The electrophysiology of language production: what could be improved

Recently,thefieldofspoken-wordproduc-tion has seen an increasing interest in theuse of the electroencephalogram (EEG),mainlyforevent-relatedpotentials(ERPs).These are exciting times to be a languageproduction researcher. However, no mat-ter how much we would like our resultsto speak to our theories, they can onlydo so if our methods are formally cor-rect and valid, and reported in ways thatallow replicability. Inappropriate practicesin signal processing and statistical test-ing, when applied to our investigations,may render our conclusions invalid ornon-generalizable. Here, we first presentsome issues in signal processing and sta-tistical testing that we think deserve moreattention when analysing data, report-ing results, and making inferences. Theseissues are not new to electrophysiology,so our sole contribution is to reiteratethem in order to provide pointers to lit-erature where they have been discussedin more detail and solutions have beenproposed. We then discuss other issuespertinent to our investigations of overtword-production because of the effects(and potential confounds) that speakingwill have on the signal. Although we can-not provide answers to some of the issuesraised, we invite researchers in the fieldto jointly work on solutions so that thetopic of the electrophysiology of wordproduction can thrive on solid grounds.