Computational methods are critical for any functional MRI study, both for carrying out the scans themselves and for extracting results during the analysis. Optimizing the hardware and software available can effectively increase the functional sensitivity of fMRI studies by improving the ability to identify task-dependent signal components and by reducing the contribution of unwanted noise components. Such improvements in functional sensitivity can come from a variety of sources, including careful control of stimuli and monitoring of behavioral and physiological variables during data acquisition, careful modeling of the response dynamics of individual voxels, and post-processing tools for segregating out different temporal components of the data. Functional resolution can also be improved by analysis methods that are capable of making the most of multiple image data sets to accurately localize MR signal changes to specific areas or structures within the brain. In this Core we will attempt to maximize the functional resolution that can be obtained in the accompanying projects by providing state of the art hardware and software resources for fMRI experimental control and data analysis. These include run-time capabilities for accurate real-time control of any experimental paradigm and immediate real-time MR image analysis to check for scan quality and evaluate preliminary results. They also include an extensive repertoire of post-processing tools for testing temporal relationships between the MR signals and the paradigm tasks, for segmenting specific brain regions of interest, and for visualizing and quantifying functional activations with respect to location in a slice, region of interest, or on the cortical surface. Although all of these capabilities are available to some extent at the outset, this Core will undertake improvements in many of these areas with the specific goal of improving their ability to resolve functionally significant signal differences. A parallel goal of the Core will be to ensure that all of the computational hardware and software resources can be used efficiently and robustly, and can be combined with each other as needed to allow the project investigators to get the most out of their fMRI data.
Showing the most recent 10 out of 116 publications
