Cognitive Neuroscience Talks

ERP correlates of person recognition in repetition priming studies

Objectives: The cognitive processes involved in person recognition are a focus of debate. For faces and names, Experiment 1 investigated differences in neural processing between primed and unprimed recognized (famous) and unfamiliar stimuli. Design and Methods: Ninety famous faces and 90 famous names were identified in a recognizability study, and the faces and names of unfamiliar people were obtained, matched to the famous ones for visual stimulus properties. Participants (n = 18) performed speeded familiarity decisions (two-choice key presses) for faces or names that were presented in separate blocks for 2000 ms each. In two priming blocks, participants viewed 120 faces and 120 names (60 famous and 60 unfamiliar in each block). Stimulus prime-test interval was approximately 20 min. At test, participants viewed 180 faces and 180 names. One third of stimuli were shown for the first time, one third were repeated using the same stimulus, and one third were repeated using a different image (faces) or font type (names). EEG was recorded from 32 channels and recalculated off-line to average reference. Results: Priming significantly reduced reaction times (RTs) for famous but not for unfamiliar stimuli. No RT differences were found for stimuli primed by repetitions using a different image or font versus those primed by the same stimuli. ERP amplitudes were analysed for time segments 100-140 ms, 160-200 ms, 200-300 ms, 300-400 ms, 500-600 ms, 600-700 ms, 700-800 ms and 800-1200 ms after stimulus onset. For both faces and names, clear familiarity effects were seen from 300 ms onwards. Significant differences related to priming were observed for famous (but not unfamiliar) names and faces at 500-600 ms. These ERP modulations were typified by increased positivity and earlier peak latency compared to unprimed stimuli. Differences were domain-independent (i.e. comparable for faces and names), and located over parietocentral sites. Conclusions: Domain-independent ERP modulations, which occur relatively late compared to perceptual processing components, may reflect the facilitation of person-specific memory retrieval that is observed behaviourally in long-term repetition priming. Preliminary findings for Experiments 2 and 3 are also discussed. Experiments 2 and 3 investigated priming for famous faces and names, respectively. In these experiments, repetition was immediate (1000 ms between prime and target), however the RT and ERP effects of within-domain priming (e.g. face prime to face target) and between-domain priming (e.g. name prime to face target) were compared. As with Experiment 1, within-domain priming could either be repetition of a stimulus with the same visual properties (primed same) or the presentation of a different portrait/different font at target (primed different). Initial comparisons of ERPs related to within- and between-domain priming show both early differences and later similarities in waveform amplitude and latency. Early differences are proposed to reflect differences in visual processing for face and name stimuli, whereas it is argued that later similarities reflect a modality-independent process of memory retrieval that is shared for both face and name stimuli.