|
|
|
Funding Sources
Title: Multivariate methods for identifying multi-task/multimodal brain imaging biomarkers
Funding Source: NIH/NIBIB: 1R01EB006841
PI: Calhoun
Abstract:
Each brain imaging modality reports on a different aspect of the brain (e.g. gray matter integrity, blood flow changes, electrical activity) and each has strengths and weaknesses. Combining multimodal imaging data is not easy since, among other reasons, each modality requires specialized expertise, and thus it is typical to analyze each imaging modality separately and interpret the results independently of one another. Many mental illnesses, such as schizophrenia, bipolar disorder, depression, and others, currently lack definitive biological markers and rely primarily on symptom assessments for diagnosis. One area which can benefit greatly from the combination of multimodal data is the study of schizophrenia. The brain imaging findings in schizophrenia are widespread and heterogeneous and have limited replicability. We show evidence that, in part, the lack of consistent findings is because most models do not combine imaging modalities in an integrated manner and miss important changes which are partially detected by each modality separately. We propose to develop multivariate methods based upon independent component analysis (ICA) to enable research on healthy versus diseased brain by identifying associations between different data types. The successful completion of this research will 1) provide a powerful set of tools [stand alone toolbox and database] for identifying relationships between multi-modal data, 2) provide a set of reliable brain imaging biomarkers for differentiating schizophrenia patients, bipolar patients (who share many symptoms with schizophrenia), and healthy controls, and 3) lay the groundwork for future work towards using imaging biomarkers for clinical purposes. In addition, the algorithms, model selection methods and anonymized data we provide will enable other investigators to use our tools and to compare their own methods with our own as well as to apply them to a large variety of brain disorders.
Title: Spatiotemporal Fusion of fMRI, EEG, and Genetic Data
Funding Source: NIH/NIBIB: R01EB005846
PI: Calhoun
Abstract:
The brain is an incredibly complex highly inter-connected organ. Each existing modality for imaging the living brain can only report upon a limited domain. For example, functional imaging provides information about dynamic blood flow changes in response to a stimulus, whereas electroencephalography (EEG) provides information about the electrical activity of the brain with centimeter spatial and millisecond temporal resolution. Finally, gene array imaging can assess specific differences at the chromosomal level that are present in individuals, some of which have functional consequences. Even though all three of these modalities can easily be collected on the same set of individuals, methods for effectively combining these different types of information are still in their infancy. All of these modalities typically involve thousands of data points per subject, and thus simple correlative approaches are of very limited nature for uncovering hidden patterns and associations in these data and can easily be computationally overwhelming. INTELLECTUAL MERIT: The goal of this project is to examine associations among fMRI, EEG, and genetic variations related to healthy and abnormal brain function. We propose to develop a set of tools based on independent component analysis (ICA) that can effectively fuse the information provided by multiple imaging modalities to span a vast range of spatial and temporal scales. The tools we develop will thus maximally exploit the information provided by multiple images and thus enable significant advances in the challenging problem of the study of brain function. We will develop these tools and apply to a large data set of healthy individuals in order to explore healthy brain function. In the last two years of the grant, we will apply our approaches to data collected from patients with schizophrenia as an example of impaired brain connectivity, and examine where the fMRl/EEG/genetic relationships have deviated from that of the healthy brain. We have recently introduced an ICA-based framework to jointly analyze data from multiple imaging sources and showed the richness of information conveyed by such a joint optimization approach. We have also introduced a number of approaches to effectively incorporate prior information into the ICA estimation in order to improve the performance. In the proposed study, we will bring these two research areas together, develop an ICA based fusion framework that enables the incorporation of prior information, specific for each data type, and will demonstrate the power of joint data analysis both between pairwise data types, and then for more than two data types for joint fusion. We will also extend the fusion ICA framework to one that incorporates prior information depending on each image type to improve the performance of joint information extraction. We will focus upon three image types, fMRI, EEG and genetic array imaging of single nucleotide polymorphisms. These three image types provide complementary information about brain function, and all can benefit from the incorporation of prior information. BROADER IMPACT: The broad impacts of the proposed work lie in its potential to substantially impact science and information technology. It also has the potential to impact public health by providing new information about healthy and abnormal brain which can then lead to new strategies for protecting and improving health. The study of human brain function is a very challenging and rich problem. The ICA-based fusion approach we believe is the key for achieving significant advances in the field. A significant broader impact of our proposal is to stimulate research at the interface between medical imaging and information processing by making the tools for the study of brain function widely available. We plan to develop a toolbox that enables fusion and analysis of various types of imaging data and allows the incorporation of prior information and make it available to the research community though a website.
