsXGC) on synthetic functional MRI data

“…This paper investigates a novel method, large-scale Extended Granger Causality (lsXGC), for discovering relations in high-dimensional dynamical systems involving observational data. The lsXGC method addresses the Ground truth Figure 2: Performance of various algorithms, namely mutual information (MI) [9], transfer entropy (TE) [8], multivariate Granger causality (GC) [16], and our proposed method (lsXGC, [2]) on the Netsim dataset [90]. As can be seen, the proposed algorithm (lsXGC) significantly outperforms other methods from the literature, suggesting robustness of the method against noise and hemodynamic response effects encountered in fMRI data.…”

“…For studying the underlying pathophysiology of neurological and psychiatric disease, functional connectivity among various cortical regions has been identified as an important research subject [1]. By providing images with sufficiently high spatial resolution and the hemodynamic response dynamics over a temporal axis, fMRI studies have demonstrated a tremendous potential to serve as a biomarker for neurologic and psychiatric disease [2][3][4][5]. Currently, most of the diagnosis in brain-related disorders relies on clinical symptom evaluations, such as based on neuropsychological testing.…”

“…Under the Gaussian assumption, transfer entropy is equivalent to Granger causality [11]. However, the computation of multivariate Granger causality for short time series in large-scale problems is challenging [2,[12][13][14][15]. Addressing this challenge, we recently proposed large-scale Extended Granger Causality (lsXGC), which is a method that combines the advantages of dimensionality reduction with the augmentation of a conditional source time-series adopted from the original space.…”

“…Addressing this challenge, we recently proposed large-scale Extended Granger Causality (lsXGC), which is a method that combines the advantages of dimensionality reduction with the augmentation of a conditional source time-series adopted from the original space. The lsXGC method uses predictive time-series modeling for estimating directed causal relationships among fMRI time-series [2], which makes it a suitable method to quantify the uncertainty associated with the stochastic process. This work extends our previous investigation on lsXGC and evaluates its ability to accurately recovering directional information from synthetic rs-fMRI data with known ground truth (GT).…”

“…Figure2: Performance of various algorithms, namely mutual information (MI)[9], transfer entropy (TE)[8], multivariate Granger causality (GC)[16], and our proposed method (lsXGC,[2]) on the Netsim dataset[90].Areas under the Receiver Operating Characteristics (AUROC) are shown in the figure, where each column represents a different algorithm (see titles). Boxplots represent [Q1, Q3]=[0.25, 0.75] quartiles and median, and the cap lines represent [minimum, maximum]=[Q1-1.5×(Q3-Q1), Q3+1.5×(Q3-Q1)].As can be seen, the proposed algorithm (lsXGC) significantly outperforms other methods from the literature, suggesting robustness of the method against noise and hemodynamic response effects encountered in fMRI data.…”