Relacs ("Relaxed ELectrophysiological data Acquisition, Control, and Stimulation", www. relacs.net) is a fully customizable software platform for data acquisition, online analysis, and stimulus generation specifically designed for electrophysiological recordings. Filters and spike detectors can be applied instantly on the recorded potentials. Freely programmable, hardware independent C++ plugins can access the preprocessed data for further online analysis and visualization. Therefore the experimental protocols can automatically adapt a stimulus (e.g. offset, variance, etc.) in a closed loop fashion and thus completely control the running experiment.
For offline data-analysis, data management, and data sharing the annotation of the raw data with metadata that specify the stimuli, the recorded neuron, the animal, as well as context of the experiment is necessary. Traditionally, such data have been written into lab books. File formats for metadata are often tailored to a specific database schema and thus assume a very detailed and specific structure and context. Within a lab this specificity is usually not a problem as long as the same database schema is used. However, these file formats for metadata are very difficult to handle, when it comes to sharing data within a larger community, in particular via the upcoming initiatives of public databases. In addition, in most cases, such metadata files do not even exist.
We here want to stress that the recording software used for acquiring the raw data knows already many of the metadata. In particular, these are properties of the generated stimuli and settings of the controlled hardware. Therefore metadata in principle can and should be immediately acquired during the acquisition of the raw data in order to minimize manual efforts in providing this important data. As a closed-loop software RELACS has all this information available and saves it to disk.
We also developed a file format for hierarchically organized key-value pairs that is independent of any specific database-schema and thus can be used as a general file format for exchanging metadata. Through customized terminologies the metadata can be structured and standardized for ensuring interoperability while at the same time not restricting the content, thus ensuring immediate flexibility. This way, metadata can be submitted to a local as well as public data-bases, like for example to the German neuroinformatics node at www.g-node.org, without any manual interference (see talk by Jan Grewe at symposium 10).
In our live demonstration we exemplify this approach by acquiring simulated data with RELACS and diretly importing the metadata to LabLog - the laboratory logbook, a datamanagement program for any neuroscientific lab (http://lablog.sourceforge.net/, see poster by Adrian Stoewer and Jan Grewe).