|Authors:||Ken Ball (ARL/UTSA), Kay Robbins (UTSA), W. David Hairston and Piotr Franaszczuk (ARL)|
|Software Language(s):||MATLAB 2014 or later|
|Software Type:||command line / EEGLAB plugin pop function|
What it does: BLASST (Band Limited Atomic Sampling With Spectral Tuning) uses a sophisticated algorithm based on matching pursuit with Gabor atoms to remove statistically anomalous spectral power in a narrow band of frequencies. The algorithm uses an iterative approach to set thresholds based on convergence criterion derived from surrounding frequencies.
Why is this important? Gamma frequency (30 Hz – 90 Hz) signals are biologically significant, but difficult to extract from EEG due to low-signal-to-noise and interference from external electrical signals. “Line noise” (50 Hz or 60 Hz) can be extremely non-stationary causing Fourier-based methods to either leave too much residual or to remove important parts of the signal.
K. Ball, W. D. Hairston, P. Franaszczuk, and K. Robbins (2017). BLASST: Band Limited Atomic Sampling with Spectral Tuning with applications to utility line noise filtering. IEEE Transaction on Biomedical Engineering, 64(9), 2276 – 2287, DOI:10.1109/TBME.2016.2632119, PMID: 27893379.
Recently the PREP pipeline has incorporated BLASST as an option for line noise removal in its implementation:
N. Bigdely-Shamlo, T. Mullen, C. Kothe, K.-M. Su, and K. Robbins (2015). The PREP pipeline: standardized preprocessing for large-scale EEG analysis, Frontiers in Neuroinformatics, 18 June 2015, PMID: 26150785, PMCID: 4471356, DOI: 10.3389/fninf.2015.00016.