Real-time detection of event-related brain activity
|Real-time detection of event-related brain activity|
|Authors:||Gerwin Schalk, Eric C. Leuthardt, Peter Brunner, Jeffrey G. Ojemann, Lester A. Gerhardt, Jonathan R. Wolpaw|
|Citation:||NeuroImage 43 (2): 245-249. 2008 November|
|Database(s):||Google Scholar cites PubMed (PMID/18718544)|
|Web:||Bing Google Yahoo! — Google PDF|
|Article:||BASE Google Scholar PubMed|
|Restricted:||DTU Digital Library|
|Format:||BibTeX Template from PMID|
Real-time detection of event-related brain activity describes a systems called SIGFRIED (SIGnal modeling For Real-time Identification and Event Detection) for real-time analysis of intracranial electrocorticographic signals recorded from an array.
SIGFRIED is also described in Brain-computer interfaces (BCIs): detection instead of classification, but applied for EEG signals.
- bandpass filtered (0.1–220 Hz)
- sampled at 500 Hz
- windowed with 400 ms
- Frequency domain conversion with the maximum entropy method with model order of 25
- different frequency bands when inputed to SIGFRIED
- Subject instructions/conditions:
- Gaussian mixture model estimated with Competitive-EM modeling (for each time point t) the vector of frequency domain amplitudes based on the data collected during resting.
- Computation fo the likelihood that new data "was produced by the resting signal distribution"
- Visualization with SUMA from AFNI.
The developed a plugin SIGFRIED plugin for the BCI2000 system.