Multi-Modal Imaging in Brain Research: The impact of Freesurfer and PETsurfer on Alzheimer’s studies’
Disciplines
Computer and Systems Architecture | Computer Engineering | Data Storage Systems | Other Computer Engineering
Abstract (300 words maximum)
Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder and the 6th leading cause of death in the United States. This project seeks to enhance AD staging prediction by integrating structural magnetic resonance imaging (MRI), genomics, cerebrospinal fluid (CSF) biomarkers, and electronic health records (EHR). Structural MRI is commonly used to observe brain changes associated with AD. Genetic studies have identified single nucleotide polymorphisms (SNPs) linked to brain structural alterations in AD patients. Abnormal levels of amyloid beta (A-beta) and tau proteins in CSF are also indicative of AD. EHR data categorizes individuals into cognitively normal (CN), mild cognitive impairment (MCI), and AD stages. However, SNPs cannot differentiate between MCI and AD, and MRI data shows similarities between early MCI and CN, as well as late MCI and AD. Thus, combining these four data modalities can improve the accuracy of AD staging. Multi-modal learning, which integrates multiple data sources, is expected to outperform single-modal approaches. However, aligning these heterogeneous data features remains a challenge, often leading to poor model fusion performance. This project uses the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and employs FreeSurfer and PETSurfer software to process MRI and positron emission tomography (PET) images for structural and functional analyses. FreeSurfer is used to extract anatomical information from MRI scans, including cortical thickness and volume, while PETSurfer processes PET images to quantify brain metabolism and protein deposition. These tools, along with genomic and CSF biomarkers, provide a comprehensive understanding of AD pathology. We anticipate that integrating these multi-modal data will significantly enhance the accuracy of AD staging predictions. This project will contribute to better diagnostic and prognostic tools, improving patient outcomes and advancing research on neurodegenerative diseases.
Academic department under which the project should be listed
CCSE - Computer Science
Primary Investigator (PI) Name
Chen Zhao
Multi-Modal Imaging in Brain Research: The impact of Freesurfer and PETsurfer on Alzheimer’s studies’
Alzheimer's disease (AD) is the most prevalent neurodegenerative disorder and the 6th leading cause of death in the United States. This project seeks to enhance AD staging prediction by integrating structural magnetic resonance imaging (MRI), genomics, cerebrospinal fluid (CSF) biomarkers, and electronic health records (EHR). Structural MRI is commonly used to observe brain changes associated with AD. Genetic studies have identified single nucleotide polymorphisms (SNPs) linked to brain structural alterations in AD patients. Abnormal levels of amyloid beta (A-beta) and tau proteins in CSF are also indicative of AD. EHR data categorizes individuals into cognitively normal (CN), mild cognitive impairment (MCI), and AD stages. However, SNPs cannot differentiate between MCI and AD, and MRI data shows similarities between early MCI and CN, as well as late MCI and AD. Thus, combining these four data modalities can improve the accuracy of AD staging. Multi-modal learning, which integrates multiple data sources, is expected to outperform single-modal approaches. However, aligning these heterogeneous data features remains a challenge, often leading to poor model fusion performance. This project uses the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and employs FreeSurfer and PETSurfer software to process MRI and positron emission tomography (PET) images for structural and functional analyses. FreeSurfer is used to extract anatomical information from MRI scans, including cortical thickness and volume, while PETSurfer processes PET images to quantify brain metabolism and protein deposition. These tools, along with genomic and CSF biomarkers, provide a comprehensive understanding of AD pathology. We anticipate that integrating these multi-modal data will significantly enhance the accuracy of AD staging predictions. This project will contribute to better diagnostic and prognostic tools, improving patient outcomes and advancing research on neurodegenerative diseases.