1.1. Aims of the present study
Our overarching aim was to investigate how changes of the N250 component in ERPs to faces across repeated exposures are associated with face memory/recognition abilities on the individual differences level. More specifically, we distinguished between an easy versus a difficult test of face memory, following the distinction established by Wilhelm et al. (2010). Due to limit testing time, we measured face memory/recognition on a manifest rather than a latent level. ERPs, in particular the N250, were derived in a paradigm first employed by Tanaka and colleagues (2006). In this paradigm, a designated target face is presented, randomly mixed with several initially unfamiliar nontarget faces and the own face of the participant. As a novelty, the nontarget faces in the present study consisted of two sets of high- and low-distinctive unfamiliar faces, as determined by an independent sample of raters.
On the group level, we tested the N250 amplitudes for changes from the first to the second half of the EEG session, expecting to replicate the increase for target faces reported by Tanaka et al. (2006) and Sommer et al. (2021). We also tested this change across session halves in the N250 to the non-target faces and differentiated between high- and low-distinctive faces. On the individual differences level, we correlated the ERP differences between the first and second half of the session for the N250 to target faces, and to high- and low-distinctive non-targets faces with the performance estimates from the two memory tests. We expected more negative N250 amplitudes in the second half of the experiment to go along with better face memory/recognition. It was of special interest to study how this relationship depends on the target-status and the distinctiveness of the N250-eliciting stimuli and on the difficulty of the memory test.
2. Methods
The study consisted of a single session with two parts, a psychometric part, employing a test from the test battery of Hildebrandt et al. (2010) in a speed and an accuracy version and an EEG part modelled after the study of Tanaka et al. (2006). We always conducted the psychometric part first, followed by a break of 40 minutes before the EEG part started.