EEG is better when cleaning effectively targets artifacts

OBJECTIVE: Electroencephalography (EEG) data are contaminated by a range of non-neural artifacts. The confounding influence of artifacts is often addressed by using independent component analysis (ICA) to decompose data into components, subtracting artifactual components, then reconstructing data into the electrode space. Due to imperfect component separation, this common approach can remove neural signals as well as artifacts. Here, we demonstrate the counterintuitive finding that this can artificially inflate event-related potential and connectivity effect sizes and bias source localisation estimates, while also removing neural signals. METHODS: We developed a novel method that targets cleaning to artifact periods of eye movement components and artifact frequencies of muscle components, and tested our method across different EEG systems and cognitive tasks. RESULTS: Our targeted artifact reduction method was effective in cleaning artifacts while also reducing the artificial inflation of effect sizes and minimizing source localisation biases. CONCLUSIONS: EEG pre-processing of Go/No-go and N400 task data is better when targeted cleaning is applied, which better preserves neural signals and mitigates effect size inflation and source localisation biases that result from subtracting artifact components. SIGNIFICANCE: These improvements enhance the reliability and validity of EEG analyses. Our method is provided in the RELAX pipeline, which is freely available as an EEGLAB plugin (https://github.com/NeilwBailey/RELAX).