A spatial explanation for synchrony biases in perceptual grouping: Consequences for the temporal-binding hypothesis

The human brain contains many functionally discrete regions involved in the processing of specific object attributes, such as form, color, motion, and location (Felleman & Van Essen, 1991; Ungerleider & Haxby, 1994; Zeki et al., 1991). A current issue in visual perception is understanding the mechanism by which these disparate representations remain in correspondence with one another and yet distinct from the attributes of other, simultaneously perceived objects (Reynolds & Desimone, 1999). A popular explanation is that all the neurons responsive to a particular object would temporally synchronize their firing across the visual cortex to bind the various representations (Malsburg, 1995; Singer & Gray, 1995). This proposal has received backing from several neurophysiological studies, although the results remain controversial (Singer & Gray, 1995; Tovee & Rolls, 1992). In addition to the neurophysiological work, several behavioral studies have described how perceptual grouping emerges from (Lee & Blake, 2001) or is influenced by (Leonards, Singer, & Fahle, 1996; Parton, Donnelly, & Usher, 2001; Usher & Donnelly, 1998), the temporal synchrony of stimulus presentation. Some authors have been guarded in their postulation of mechanisms that could underlie such sensitivity (Lee & Blake, 2001), whereas others see it as evidence for the time-based binding mechanism described above (Leonards et al., 1996; Parton et al., 2001; Usher & Donnelly, 1998). The behavioral studies have themselves generated a great deal of correspondence and controversy over the last few years, as has been detailed in a succinct review by Farid (2002). Early reports of temporal sensitivity by Fahle (1993), for example, were later attributed to spurious motion cues (Morgan & Castet, 2001). Indeed, as Fahle himself described, some subjects reported perceiving apparent motion in the displays, consistent with the earlier work of Westheimer and McKee (1977), who attributed their temporal hyperacuity effects to motionprocessing mechanisms. Later work by Lee and Blake (1999b), aimed at circumventing these types of motion cue, has itself been criticized on the basis of contrast and motion artifacts (Adelson & Farid, 1999; Farid & Adelson, 2001; Lee, & Blake, 1999a; Morgan & Castet, 2001), and subsequent results from the same group (Lee & Blake, 2001) have also produced controversy (Farid, 2002). In more recent work, Fahle and colleagues have once again attempted to counter earlier criticisms (Kandil & Fahle, 2001), but they have reported effects that were, at times, an order of magnitude too long to be of direct relevance to a temporal-binding mechanism (Farid, 2002; Morgan & Castet, 2001). Other recent studies have offered alternative criticisms. Beaudot (2002), for example, suggested that in the case of contour detection, the order of presentation (contour/background) might play a role in a subject’s sensitivity to stimulus asynchrony. Dakin and Bex (2002) reported that altering stimulus contrast, or the use of a mask, reduces or even abolishes