The Competence Center Biomedical Data Science works with diverse biomedical data – big data, small data and traces. In some cases, the center carries out data analyses on behalf of its partners from medicine, psychology and industry. However, most data are acquired in its own experimental or clinical studies using protocols and paradigms developed in projects. The different types of data are described below. Researchers can receive a trial-set on request. Please write to firstname.lastname@example.org.
FUNCTIONAL MAGNETIC RESONANCE IMAGING
The functional magnetic resonance imaging (fMRI) is a neuroimaging technique that is based on the key assumption that neural activity is coupled with increased blood flow. As the name suggest, fMRI is based on MRI technology. To be more precise, it uses the blood-oxygen-level dependent (BOLD) contrast to determine which brain areas show increased neural activity during a certain task.
The center acquires anatomical T1 weighted MRI scans with an isotropic voxel size of 1 mm of the brain. This data is well suited for anatomical segmentation of the brain into anatomical regions as well as high quality extraction of gray matter and white matter surfaces. Given these surfaces, individual EEG source localization and many other analyses can be augmented.
To investigate white matter structure, diffusion-weighted MRI with 64 non-collinear gradient directions, a single high b-value of 1000 s/mm and an isotropic voxel size of 2 mm is acquired. Such data provides insight into the anisotropic structure of the brain and the structural neuronal wiring.
For accurate structural connectivity analyses, high angular resolution diffusion-weighted MRI (HARDI) is acquired. The data acquisition protocol includes 256 non-collinear gradient directions, a single high b-value of 1500 s/mm and an isotropic voxel size of 2 mm. This makes conclusions about the anisotropic structure of the brain even more accurate enabling deeper insight into nerve fiber bundles especially in case of crossing fiber pathways.
Mobile sensor devices become more and more important in research and medical applications. Even consumer devices are utilized. They measure, for instance, acceleration, rotational speed, heart rate, electrodermal activity or skin temperature. Under certain circumstances (for instance, during rest, sleep or mild physical activity) the quality of the mobile data is comparable to the quality of data recorded in the lab with laboratory equipment. In some other cases, however, special preprocessing techniques need to be applied.