This is the authors' abstract. We don't add key points for this paper.
Appropriate preprocessing is essential in electroencephalographic (EEG) research, particularly when neural activity near the 50/60-Hz power-line frequency is of interest. Although several algorithms are available for suppressing line noise, their evaluation has often emphasized power spectral density (PSD) flattening, with less attention to transient signal distortion in the time domain. Here, we compared seven approaches: narrow- and wide-band notch filtering, spectrum interpolation, CleanLine, band limited atomic sampling with spectral tuning (BLASST), ZapLine, and ZapLine-plus. In simulations, transient neural signal bursts at the line frequency were embedded in multichannel EEG-like data contaminated by nonstationary line noise. Root mean squared error was quantified separately during signal-burst, peri-burst, and no-signal periods to distinguish signal preservation, temporal smearing, and artifact rejection. Most algorithms showed a trade-off between signal preservation and artifact rejection. CleanLine best preserved sparse narrowband activity at the line frequency, whereas BLASST provided strong artifact rejection with intermediate signal preservation. ZapLine and ZapLine-plus avoided temporal smearing and preserved the neighboring spectrum but depended more strongly on the adequacy of spatial information. In empirical resting-state EEG recordings from 99 participants, BLASST produced the flattest PSD and the smallest median error at 10-s temporal resolution, with ZapLine and ZapLine-plus close behind. These results show that line-noise removal is not a single-objective denoising problem: performance depends on the time–frequency and spatial information available to distinguish neural activity from interference. Algorithm selection should therefore be guided by the characteristics of the target signal and the analysis goal rather than by PSD flattening alone.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Maruyama et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: