Distinct neural correlates of attending speed vs. coherence of motion S. Kau a, ⁎, H. Strumpf a , C. Merkel a , C.M. Stoppel a , H.-J. Heinze a, b , J.-M. Hopf a, b , M.A. Schoenfeld a, b, c a Department of Neurology, Otto-von-Guericke University, 39120 Magdeburg, Germany b Leibniz Institute for Neurobiology, 39118 Magdeburg, Germany c Kliniken Schmieder, 78476 Allensbach, Germany abstract article info Article history: Accepted 27 August 2012 Available online 4 September 2012 Keywords: Feature-based attention Motion Speed Coherence fMRI EEG/MEG Attention to specific features of moving visual stimuli modulates the activity in human cortical motion sensitive areas. In this study we employed combined event-related electrophysiological, magnetencephalographic (EEG, MEG) and hemodynamic functional magnetic resonance imaging (fMRI) measures of brain activity to investigate the precise time course and the neural correlates of feature-based attention to speed and coherence. Subjects were presented with an aperture of dots randomly moving either slow or fast, at the same time displaying a high or low level of coherence. The task was to attend either the speed or the coherence and press a button upon the high speed or high coherence stimulus respectively. When attention was directed to the speed of mo- tion enhanced neural activity was found in the dorsal visual area V3a and in the IPL, areas previously shown to be specialized for motion processing. In contrast, when attention was directed to the coherence of motion signifi- cant hemodynamic activity was observed in the parietal areas fIPS and SPL that are specialized for the processing of complex motion patterns. Concurrent recordings of the event-related electro- and magnetencephalographic responses revealed that the speed-related attentional modulations of activity occurred at an earlier time range (around 240–290 ms), while the coherence-related ones occurred later (around 320–370 ms) post-stimulus. The current results suggest that the attentional selection of motion features modulates neural processing in the lowest-tier regions required to perform the task-critical discrimination. © 2012 Elsevier Inc. All rights reserved. Introduction Attention selection can be based on various visual features, such as color or motion. Previous studies showed that attending to such fea- tures enhanced neural activity in corresponding feature specific cortical areas along the dorsal and ventral visual stream (Corbetta et al., 1990, 1991; O'Craven et al., 1999; Saenz et al., 2002; Schoenfeld et al., 2007; Stoppel et al., 2007; Stoppel et al., 2011; Treue and Martinez Trujillo, 1999; Treue and Maunsell, 1996). Most studies reported enhanced neural activity in the dorsal stream visual area hMT/V5 (Chawla et al., 1999; Corbetta et al., 1990, 1991; Culham et al., 2001; Heeger et al., 1999; O'Craven et al., 1997; Saenz et al., 2002; Schoenfeld et al., 2007; Stoppel et al., 2011; Tootell et al., 1995; Watson et al., 1993) when a motion stimulus was attended. However, within the visual motion feature dimension there are simple sub-features such as speed or direction but also complex ones like coherence that can be selectively attended to. These simple sub- features can already be processed at single neuron level of motion re- sponsive neurons that have a preferred direction and speed. Attending to the direction of a moving stimulus enhances the baseline activity and firing rate of neurons in area MT tuned to that direction (Treue and Martinez Trujillo, 1999). The same effect in area hMT could recently also be shown at population level using fMRI (Stoppel et al., 2011). Nev- ertheless, the discrimination of simple motion sub-features like speed or direction does not necessarily rely on processing in area MT and can easily be performed in lower tier motion sensitive areas such as V3 and V3a (Sunaert et al., 1999; Tootell et al., 1997). The activity in these areas has been shown to be modulated by visual attention (Buchel et al., 1998; Stoppel et al., 2011; Sunaert et al., 2000). For more complex properties such as motion coherence it is less clear in which areas the processing takes place. In order to discrimi- nate between different coherences information from several neurons with different preferred directions has to be integrated. Typically such stimuli elicit strong responses in higher-tier parietal areas fIPS, SPL and IPL (Orban et al., 2004; Peuskens et al., 2004; Sunaert et al., 1999; Vanduffel et al., 2002). On the other hand, hMT itself (reviewed in (Orban, 2011)), as well as some other neighboring regions in the lateral occipital cortex such as LO (Grill-Spector et al., 1998; Grill-Spector et al., 1999; Grill-Spector et al., 2000; Grill-Spector et al., 2001; Kourtzi and Kanwisher, 2000, 2001; Kourtzi et al., 2002; Malach et al., 1995) or KO (Dupont et al., 1997) could perform the processing. Importantly, the activity in these regions is typically increased when attention is paid to visual motion (Braddick et al., 2000; Heeger et al., 2000; Saenz et al., 2002; Schoenfeld et al., 2007; Stoppel et al., 2011; Treue and Martinez Trujillo, 1999) especially when the stimulus contains complex motion sub-features such as coherence or 3D structures from motion NeuroImage 64 (2013) 299–307 ⁎ Corresponding author. E-mail address: stefanie.kau@med.ovgu.de (S. Kau). 1053-8119/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.neuroimage.2012.08.080 Contents lists available at SciVerse ScienceDirect NeuroImage journal homepage: www.elsevier.com/locate/ynimg