Four‐way multimodal fusion of 7 T imaging data using an mCCA+jICA model in first‐episode schizophrenia

Abstract

Acquisition of multimodal brain imaging data for the same subject has become more common leading to a growing interest in determining the intermodal relationships between imaging modalities to further elucidate the pathophysiology of schizophrenia. Multimodal data have previously been individually analyzed and subsequently integrated; however, these analysis techniques lack the ability to examine true modality inter-relationships. The utilization of a multiset canonical correlation and joint independent component analysis (mCCA + jICA) model for data fusion allows shared or distinct abnormalities between modalities to be examined. In this study, first-episode schizophrenia patients (n =19) and matched controls (n =21) completed a resting-state functional magnetic resonance imaging (fMRI) scan at 7 T. Grey matter (GM), white matter (WM), cerebrospinal fluid (CSF), and amplitude of low frequency fluctuation (ALFF) maps were used as features in a mCCA + jICA model. Results of the mCCA + jICA model indicated three joint group-discriminating components (GM-CSF, WM-ALFF, GM-ALFF) and two modality-unique group-discriminating components (GM, WM). The joint component findings are highlighted by GM basal ganglia, somatosensory, parietal lobe, and thalamus abnormalities associated with ventricular CSF volume; WM occipital and frontal lobe abnormalities associated with temporal lobe function; and GM frontal, temporal, parietal, and occipital lobe abnormalities associated with caudate function. These results support and extend major findings throughout the literature using independent single modality analyses. The multimodal fusion of 7 T data in this study provides a more comprehensive illustration of the relationships between underlying neuronal abnormalities associated with schizophrenia than examination of imaging data independently.

Patterns of co‐altered brain structure and function underlying neurological soft signs in schizophrenia spectrum disorders

“…GMV reduction in these regions, which are involved in important sensorimotor and cognitive functions, is consistent with previous MRI studies on SSD patients (Iwashiro et al, 2012;Kikinis et al, 2010;Koo et al, 2008;Shimizu et al, 2007;Tang et al, 2012;Torii et al, 2012;Zhou et al, 2005). Furthermore, the present study is in line with the recent mCCA + jICA study by Lottman and colleagues (Lottman et al, 2018), once again highlighting the relationship between GMV alterations and aberrant INA in SSD.…”

“…The procedure was similar to that described by and Lottman et al (2018). In a first step, two-tailed for multiple comparisons using the false discovery rate (FDR) method (Benjamini & Hochberg, 1995).…”

“…Still, one of the major benefits of using fALFF and combining both modalities is getting the information about both GMV and INA as well as increasing the power of data fusion analysis. There are also several recent fMRI studies that have successfully used fusion ICA methods with fALFF in healthy and SSD samples Hirjak, Rashidi, et al, 2019;Kubera et al, 2019;Lottman et al, 2018;Xu et al, 2015). Finally, GMV volume changes can cause spatiotemporal changes in INA and these joint network alterations might result in aberrant sensorimotor functioning and the development of NSS in SSD patients (Mittal, Bernard, & Northoff, 2017;Northoff & Duncan, 2016).…”

Fusion analysis of gray matter and white matter in subjective cognitive decline and mild cognitive impairment by multimodal CCA-joint ICA

“…In other word, the spatial maps of different modalities in one IC are correlated between modalities, while spatial maps of different ICs are independent with each other. A previous study conducted multimodal fusion analysis using mCCA-jICA and found results supporting and extending previous findings derived from single modality analyses, and also provided a more comprehensive understanding of the underlying relationship between functional and structural abnormalities in schizophrenia ( Lottman et al, 2018 ). Furthermore, these selected group-discriminating components might be useful for diagnosing schizophrenia ( Sui et al, 2013a , Yang et al, 2010 ).…”

Machine learning approaches for parsing comorbidity/heterogeneity in antisociality and substance use disorders: A primer

“…One popular approach for combining multimodal biomarkers of disease states is joint ICA (jICA), which concatenates two or more modalities into a single data stream (Sui et al, 2011). Lottman et al (2018), for instance, combined measures of gray matter, white matter, cerebrospinal fluid, and the amplitude of low-frequency fluctuations into a single jICA model, toward successful differentiation of first-episode schizophrenia patients and controls (Lottman et al, 2018). Similarly, Ouyang et al (2015) combined both gray matter volume and white matter functional anisotropy profiles into a single jICA model, toward the identification of multimodal features that distinguished Alzheimer's patients from controls.…”