NeuroRT Studio is a block diagram environment for the design of real-time experiments and analysis for neuroscience. It provides a graphical editor to drag and drop boxes from a library covering the essential steps of such analysis: data read and write (from/to a device and/or a file), temporal and spatial filtering, epoching, selection, mathematical operations, visualization, and many others. These processing steps can then be connected together with wires in order to create a signal processing pipeline.
The advanced signal processing toolbox offers cutting-edge signal processing techniques. The identification and the correction of artefacts in the data is available and runs in real-time for EMG, EOG, and electrode contact in a non-supervised fashion. In particular, we have implemented for real-time use : filter banks, blind sources separation methods, and tools for covariance matrices.
The advanced visualization toolbox allows illuminating display of brain activity in real-time. The available displays include 1D, 2D, and 3D, as well as the most famous xLORETA reconstruction in real-time.
Some experiments require high-density EEG (from 64 channels) needing major improvements beyond state-of-the-art techniques in order to run the processing and/or the visualization in real-time.
While the NeuroRT Studio is designed to analyze and display data in real-time, it is often desirable to batch process several files retrospectively. The Offline analysis toolbox was designed for this and implements essential machine learning tools including cross-validation framework and heuristics for parameters explorations. In order to compare prediction with clinically identified events, a review application enables signal annotation.