Masking in RSVP Task difficulty in RSVP Easy letters: G,K,U,V; mean accuracy 93% - Hard letters: B,C,J,P; mean accuracy 64% ` Conscious Perception: All or none? Skeletal is not a perfect substitute Homogeneous More variance Skeletal presentation (e.g. Ward et al., 1997) - the best of both worlds? RSVP effects without interference from surrounding distractors? EEG analysis reveals there is a cost: Considerable differences in how targets are processed Skeletal stacked plot RSVP stacked plot Simultaneous Type Serial Token (ST 2 ) model (Bowman & Wyble, 2006; Wyble & Bowman, 2005) Stage 2 - Tokens Episodic contexts in working memory Stage 1 - Types Features (letters, colours, ..) Working memory encoding ` Accelerated P3 onset & latency for skeletal targets RSVP Skeletal Task filter determines targets Targets marked by visual onset Stats: RSVP/Skeletal comparison Onset (1/3 of max. amplitude) F(1,11) = 13, p < .01 50% area latency F(1,11) = 6., p < .05; area range 0ms-800ms ` Methods Presentation • 12 under- and postgrad students with normal or corrected-to-normal vision (mean age 24.1; 6 male / 6 female; 11 right-/1 left-handed; paid 10 GBP) • Alphanumeric characters (Arial font, 5 cm mean height) presented at a distance of 100cm (2.86° visual angle) on 21” CRT computer screen (1024x768 @ 85Hz) using Psychophysics toolbox (Brainard, 1997) on Matlab 6.5 under Microsoft Windows XP • 4 blocks (3 RSVP/1 Skeletal) of 100 trials; Each block: 96 target & 4 distractor only trials; 50% of targets masked and 50% unmasked; 5 practice trials before first RSVP and Skeletal block • Target letters: B, C, D, E, F, G, J, K, L, P, R, T, U, V ; Distractor digits: 2, 3, 4, 5, 6, 7, 8 ; Last item: . (dot) or , (comma) • 500ms fixation cross before each stream; RSVP stream: 70 items (47.1ms SOA; no inter-stimulus interval; total stream length 3.3 seconds); Skeletal stream: Blank screen for 471ms to 2.5 seconds - then the target (and its mask in the masked condition) for 41.7ms - then blank screen for another 706ms to 2.8 seconds • After viewing the stream subjects entered the target letter or pressed space if they did not see a target; then pressed dot or comma depend- ing on what the last item of the stream was (included to keep subjects’ attention with the stream) EEG • Quickamp amplifier (22-bit digital-to-analog converter; BrainProducts, Munich, Germany) • 2000Hz sampling rate, digitally filtered 0.5Hz (low-pass) to 85Hz (high-pass) at recording • Electrodes: Fp1, Fp2, F3, F4, C3, C4, P3, P4, O1, O2, F7, F8, T7, T8, P7, P8, Fz, Cz, Pz, Oz; EOG (above/below right eye) • Referenced to common average online and re-referenced to linked earlobes offline. Left mastoid acted as ground. • Eye movement artefacts were removed by rejecting data in the window of 200ms before and after an eye blink. • Data was inspected for sudden high/consistently low activity. Epochs from 500ms prior to 500ms after an artefact were marked as bad & removed from further analysis. • Channel data was grouped into five regions of interest: Frontal, Central, Temporal, Parietal & Occipital. • EEG data analysed using BrainProducts Analyzer software and the Matlab EEGLab toolbox (Delorme & Makeig, 2004) • ERP: negative plotted upward; 8Hz low-pass filter to enhance visualisation. RSVP-Skeletal stacked plot: Smoothed by a factor of 5. • Stats: ERP component sizes - ANOVA on peak amplitudes; ERP component latencies - ANOVA on area measures (Luck & Hillyard, 1990) An EEG study of masking effects in RSVP Patrick Craston, Brad Wyble and Howard Bowman Centre for Cognitive Neuroscience and Cognitive Systems Computing Laboratory, University of Kent Canterbury, UK {pc52,bw5,hb5}@kent.ac.uk References Bowman, H. & Wyble, B. (2006), 'The Simultaneous Type, Serial Token Model of Temporal Attention and Working Memory‘, in preparation Brainard, D. (1997), 'The Psychophysics Toolbox', Spatial Vision 10, 433-436 Delorme, A. & Makeig, S. (2004), 'EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent component analysis', Jour- nal of Neuroscience Methods 134(1), 9-21 Kessler, K.; Schmitz, F.; Gross, J.; Hommel, B.; Shapiro, K. & Schnitzler, A. (2005), 'Target consolidation under high temporal processing demands as re- vealed by MEG', NeuroImage 26(4), 1030-1041 Luck, S. & Hillyard, S. (1990), 'Electrophysiological evidence for parallel and serial processing during visual search.', Perception & Psychophysics 48, 603-617 Raymond, J.; Shapiro, K. & Arnell, K. (1992), 'Temporary Suppression of Visual Processing in an RSVP Task: An Attentional Blink?', Journal of Experimental Psychology: Human Perception and Performance 18(3), 849-860 Vogel, E.; Luck, S. & Shapiro, K. (1998), 'Electrophysiological Evidence for a Postperceptual Locus of Suppression During the Attentional Blink', Journal of Ex- perimental Psychology: Human Perception and Performance 24(6), 1656-1674 Ward, R.; Duncan, J. & Shapiro, K. (1997), 'Effects of similarity, difficulty, and nontarget presentation on the time course of visual attention‘, Perception & Psy- chophysics 59(4), 593-600 Wyble, B. & Bowman, H. (2005) 'Computational and Experimental Evaluation of the Attentional Blink: Testing The Simultaneous Type Serial Token Model, in Bara; Barsalou & Bucciarelli, ed.,‘CogSci 2005, XXVII Annual Conference of the Cognitive Science Society‘, Cognitive Science Society, 2371-2376 Acknowledgement This work was supported by the UK Engineering and Physical Research Council under grant number GR/S15075/01 awarded to Howard Bowman and a Doctoral Training Account awarded to Patrick Craston. RSVP masked RSVP unmasked Skeletal masked Skeletal unmasked Introduction The limits of temporal attention are commonly studied by presenting subjects with tar- get items that are difficult to perceive. Rapid Serial Visual Presentation (RSVP), where items are presented in the same spatial location at a rapid rate, evokes the Attentional Blink (AB; Raymond et al., 1992), an impairment in detecting a second target if it follows within short temporal proximity of an attended first target. In addition to behavioural experiments, the AB has also been explored by means of EEG (Vogel et al., 1998) and MEG (Kessler et al., 2005). However, for the AB to occur, targets have to be presented within 700ms of each other and each target item generates an EEG/MEG signal, which lasts up to 800ms. Hence, the close temporal proximity causes Event-Related Potential (ERP) signals to overlay. Our study explores the processing of single targets in RSVP to aid the understanding of results gained from multiple target paradigms. ST 2 model: Unmasked (black) - Masked (red) ST 2 Neuron Stronger activation for unmasked targets ERP: Unmasked (black) - Masked (red) ST 2 model: Easy (black) - Hard (red) ERP: Easy (black) - Hard (red) P3 component significantly larger for unmasked targets Stats: Unmasked/Masked comparison P3 peak amplitudes F(1,11) = 7, p < .05 ST 2 Neuron Stronger activation for easy targets P3 component significantly larger for easy targets Stats: Easy/Hard comparison P3 peak amplitudes F(1,11) = 27, p < .001 ERP: RSVP (black) - Skeletal (red) RSVP Skeletal Minimal P1/N1 Clear P1/N1 Visual percept is continuous Targets are marked by visual onset Stats: RSVP/Skeletal comparison P1/N1 mean value F(1,11) = 13, p < .01 Mean value determined by taking baseline corrected absolute values of area underneath the ERP curve in the range of 60-200ms Occipital - Sensory components (P1/N1) ERP: RSVP (black) - Skeletal (red) Parietal - P3 component For better comparison with ERPs negative is plotted upward Negative plotted upward For better comparison with ERPs negative is plotted upward ERP grand average suggests P3 for perceived targets - no P3 for missed targets Stacked plot shows distribution of P3 sizes is not bimodal "All or none" may be an artefact of ERP averaging ERP: Correct (black) - Incorrect (red) EEG stacked plot