Novel image analysis techniques for multicenter magnetic resonance imaging (MRI) study : neuroimaging measure of neural resource allocation and protocol-specific w-score standardization = 다기관 MRI 연구를 위한 새로운 이미징 분석기술 : 신경자원 할당에 대한 뉴로이미징 지표와 프로토콜별 w-score 표준화 방법
neuroimaging measure of neural resource allocation and protocol-specific w-score standardization
Nowadays, many neuroimaging studies of neurological or psychiatric disorders are conducted using magnetic resonance imaging (MRI) across multiple research centers. This approach makes it possible to acquire a large number of data in a short period of time. However, a major drawback of multi-center studies is data incompatibility caused by diverse task designs and imaging protocols at different centers. This disturbs multi-center studies to get consistent and reproducible results. Thus, I developed novel methods to improve data compatibility across centers. First, I proposed a novel neuroimaging measure of neural resource allocation for quantifying task-independent attention. It helps us to overcome the effect of task design on functional MRI data incompatibility for exploring attention in various cognitive tasks. The proposed measure consistently correlates with behavioral, physiological, and neural indicators of attention in various cognitively demanding tasks. And neural substrates of the measure also show a consistent spatial pattern, which exhibited spatial correspondence with attention networks, regardless of diverse task designs. These results indicate that the neuroimaging measure of neural resource allocation could be used as a quantitative indicator for attention to get consistent and reproducible results using functional MRI data from multi-center studies. Second, I developed protocol-specific w-score standardization to control the protocol effect on structural MRI data. W-score standardization creates quality reference values by using a healthy group to obtain normalized disease values. The protocolspecific w-score standardization applies it to each protocol separately to remove protocol-specific effects. This method reduced both univariate and multivariate differences in the structural MRI data while maintaining the disease effect, which ideally is not affected by the method. This result indicates that the data incompatibility of structural MRI is effectively resolved by the protocol-specific w-score standardization.