1. INTRODUCTION Emotional stimuli that are irrelevant to task performance may redirect attentional resources, particularly among individuals evidencing high trait anxiety. EEG markers for this emotional capture of attention and their relation to attention performance remain unclear. This study included a sample of normal adults and examined two questions: First, whether ERP components reflecting initial emotional detection (N1), emotional significance processing (early posterior negativity (EPN)), and conflict monitoring (N2) differed by level of trait anxiety and emotional stimulus type; second, whether these ERP measures of emotional processing were related to subsequent attention performance. Hypotheses: 1. ERP amplitudes reflecting relatively more rapid and automatic emotional detection and salience processing (N1, EPN) would be less sensitive to trait anxiety compared to more “top-down” executive processing (N2). 2. ERP effects would be related to decrements in subsequent attentional performance, but this may depend on attention system: alerting, orienting, and conflict resolution. 3. We explored these processes in a sub-group of the sample showing only high depression, and expected unique effects: Relatively high levels of depression would be associated with reduced ERPs overall, particularly to sad faces, due to withdrawal of task engagement associated with depressed mood. METHOD Participants. Non-clinical sample including 18 adults, aged 19-34. Trait Anxiety and Depression were measured via self-report using the STAI and the BDI. The Attention Network Task (Fan et al., 2002). This 30-minute task is a combination of a cued reaction time and a flanker task that requires the subject to determine whether a central arrow points to the left or right. Emotional Faces were 48 black-and-white photographs of fearful, sad, and neutral faces (Tottenham et al., 2002). Attention scores. The ANT generates a reaction time score from the flanker task. Two blocks of 192 trials were defined by whether trials were randomly preceded by fearful versus neutral faces or sad versus neutral faces. Each trial lasts 4000 msec plus the 50 msec inter-trial face. Alerting (RT no cues – double cues), Orienting (RT center cues – spatial cues), Conflict Interference (RT incongruent – congruent flankers). ERP recording and analysis EEG activity was recorded from 64 scalp sites, left mastoid reference, re- referenced to linked mastoids, using AgCl electrodes in an elasticized cap. 0.01 – 100 Hz input range, with a subsequent 30 Hz low-pass filter. Quantified data into 40 20ms epochs covering an interval from 0-800 ms. PCA conducted revealing three components: N1 maximal at 200 msec, EPN, maximal at 250 msec, and an anterior N2, maximal at 320 msec. Using Source Analysis (BESA 5.1), these components were localized to three dipoles in the areas of the rostral ACC, the posterior attention system, and the right hemisphere. RESULTS Examination of Trait Anxiety Table 1 Repeated Measures ANOVA at Target Electrode Sites (FCz, Cz, PCz, Pz, POz) for Effects of Trait Anxiety and Face Type on Peak Amplitudes in Three Time Windows (N1, EPN, N2) Time window Variable dfN1 (partial eta²)EPN (partial eta²)N2 (partial eta²) ELEC3, 146.51 (.58)**10.02 (.68)***7.01 (.60)** HEM * ANX1, 165.44 (.25)* EM * ELEC * ANX¹3, 14 3.45 (.43)* EM * ELEC * HEM² 3, 146.55 (.58)** p <.05.** p <.01. *** p <.001. ¹ indicates the analysis comparing fear and neutral-f trials. ² indicates the analysis comparing fear and sad trials. Trait Anxiety versus Depression: Distinct Associations with ERP Measures of Emotional Processing and Attention Tracy Dennis, Chao-Cheng Chen, Melville Malone, Tara Nolan & Michael Parker Hunter College, City University of New York Figure 1. The EPN was most pronounced in posterior versus anterior electrodes, particularly in the right hemisphere following fearful faces. *High (solid line) versus low (dashed line) trait anxiety participants Left Hemisphere Right Hemisphere Posterior P7 Fear Sad P8 Fear Sad +4 μv -4 μv Anterior C3C4 +4 μv -4 μv Figure 2. The N2 was most pronounced in anterior electrodes (FCz) among high versus low trait anxiety participants following fearful and neutral-f faces. Though the same trend is apparent in waveforms for sad and neutral-s faces, statistical tests did not reach significance. Fear and Neutral-fear Fear (high – low anxiety) +4 μv -4 μv Sad and Neutral-sad Sad (high – low anxiety) +4 μv -4 μv Figure 3. As N2 and EPN amplitudes increased, high but not low trait anxiety participants showed (a) decreasing orienting efficiency and (b) greater conflict interference, respectively. (a) r =.39 for high and -.64 for low trait anxiety. (b) r = -.66 for high and -.38 for low trait anxiety. Preliminary Examination of Depression Groups Figure 4. Grand-averaged ERPs at FCz for high (solid line) versus low (dashed line) depression groups. The N2 was most pronounced in anterior electrodes among low versus high depression participants across all conditions, but this only reached significance following sad faces. These N2 amplitudes were negatively correlated with orienting (r = -0.36). FCz High Depression Low Depression +4 μv -4 μv DISCUSSION Summary of Findings: For all participants, an N1 component emerged around 200 msec in anterior electrodes. Later, around 250 msec, an EPN was evident, which was greater following fearful but not sad faces. In the high versus low trait anxiety group, N2 amplitudes around 320 msec were greater during blocks of trials with fearful faces. As N1, EPN, and N2 to fearful faces increased, orienting and executive attention decreased, but mainly in the high trait anxiety group. High depression linked to reduced N2 amplitudes, particularly following sad faces, suggesting withdrawal from negative emotional stimuli. Implications: Results suggest that ERP components associated with relatively rapid and automatic stimulus (N1) and emotional salience (EPN) processing were similar for all participants. In contrast, high trait anxiety was characterized by increased executive, top-down processing of fear (conflict monitoring reflected in N2), a potential mechanism in the effects of anxiety on attention performance and emotional distractibility during cognitive tasks (Compton, 2003). Future Directions: Emotion induction techniques. Block-presentation of faces in addition to randomly presented faces in order to examine whether N2 emerges under both conditions. Sample of individuals showing only high levels of depressive symptoms in order to tease apart unique emotion processing and attentional tendencies associated with trait anxiety versus depression. REFERENCES Compton, R. J. (2003). The interface between emotion and attention: a review of evidence from psychology and neuroscience. Behavioral and Cognitive Neuroscience Reviews, 2, 115-129. Fan, J., McCandliss, B., Somer, T., Raz, A., & Posner, M.I. (2002). Testing the efficiency and independence of attentional networks. Journal of Cognitive Neuroscience, 14, 340- 347. Tottenham, N., Borscheid, A., Ellertsen, K., Marcus, D. J., & Nelson, C. A. (2002, April). Categorization of facial expressions in children and adults: Establishing a larger stimulus set. Paper presented at the meeting of Cognitive Neuroscience Society, San Francisco. Acknowledgement. This research was supported by NIH grant 5S06GM060654-04.