Academic journal article Attention, Perception and Psychophysics

Further Evidence for the Spread of Attention during Contour Grouping: A Reply to Crundall, Dewhurst, and Underwood (2008)

Academic journal article Attention, Perception and Psychophysics

Further Evidence for the Spread of Attention during Contour Grouping: A Reply to Crundall, Dewhurst, and Underwood (2008)

Article excerpt

In a contour-grouping task, subjects decide whether contour elements belong to the same or different curves. Houtkamp, Spekreijse, and Roelfsema (2003) demonstrated that object-based attention spreads gradually over contour elements that have to be grouped in perception. Crundall, Dewhurst, and Underwood (2008) challenged this spreading-attention model and suggested that attention in the contour-grouping task is not object based but rather has the shape of a zoom lens that moves along the relevant curve. To distinguish between object-based and spatial attention, they changed the stimulus and measured the impact on performance. Subjects were not able to correct for changes at the start of the relevant curve toward the end of the trial. They suggested that attention did not stay at the beginning of the curve, in accordance with a moving zoom lens model. Here, we examine the task of Crundall et al. and find that subjects perceive the changes but fail to correct their response. By measuring change detection directly, we find that performance is much better for the start of the relevant curve than for an irrelevant curve, at all times. Our findings do not support the zoom lens model but provide further support for the spreading-attention model.

The typical visual scene that we perceive is cluttered with many objects embedded in a complex background. To analyze this wealth of information, our visual system starts with a decomposition of the image into small, digestible parts. The first steps of image processing are carried out by neurons with small receptive fields that oversee only a small fraction of the incoming information and that are tuned to simple features, such as line elements of a particular orientation and surface patches with a specific color or texture. This piecemeal analysis is very different from our subjective perception. We do not perceive a set of small image fragments but interact with spatially extended objects that may fill hundreds or thousands of receptive fields. Thus, our visual system must be equipped with powerful perceptual-grouping processes that reconstruct objects from these image fragments. We appear to perceive the objects immediately, even if they are large, and it is easy for us to tell where in the picture one object ends and the next one begins.

Neurons in higher areas of the visual cortex presumably account for some of this efficiency. These neurons have large receptive fields and are tuned to complex objects, like faces and other shapes (Oram & Perrett, 1994; Riesenhuber & Poggio, 1999; Roelfsema, 2006; Tanaka, 1993; Tsao, Freiwald, Tootell, & Livingstone, 2006), and they are activated only a few tens of milliseconds after the neurons in early visual areas (Hung, Kreiman, Poggio, & DiCarlo, 2005; Oram & Perrett, 1992; Sugase, Yamane, Ueno, & Kawano, 1999). The activity of a neuron tuned to, say, a face may explain how the visual brain rapidly detects the group of image elements that belong to such a familiar object (Rousselet, Macé, & Fabre-Thorpe, 2003; Thorpe, Fize, & Marlot, 1996). The efficient detection of groups of image elements by selective neurons in higher areas of the visual cortex has been called base grouping (Roelfsema, 2006).

However, there are also perceptual-grouping tasks that demand more processing time. An example of such a task is shown in Figure 1A. Imagine approaching the desk because you want to switch on the light. To determine which plug to put into the socket, you have to know which plug is attached to the lamp. This task could be solved by a perceptual-grouping operation that groups together all the contour elements of the lamp cable and segregates them from the contour elements of the other cable. In an elegant series of studies, Jolicoeur and colleagues (Jolicoeur & Ingleton, 1991; Jolicoeur, Ullman, & Mackay, 1986, 1991) demonstrated that a laboratory version of this task requires many hundreds of milliseconds, and that the response time (RT) of subjects increases linearly with the number of contour elements that need to be grouped in perception. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.