Neuroimaging data is being increasingly used for investigating the neurobiological underpinnings of complex disorders ranging from brain lesions due to glioma or multiple sclerosis, to mental disorders such as Alzheimer's disease or posttraumatic disorder. However, such disorders are complex, and information based on one neuroimaging modality or a single experiment, or single visit, may not be sufficient to characterize the etiology of the disease. Hence, it is of paramount importance to combine data from multiple imaging modalities, or experimental sessions, and/or longitudinal visits, to accurately map the trajectory of the disorder and to discover the underlying neurobiological causes. Unfortunately, there are limited methods and computational tools available for the integrative analysis of neuroimaging data that are systematically able to combine information across multiple modalities, experiments or longitudinal visits, so as to extricate underlying complex patterns that just would not be revealed under a standard neuroimaging analysis.
The goal of this Research Topic is to develop statistical and machine learning methods for integrative analysis of multi-modal, and/or multi-task, and/or longitudinal neuroimaging data. Bayesian as well as penalized approaches that are able to develop novel methodology or computational methods that are able to systematically borrow information from modalities/experiments/longitudinal visits in an unsupervised manner are desirable. These approaches must account for heterogeneity across datasets, and where needed, within datasets. The integrative analysis may include other data types such as -omics data that are combined with neuroimaging data so as to result in an imaging genetics analysis; however, the imaging data must contain more than one modality, must come from more than one experiment or should correspond to multiple longitudinal visits. The analysis may involve data from single individuals or multiple individuals. Other types of derived imaging data such as brain connectome are also welcome as potential themes for analysis.
Some potential themes for this Topic include, but are not limited to, the following:
1. Joint modeling of neuroimaging data coming from task and rest experiments
2. Approaches for incorporating structural brain information for analysis of brain function (e.g brain functional connectivity)
3. Approaches for joint analysis of neuroimaging data coming from multiple related disease subgroups
4. Prediction and/or classification approaches that incorporate data from multiple modalities/experiments/longitudinal visits, including data coming from different scanners and experimental sites
5. Approaches for mapping the trajectory of a disorder using longitudinal neuroimaging data
We would note that authors are welcome to combine additional data such as -omics data with the above analysis themes, although this is not a strict requirement.
Neuroimaging data is being increasingly used for investigating the neurobiological underpinnings of complex disorders ranging from brain lesions due to glioma or multiple sclerosis, to mental disorders such as Alzheimer's disease or posttraumatic disorder. However, such disorders are complex, and information based on one neuroimaging modality or a single experiment, or single visit, may not be sufficient to characterize the etiology of the disease. Hence, it is of paramount importance to combine data from multiple imaging modalities, or experimental sessions, and/or longitudinal visits, to accurately map the trajectory of the disorder and to discover the underlying neurobiological causes. Unfortunately, there are limited methods and computational tools available for the integrative analysis of neuroimaging data that are systematically able to combine information across multiple modalities, experiments or longitudinal visits, so as to extricate underlying complex patterns that just would not be revealed under a standard neuroimaging analysis.
The goal of this Research Topic is to develop statistical and machine learning methods for integrative analysis of multi-modal, and/or multi-task, and/or longitudinal neuroimaging data. Bayesian as well as penalized approaches that are able to develop novel methodology or computational methods that are able to systematically borrow information from modalities/experiments/longitudinal visits in an unsupervised manner are desirable. These approaches must account for heterogeneity across datasets, and where needed, within datasets. The integrative analysis may include other data types such as -omics data that are combined with neuroimaging data so as to result in an imaging genetics analysis; however, the imaging data must contain more than one modality, must come from more than one experiment or should correspond to multiple longitudinal visits. The analysis may involve data from single individuals or multiple individuals. Other types of derived imaging data such as brain connectome are also welcome as potential themes for analysis.
Some potential themes for this Topic include, but are not limited to, the following:
1. Joint modeling of neuroimaging data coming from task and rest experiments
2. Approaches for incorporating structural brain information for analysis of brain function (e.g brain functional connectivity)
3. Approaches for joint analysis of neuroimaging data coming from multiple related disease subgroups
4. Prediction and/or classification approaches that incorporate data from multiple modalities/experiments/longitudinal visits, including data coming from different scanners and experimental sites
5. Approaches for mapping the trajectory of a disorder using longitudinal neuroimaging data
We would note that authors are welcome to combine additional data such as -omics data with the above analysis themes, although this is not a strict requirement.