Title: SEI: Independent Component Analysis of Complex-Valued Brain Imaging Data
Funding Source: NSF: 0715022
PI: Calhoun
Abstract:
Independent component analysis (ICA) has emerged as an attractive analysis tool for discovering hidden factors in observed data and has been successfully applied for data analysis in a wide array of applications such as biomedicine, communications, finance, and remote sensing. In a good number of these application domains, the data are typically complex valued. This is also the case in biomedical image analysis where ICA has been recognized as a promising tool for studying the brain function. Most biomedical image analysis techniques, however, use only the magnitude information and discard the phase, resulting in an unnecessary loss of information. Moreover, most brain imaging studies collect multiple data types where each existing modality for imaging the brain reports upon a limited domain and provides complementary information. Thus processing of imaging data in its native, complex form and by utilizing multiple modality images promises significant advances in our understanding of the brain function. We propose to develop a class of complex ICA algorithms, in particular for analysis of biomedical imaging data and demonstrate the power of joint data analysis as well as performing the analysis on the complete set of data, i.e., by utilizing both the magnitude and the phase information. We focus upon three image types, functional magnetic resonance imaging (fMRI), structural MRI (sMRI) and diffusion tensor imaging (DTI). These three imaging data provide complementary information about brain connectivity, and all can benefit from the incorporation of a complex-valued data processing approach. The broad impact of the proposed work lies in its potential to substantially impact science and information technology as well as in its educational features. Study of human brain connectivity is a very challenging and rich problem. The ICA-based fusion approach as well as the use of imaging data in its native, complex form, we believe is the key for achieving significant advances in the field. Successful demonstration of our approach for medical imaging data will also benefit other areas of science and technology where data from multiple sources and/or data in complex form need to be jointly analyzed for inferences. A significant broader impact of our proposal is to stimulate research at the interface between medical imaging and information processing by making the tools for the study of brain connectivity widely available through a toolbox and a medical imaging database.
Title: A unified framework for flexible brain image analysis (formally Multi-group Semi-blind ICA of fMRI)
Funding Source: NIH/NIBIB: 2R01 EB000840-06
PI: Calhoun
Abstract:
Data driven methods are being increasingly used to analyze brain imaging data. FMRI analyses can be put on an analytic spectrum with heavily model-based approaches (like the general linear model (GLM) implemented in the SPM software) on one end and flexible data-driven approaches like independent component analysis (ICA), principal component analysis (PCA), or clustering on the other end. In between there is a gap, which we and others have been trying to fill. In particular, methods such as ICA are particularly useful for reducing the multivariate fMRI problem down to one that is both tractable and also enables the incorporation of prior information. In the first period of this competing renewal, we focused our efforts upon developing ICA of fMRI methods which would be suitable for making group inferences, and which would allow the incorporation of prior information, hence moving from a 'blind' ICA approach to a semi-blind ICA approach. Despite the progress we have made, there is still considerable work to be done in the analysis of fMRI data with ICA. In this competing renewal, we propose to continue and significantly expand this work. First, we will extend our semi-blind ICA (sbICA) framework to provide a general framework for incorporating prior information from multiple spatial and temporal sources. In the second aim we will focus upon statistical inference and develop a framework for integrating the relevant functional components. In the third aim, we will validate the algorithms in aims 1 and 2, including using fMRI data collected on multiple days from a variety of paradigms. In this aim we develop a decision mechanism for selecting the best combination of methods given a particular problem. For the fourth aim, we will apply our methods to data collected during four well-studied paradigms in healthy controls and patients with schizophrenia. Our final aim involves the continuing development of our GIFT toolbox, and incorporation of the above algorithms, constraint selection mechanisms, and visual interfaces into the software. The successful completion of this research will provide a powerful set of tools for the research community to increase the sensitivity and specificity of BOLD analysis methods by drawing upon the strengths of both model-based and data-driven approaches. These tools will also provide a way to study the inter-relationship among functional networks in a flexible manner. This has application not only in schizophrenia but in many other diseases such as Alzheimer's, attention deficit hyperactivity, and psychopathy. |
|
|
|
|
|
|
Funding Sources