The Role of Perceptual Organization in Biasing Competition (McMains)

When multiple stimuli appear simultaneously in the visual field, they are not processed independently, but rather interact in a mutually suppressive way suggesting that they compete for neural representation. The biased competition model of selective attention suggests that both top-down and bottom-up processes can help resolve this competition by reducing suppressive interactions between competing stimuli. Both top-down attention (Kastner el al. 1998) and bottom-up visual salience (via pop-out visual stimuli, (Beck and Kastner 2005) have been found to reduce competition in extrastriate cortex. We are currently investigating whether mechanisms that subserve illusory contour formation and collinearity can reduce the competitive interactions between competing stimuli in visual cortex in a bottom-up fashion. Our work suggests that perceptual grouping is a general principle by which competition is biased in favor of foreground elements, leading to the formation of candidate objects that will be more likely to be selected by top-down attention.

Functions of Human Posterior Parietal Cortex (Konen)

Most of our current knowledge about the primate PPC has been derived from invasive methods in macaque monkeys. The PPC is an association area and contains a complex mosaic of numerous subregions associated with the transformation of sensory input to motor output in order to form an abstract representation of space. For instance, the lateral intraparietal area (LIP) is crucial for the process of spatial attention and acts as a controller for saccadic eye movements (Andersen 1997, Colby & Goldberg 1999). However, it is still an open question which area in the human PPC is equivalent to area LIP (Sereno et al. 2001, Schluppeck et al. 2005). We are currently using the delayed saccade task to define topographic maps in human PPC. The anatomical and functional definition of these areas gives important information about the architecture of the human PPC and forms the basis for probing functionality related to a multitude of cognitive tasks. Thus, the fMRI experiments will reveal whether the human parietal cortex is organized in a similar or completely different fashion compared to non-human primates.

Attentional Modulation of the Human LGN (Kung)

The lateral geniculate nucleus (LGN) appears to be the first processing stage that is affected by attentional modulation of neural activity. We have shown that selective attention modulated neural activity in the LGN in multiple ways; by enhancing nerual responises to attended stimuli, by attenuating those to ignored stimuli, and by increasing baseline activity in the absence of visual stimulation (O'Connor et al. 2002). These findings suggest that the LGN plays an important role in attentional gain control. The nature of the gain mechanism is unknown. Building on our work in which we have been able to distinguish the magnocellular and parvocellular portions of the LGN, as well as the retinotopic structure of the LGN, we are currently investigating how the effects of attention are distributed throughout the LGN.

Attentional Control Functions of Frontal and Parietal Cortex Revealed by Phased-encoded Attention Task (Szczepanski)

Our lab, among several others in the field, has recently revealed the existence of topographically organized regions within higher- order cortex using a memory-guided saccade task (Kastner et al, 2007; Schluppeck et al, 2005). Six separate maps have been identified bilaterally within and surrounding the intraparietal sulcus (IPS1-IPS5) and in the superior parietal lobule (SPL), each containing a representation of the contralateral visual field and separated by reversals in the visual field orientation. We have also identified two topographically organized regions bilaterally within frontal cortex: one in the region of the superior branch of precentral cortex and superior frontal sulcus (PreCC/SFS) and the other in the region of the inferior branch of precentral cortex and the inferior frontal sulcus (PreCC/IFS). These topographic maps overlap considerably with the fronto-parietal attention control network, which includes the frontal eye fields, supplementary eye fields, and portions of parietal cortex (SPL and IPS) in the human. Based on this discovery, we are able to define regions within the attention control network that are topographically organized (e.g. PreCC/SFS, PreCC/IFS, IPS 1-5, SPL) and regions of frontal and parietal cortex that are part of the attention network, but do not show any topographic organization. Such a region of interest (ROI) approach allows us to further divide the massive attention network activation into more discrete areas and to probe the attentional control functions of each of these areas independently.

Functional organization of the magnocellular and parvocellular LGN in dyslexic populations (Kung, McCandliss)

Candidate magnocellular (M) and parvocellular (P) human LGN voxels can be reliably identified by their differential contrast responses using high-resolution fMRI (Schneider et al., 2004). This methodology can be applied to test the magno-deficit hypothesis, one of the major theories underlying reading difficulty spectrum (i.e., dyslexia. See Stein 2001 for review). In collaboration with Dr. McCandliss (Sackler Institute for Developmental Psychobiology), who has extensive experience in developmental dyslexia assessment and diagnosis, we compare the functional organization of LGN between dyslexic andnormal subjects, to better understand the relationship between the functional deficiencies and/or anatomical anomalies of dyslexic's LGN, and their reading performances.

Topography in Posterior Parahippocampal Cortex (Arcaro)

Recent studies have demonstrated two retinotopically organized areas, VO-1 and VO-2, in human ventral occipital cortex adjacent to hV4 (see Wandell et. al. 2007 for review). However, it is not clear whether the VO complex is the anterior extent of topographic organization within the human ventral stream. We are currently exploring topographic organization in areas adjacent to the VO complex by combining a salient bottom-up stimulus with a top-down cognitive task. Such stimuli have been used to identify topographic organization, beyond retinopically defined visual cortex, within the intraparietal sulcus (Schluppeck et. al. 2006) and frontal cortex (Kastner et. al. 2007). By establishing topographic borders, we will be able to probe functionality within ventral-temporal cortex without relying on category-specific stimuli (e.g. faces, objects, color) to establish regions of interest.

Rapid detection of objects in natural scene (Peelen)

Natural scenes can be recognized and categorized with remarkable ease. For example, when presented very briefly with photographs of complex cluttered environments, subjects are extremely good in detecting the presence of highly familiar objects such as animals and cars (Thorpe et al., 1996). This fast detection has also been observed in situations where attention is drawn away from the image, suggesting that little attention is needed to process natural scenes, which is in stark contrast to seemingly simpler artificial stimuli that need focal attention to be discriminated (Li et al., 2002). How does the visual system manage to recognize complex images so rapidly? Specifically, at what processing stage of the visual system do patterns of fMRI activity in response to target and distracter images start to diverge? We currently investigate these issues in collaboration with Dr. Fei-Fei Li from the Computer Science department.

Sources of Spatial and Feature-Based Attention (Mruczek, Peelen)

Studies of spatial attention have identified a network of higher-order areas in frontal and parietal cortex that appear to be the source of modulating attention signals (Kastner and Ungerleider, 2000).  In contrast, there is no compelling evidence for the existence of a complimentary and independent network for feature-based attention.  Regions of cortex that contribute to top-down feature-based attentional effects should contain a map of the feature dimensions that could be attended (Maunsell and Treue, 2006), but in contrast to the topographic representation of visual space (Kastner et. al. 2007, Schluppeck et al, 2005) there is little evidence for feature-selective responses in frontal or parietal cortex.  We are currently studying the mechanisms of spatial and feature-based attention, and their interactions across cortex, to more clearly define the role of the frontoparietal network in visual attention.

Overview

Methods

Back