Over the last two decades neuroimaging has made significant contributions to our understanding of human brain function. The indirect nature of the data requires sophisticated modeling and analysis approaches in order to infer interpretable quantities of interest. The 'Statistical Imaging Neuroscience' group at the DCCN researches analytical approaches for the analysis of functional neuroimaging data. Main research efforts involve:
- Computational methods for multimodal brain imaging using multivariate generative models
- Exploratory Data Analysis, specifically Independent Component Analysis, multi-way generalisations to bilinear linear models (Parallel Factor Analysis and Tensor-ICA)
- Bayesian time-series modeling, statistics and data fusion
- Statistics for Imaging Genetics
- Methods for pharmacological FMRI (phFMRI)
- Neural networks, data mining and statistical pattern recognition for neuroimaging data
- Biomarker development for clinical neuroscience
- Connectomics and resting-state functional connectivity
We participate in the development of the FMRIB Software Library (FSL), to provide powerful multimodal research tools for imaging neuroscientists, and advanced practical tools for clinicians using multimodal brain imaging in medicine.