Please cite this article in press as: Conci, M., et al. The time-course of global and local attentional guidance in Kanizsa-figure detection. Neuropsy- chologia (2011), doi:10.1016/j.neuropsychologia.2011.04.023 ARTICLE IN PRESS G Model NSY 4117 1–9 Neuropsychologia xxx (2011) xxx–xxx 1 Contents lists available at ScienceDirect Neuropsychologia journal homepage: www.elsevier.com/locate/neuropsychologia The time-course of global and local attentional guidance in Kanizsa-figure detection 1 2 Markus Conci a,∗ , Thomas Töllner a , Marcin Leszczynski a , Hermann J. Müller a,b 3 a Department of Psychology, Ludwig-Maximilians-University, Munich, Germany 4 b Department of Psychology, Birkbeck College London, UK 5 6 article info 7 8 Article history: 9 Received 5 January 2011 10 Received in revised form 30 March 2011 11 Accepted 16 April 2011 12 Available online xxx 13 Keywords: 14 Visual attention 15 Kanizsa figure 16 Perceptual grouping 17 Local/global processing 18 Event-related potentials 19 abstract Object configurations can be perceptually represented at various hierarchical levels. For example, in visual search, global Kanizsa figures are detected efficiently, whereas search for local groupings is inefficient, with similarity-dependent nontarget interference arising at the hierarchical level that defines the tar- get (Conci, Müller, & Elliott, 2007). The present study was designed to examine the electrophysiological correlates of this global–local search asymmetry. The results revealed differences between hierarchical object levels to be evident throughout a number of processing stages: search for a global, versus a local, target elicited larger amplitudes in early sensory components (P1, N1). Moreover, the efficiency of atten- tional orienting towards a target was mirrored in the Posterior Contralateral Negativity (PCN), with PCN latencies being substantially delayed (by ∼70 ms) with local, versus global, targets. Finally, late compo- nents (P3 and slow wave—SW) reflected the overall search efficiency, which was determined by both the hierarchical level at which the target was defined and the similarity-based nontarget interference. Taken together, this pattern shows that multiple, sequential processes of object completion contribute to the attentional precedence of a globally bound object over a mere local element grouping. © 2011 Published by Elsevier Ltd. 1. Introduction 20 Natural scenes typically consist of multiple objects that can be 21 described at various, hierarchical levels of organization (or repre- 22 sentation). For instance, a forest has trees, and the trees, in turn, 23 have leaves. A similar hierarchical relationship between parts and 24 wholes has also been demonstrated using a variety of composite fig- 25 ures. For example, the Navon letter (Navon, 1977) in Fig. 1A consists 26 of a local arrangement of the letters “H”, which combine to form the 27 global letter “U”. Similarly, Kanizsa figures (Kanizsa, 1976) com- 28 prise a local arrangement of “pacman”-type inducers. When the 29 local inducers face inwards (see Fig. 1B), a salient, global figure (in 30 the example, a square) emerges (referred to as an ‘illusory figure’). 31 By contrast, when the local inducers face outwards (see Fig. 1C), no 32 corresponding global figure is induced. Both examples of groupings 33 illustrated in Fig. 1B and C are in principle identical with regard to 34 their physical stimulus attributes (i.e., with respect to the inducer 35 elements or, respectively, the ‘local’ level of representation). How- 36 ever, only the Kanizsa figure gives rise to an illusory square, which 37 can be considered a ‘global’ object representation (see Fig. 1 for 38 ∗ Corresponding author at: Allgemeine und Experimentelle Psychologie, Depart- ment Psychologie, Ludwig-Maximilians Universität, Leopoldstr. 13, D-80802 München, Germany. Tel.: +49 089 2180 4889; fax: +49 089 2180 5211. E-mail address: conci@psy.lmu.de (M. Conci). an illustration of global and local object representations; see also 39 Rauschenberger & Yantis, 2001). Thus, a variety of objects can be 40 represented at multiple levels, with global properties being depen- 41 dent on the existence and arrangement of more elementary local 42 parts. 43 Also, generally, object information can be accessed at vari- 44 ous hierarchical levels. However, the default level of attentional 45 orienting has often been described in terms of “global prece- 46 dence”, illustrating that the ‘forest comes before the trees’ (Navon, 47 1977; Rauschenberger & Yantis, 2001; Rensink & Enns, 1995). For 48 instance, Rensink and Enns (1995) demonstrated that visual search 49 is based on the global level of representation that comprises inte- 50 grated objects, while corresponding fragmentary (local) parts could 51 not be accessed with comparable efficiency. In general agreement 52 with this finding, Conci, Müller, and Elliott (2007) showed that hier- 53 archical levels in Kanizsa figures give rise to a search asymmetry 54 in favor of global precedence: search for a global Kanizsa square 55 (Fig. 1B) was more efficient than search for a comparable local con- 56 figuration (Fig. 1C; mean search reaction time (RT)/set size slopes 57 were 16 and 46 ms/item, respectively). Consequently, detection of a 58 global shape precedes detection of a configuration that is uniquely 59 defined at a local hierarchical level (see also Conci, Müller, & Elliott, 60 2009, for a comparison of visual search for different types of global 61 shape). 62 Moreover, the study of Conci et al. (2007) demonstrated that 63 variations of target–nontarget similarity (cf. Duncan & Humphreys, 64 0028-3932/$ – see front matter © 2011 Published by Elsevier Ltd. doi:10.1016/j.neuropsychologia.2011.04.023