The fundamental concept of this platform is that data are first transformed into a standard layout using common file formats, so that all processing pipelines can expect a deterministic, well described data structure. This setup enables the development of data processing components that are agnostic of the peculiarities of individual studies, as long as all relevant aspects of a study are properly described. The necessary data conversion and layout are performed largely automatic. The procedures are tuned for data acquired in Magdeburg, but with minor extra effort data from other sources can be processed too.
For each study the following steps will be performed:
- Initial study setup: create a study raw dataset
- Import a new acquisition (repeated as necessary):
- Complete study description (if necessary):
- Identify data that were acquired simultaneously during an acquisition
- Associate stimulation and behavior logs with (MR) data acquisitions
- Identify data formats/converters for non-standard acquisitions (e.g. custom hardware)
- possibly use datalad-hirni's webUI to edit the specification accordingly
- Convert raw data into the common data structure
- Verify result of automatic conversion (automatic validation tools are available)
- Apply analysis procedure
Detailed step-by-step demos are included in many section to help getting started with this platform.