DSS Examples#
Overview#
Examples demonstrating Denoising Source Separation (DSS) across evoked, spectral, temporal, and blind-separation use cases.
Files#
plot_01_dss_fundamentals.py: Core DSS concepts with trial-average and bandpass biases.plot_02_artifact_correction.py: Blink and heartbeat correction with DSS.plot_03_evoked_responses.py: Evoked-response denoising and contrast-focused DSS.plot_04_spectral_dss.py: Frequency-specific component extraction on synthetic and real data.plot_05_periodic_dss.py: Periodic signal extraction for SSVEP and quasi-periodic structure.plot_06_temporal_dss.py: Time-shift and smoothness biases for temporally structured signals.plot_07_spectrogram_dss.py: Time-frequency masking with spectrogram-based DSS.plot_08_blind_source_separation.py: Blind source separation and FastICA equivalence.plot_09_custom_bias.py: Defining custom DSS biases.plot_10_benchmarking.py: Efficiency benchmarking against PCA, ICA, and averaging.plot_11_wiener_masking.py: Adaptive Wiener masking for bursty signals.plot_12_joint_dss.py: Joint DSS for multi-dataset repeatability.
Data Requirements#
Synthetic sections run directly with no external data.
Examples using MNE datasets download and cache them through MNE when needed.
References#
Särelä & Valpola (2005). Denoising Source Separation. J. Mach. Learn. Res.
de Cheveigné & Simon (2008). Denoising based on spatial filtering. J. Neurosci. Methods.
de Cheveigné & Parra (2014). Joint decorrelation. NeuroImage.
Efficiency Benchmark: DSS vs PCA, ICA, and Averaging.