Multimodal MRI assessment for first episode psychosis: A major change in the thalamus and an efficient stratification of a subgroup

Abstract

Multi-institutional brain imaging studies have emerged to resolve conflicting results among individual studies. However, adjusting multiple variables at the technical and cohort levels is challenging. Therefore, it is important to explore approaches that provide meaningful results from relatively small samples at institutional levels. We studied 87 first episode psychosis (FEP) patients and 62 healthy subjects by combining supervised integrated factor analysis (SIFA) with a novel pipeline for automated structure-based analysis, an efficient and comprehensive method for dimensional data reduction that our group recently established. We integrated multiple MRI features (volume, DTI indices, resting state fMRI-rsfMRI) in the whole brain of each participant in an unbiased manner. The automated structure-based analysis showed widespread DTI abnormalities in FEP and rs-fMRI differences between FEP and healthy subjects mostly centered in thalamus. The combination of multiple modalities with SIFA was more efficient than the use of single modalities to stratify a subgroup of FEP (individuals with schizophrenia or schizoaffective disorder) that had more robust deficits from the overall FEP group. The information from multiple MRI modalities and analytical methods highlighted the thalamus as significantly abnormal in FEP. This study serves as a proof-of-concept for the potential of this methodology to reveal disease underpins and to stratify populations into more homogeneous sub-groups.

Keywords: DTI; factor analysis; first-episode psychosis; multimodal MRI; resting state fMRI; schizophrenia.

FIGURE 1

Schematic representation of the automated image parcellation using MRICloud (www.MRICloud.org). Each brain image is mapped to a set of multiple atlases and the pre‐defined labels are applied to each original brain. T1‐weighted images (for volumetric analysis) and DTI pipelines run in parallel. For the low‐resolution modalities (e.g., rs‐fMRI), the labels are brought to the original space by co‐registering the T1‐WIs. Through this process, the multiple MRI modalities are converted to a matrix of structures by image features, which represent each individual

FIGURE 2

Differences in imaging features between groups. Regions with abnormal DTI indices [FA (top row), MD (middle row)] and edges of abnormal rs‐fMRI synchrony (bottom row) in FEP (left column), S‐FEP (middle column), and M‐FEP (right column) compared with HC. Blue are lower mean values in FEP groups compared with controls; red are higher mean values in FEP compared with controls. Visualization with the BrainNet Viewer (http://www.nitrc.org/projects/bnv/, by Xia, Wang, & He, 2013)

FIGURE 3

Characterization of FEP group and subgroups (S‐FEP and M‐FEP), compared with controls, by the SIFA. Representation of the regional loadings of the common factors that show significant difference between groups (two in the all FEP vs. HC, one in the S‐FEP vs. HC, and one in the M‐FEP vs. HC), in a glass brain. The loading values are reported in Table 3. Visualization with the BrainNet Viewer (http://www.nitrc.org/projects/bnv/, by Xia et al., 2013)

FIGURE 4

Leave‐one‐out cross‐validated ROC curve for the classification of S‐FEP and controls. The logistic models were trained using the factors (both common and individual) estimated from the SIFA. The model including multimodality‐imaging features (volumes, FA, MD, and rs‐fMRI synchrony) was the most effective on correctly classifying individuals with S‐FEP, achieving an accuracy of 77% (Table 4)