A practical introduction to EEG Time-Frequency Principal Components Analysis (TF-PCA)

This EEG methods tutorial provides both a conceptual and practical introduction to a promising data reduction approach for time-frequency representations of EEG data: Time-Frequency Principal Components Analysis (TFPCA). Briefly, the unique value of TF-PCA is that it provides a data-reduction approach that does not rely on strong a priori constraints regarding the specific timing or frequency boundaries for an effect of interest. Given that the time-frequency characteristics of various neurocognitive process are known to change across development, the TF-PCA approach is thus particularly well suited for the analysis of developmental TF data. This tutorial provides the background knowledge, theory, and practical information needed to allow individuals with basic EEG experience to begin applying the TF-PCA approach to their own data. Crucially, this tutorial article is accompanied by a companion GitHub repository that contains example code, data, and a step-by-step guide of how to perform TF-PCA: https://github.com/NDCLab/tfpca-tutorial. Although this tutorial is framed in terms of the utility of TF-PCA for developmental data, the theory, protocols and code covered in this tutorial article and companion GitHub repository can be applied more broadly across populations of interest.

Automated summary

This tutorial introduces a promising data reduction method, Time-Frequency Principal Components Analysis (TFPCA), for EEG data. TFPCA offers a data-reduction approach that does not rely on specific timing or frequency boundaries for an effect of interest. The tutorial provides background knowledge, theory, and practical information, and is accompanied by a GitHub repository with example code and a step-by-step guide for performing TFPCA.