DIAGNOSIS STATUS GUIDED BRAIN IMAGING GENETICS VIA INTEGRATED REGRESSION AND SPARSE CANONICAL CORRELATION ANALYSIS

Area of Science:

• Neuroscience
• Genetics
• Biostatistics

Background:

• Brain imaging genetics links brain structure/function to genetic factors using quantitative traits (QTs).
• Regression and sparse canonical correlation analysis (SCCA) are common but have limitations in feature selection and utilizing diagnostic information.

Purpose of the Study:

• To develop a supervised sparse bi-multivariate learning model for enhanced feature selection in brain imaging genetics.
• To integrate regression and SCCA to leverage diagnostic information and improve the identification of genetic and imaging markers.

Main Methods:

• Proposed a novel supervised sparse bi-multivariate learning model combining regression and SCCA.
• Employed an efficient algorithm based on the alternative search method for feature selection.
• Validated the method on synthetic and real neuroimaging datasets.

Main Results:

• The proposed method demonstrated superior feature selection performance compared to traditional regression and SCCA.
• Successfully identified relevant genetic markers and imaging QTs.
• Showcased improved accuracy in analyzing brain imaging genetics data.

Conclusions:

• The novel supervised method offers a promising advancement for brain imaging genetics research.
• This bi-multivariate tool enhances the understanding of the genetic basis of brain structure, function, and abnormalities.
• The integrated approach effectively utilizes diagnostic information for more robust genetic and imaging marker discovery.