This chapter provides a tutorial style guide to analyzing electroencephalogram (EEG) data contingent on feature-based attentional selection. It is targeted at researchers that currently investigate attentional processes using univariate methods but consider moving to multivariate analyses. The chapter starts by providing examples of classical univariate analysis, in which the EEG signal occurring ipsilateral to the target is subtracted from the signal that occurs in a contralateral electrode (i.e., the classical N2pc, an interhemispheric posterior negativity emerging around 180–200 ms). Next, it shows how the same type of information can also be identified using multivariate pattern analysis (MVPA). MVPA does not restrict one to contrast attentional selection in opposite hemifields but also allows one to assess attentional selection on the vertical meridian, or even within a quadrant of the visual field, opening up new avenues for research. The chapter demonstrates how to visualize topographic maps of attentional selection when using MVPA and shows how to assess timing onsets using the percent-amplitude latency method. Finally, it shows how a forward encoding model enables one to characterize the relationship between a continuous experimental variable (such as attended targets positioned on a circle) and EEG activity. This allows one to construct brain patterns for positions in the visual field that were never attended in the data that was used to create the forward model. This chapter is intended as a practical guide, explaining the methods and providing the scripts that can be used to generate the figures in-line, thus providing a step-by-step cookbook for analyzing neural time series data in the field of feature-based attentional selection.
Feature-based attention Attentional selection EEG Univariate analysis N2pc MVPA Multivariate pattern analysis Classification Decoding BDM Forward encoding model Inverted encoding model FEM
This is a preview of subscription content, log in to check access.
Springer Nature is developing a new tool to find and evaluate Protocols. Learn more
I would like to thank Anna Grubert, Martin Eimer, and Chris Olivers for allowing me to freely use and share the data from these experiments, as well as the analysis plans that we used on these data. This chapter would not have existed without them.