In the analysis event-related potentials (ERPs), once the data has been pre-processed, epoched and cleaned the next step is to extract the grand-average at both the subject and group levels for each experimental condition in the hope of revealing condition-dependent differences between the grand average data.
Very often, our analysis of ERPs is guided by prior related research which we can use to define the relevant ERP activity, electrodes of interest, time-windows of interest and activity latencies. Given a sufficient amount of extant research on our subject of interest, we can define our ERPs of interest, our electrodes of interest and our critical time-windows a priori and carry out our statistical analysis (e.g. ANOVA) on this data. This means of ERP analysis has been very widely applied.
However, if our study is novel or departs from the approach used in extant research or, if our grand average results reveal a pattern of results that does not correspond with expectations on the basis of prior research, then we need to apply more exploratory data-driven methods to define our electrodes of interest, our time-windows of interest and the condition dependent activity (ERPs).
This page provides links to documents and pages describing various methods and tools that facilitate an exploration of grand average ERP data in a data-driven manner.