Person. indicid. DiA: Vol. Il. No. 7, pp. 741-749. 1990 Printed in Great Britain. All rights reserved 0191-X869/90 $3.00+0.00 Copyright (1 1990 Pergamon Press plc ANXIETY AND INDIVIDUAL DIFFERENCES IN EVENT-RELATED POTENTIALS DURING THE RECOGNITION OF SENSE AND NONSENSE WORDS VILFREDO DE PASCALIS and ALESSANDRA MORELLI Dipartimento di Psicologia. University of Rome. Via degli Apuli 8. 00185 Rorna, Italy zyxwvutsrqponmlkjihgfe (Received 27 October 1989) Summary-Event-related potentials (ERPs) to binaurally presented tone pips were elicited while male and female subjects with high and low level of anxiety were engaged in sense and nonsense word recognition tasks. ERPs were recorded from left and right frontal and temporo-parieto-occipital scalp locations. Peak latencies and amplitudes of N2 and P3 components were obtained for each subject during sense and nonsense word recognition tasks. High anxiety subjects exhibited a smaller N2 peak amplitude as compared with the lows. These subjects also showed smaller N2 and P3 peak amplitudes on the left hemisphere with respect to the right in the recognition of sense words and no hemisphere differences in the recognition of nonsense ones. Low subjects, in contrast, did not exhibit a hemisphere difference for sense words. while they, for nonsense words, showed smaller N2 and P3 amplitudes on the left hemisphere as compared with the right. Word recognition time was longer for nonsense words as compared with that obtained for sense ones. P3 amplitude and word recognition time were parallel among male and female subjects and across task conditions. An interpretation of these findings according to information arousal models is given. INTRODUCTION Individual differences in trait anxiety or neuroticism, as reflected in individual differences in arousal responses to stimulation, have been considered central to the study of personality and psycho- pathology. Due to the lack of any adequate theory of the behavioural and physiological effects of anxiety, and because of contrasting effects of anxiety across studies, measurement efforts have often been frustrated. Michael Eysenck (1982, 1983) reviewing the literature reported that the most common result across studies was a significant interaction between trait anxiety and task difficulty. Eysenck (1985) demonstrated that anxiety adversely affects the central executive component of the working-memory system impairing the rehearsal and storage of task-relevant information. The present study examined the possibility that anxiety may differentially affect human event-related brain potentials (ERPs) when they are elicited during the recognition of sense and nonsense words. A theoretic trend assumes that the information processing system is more readily respondent to familiar stimuli than to unfamiliar ones (Broadbent & Broadbent, 1975; Treisman, 1978; O’Connor & Forster, 1981). On this assumption we expected that anxiety might interfere in a differential way with the word recognition process in the case of sense and nonsense words. More specifically, anxiety could mainly affect the late ERP components elicited during the recognition of nonsense words which would require a greater attentional processing capacity than that required by sense words. In this respect, some studies have reported that the recognition of ‘uncommon’, as compared with ‘common’ words, requires a greater amount of lexical search (Glanzer & Ehrenreich, 1979; Taft, 1979; Stanners, Jastrzembski & Westbrook, 1975). The small amount of research comparing the effects of trait anxiety and neuroticism on ERP reponses has been mainly carried out on anxiety neurotic patients, and has not produced consistent findings. Bond, James 8~ Lader (1974) reported shorter latencies for peaks between 100 and 200 msec and a smaller amplitude for P200 and N280 peaks when they elicited ERPs with clicks. In contrast, Chattopadhyay, Cooke, Toone & Lader (1980) found a larger P300 peak and no difference in latency for Ss experiencing high anxiety. However, Beech, Ciesielski & Gordon (1983) partially confirmed previous results showing smaller amplitudes and shorter latencies of the N200 and P300 peaks in obsessional patients than in controls. Moreover, Plooij-van Gorse1 (198 1) elicited ERPs by warning stimuli in a reaction time (RT) paradigm and compared a group of anxiety 741