Object-Centered Visual Neglect, or Relative Egocentric Neglect?

Abstract

There has been much recent interest in the spatial frame(s)-of-reference that may influence visual neglect after brain injury (Pouget & Driver, 1999; Bisiach, 1996; Driver, Baylis, Goodrich, & Rafal, 1994). In principle, a patient with left neglect after right-hemisphere damage might neglect information towards the left of their retina, towards the left of their head, or their body (Moscovitch & Behrmann, 1994; Karnath, Christ, & Hartje, 1993; Vallar, Guariglia, Nico, & Bisiach, 1993; Farah, Brunn, Wong, Wallace, & Carpenter, 1990; Ladavas, 1987; Bisiach, Capitani, & Porta, 1985), and so on. Neglect operating relative to some part(s) or other of the patient has been termed egocentric neglect. Recent research has contrasted such egocentric deficits with a form of neglect that initially seems to be more ‘‘allocentric,’’ applying to one side of individual objects (Tipper & Behrmann, 1996; Behrmann & Moscovitch, 1994; Arguin & Bub, 1993; Driver & Halligan, 1991; Caramazza & Hillis, 1990). For instance, several studies have now found that patients may neglect the left side of an object (or figure, or perceptual group) whether or not it appears in the left or right hemifield with respect to the patient’s eyes, head, and body (e.g., Humphreys, Olson, Romani, & Riddoch, 1996; Walker, 1995; Halligan & Marshall, 1994; Arguin & Bub, 1993; Young, Hellawell, & Welsh, 1992; Driver & Halligan, 1991; Driver, Baylis, & Rafal, 1992; Rapcsak, Varfaelli, Fleet, & Heilman, 1989). Such findings have often been interpreted (e.g., see Vallar, 1998) as indicating ‘‘allocentric object-centered’’ neglect, as opposed to neglect within egocentric representations of space. A recent study by Pavlovskaya, Glass, Soroker, Blum, and Grosswasser (1997) made strong new claims in this respect. Here, we argue that their data, and, likewise, many previous reports of putatively ‘‘object-centered’’ neglect, can actually be explained in purely egocentric terms, provided that relative egocentric position matters in addition to absolute egocentric position (Driver, 1998; Pouget & Sejnowski, 1997; Driver et al., 1994). The reason for this is illustrated in Figure 1. Figure 1A represents the popular notion of an egocentric gradient of impairment in neglect following right-parietal injury (Pouget & Sejnowski, 1997; Anderson, 1996; Driver et al., 1994; Mozer & Behrmann, 1990; Kinsbourne, 1987). The neural response to stimuli is increasingly impaired toward the patient’s egocentric left. Figure 1A represents this by a decreasing number of parietal cells in the lesioned system for positions toward the retinal left, in accordance with neurophysiological data on the distribution of receptive fields between the hemispheres (see Pouget & Sejnowski, 1997). However, the general points made below would still hold if the graded deficit affected leftward positions relative to the head or body (i.e., within other egocentric coordinates), as well as on the retina; and hold regardless of the exact shape of the depicted function, provided that it declines monotonically towards the egocentric left. Figure 1B illustrates the hypothetical response strength for different lateral regions of a schematic object, shown in either the left or right egocentric hemispace, given the pathological gradient shown in Figure 1A. Note that while the overall level of response is lower for the left visual field, the left side of the object still induces a weaker response than the right side of the object within both fields. This alone is sufficient to explain many reports of putatively ‘‘objectcentered’’ neglect, within purely egocentric terms, without the need to postulate specifically object-centered representations. Neglect for the retinally left side of an object within either visual field might, thus, more properly be called ‘‘relative egocentric neglect,’’ rather than truly ‘‘object-centered neglect.’’ The results from a few studies (e.g., Driver et al., 1992) do require that the affected visual object gets segmented from its background before the graded impairment applies. Nevertheless, the latter impairment, which causes the neglect, would still be entirely egocentric; that is, the side of the segmented object that is further to the patient’s left is the side that suffers. Pavlovskaya et al. (1997) claim to have obtained new evidence for truly object-centered, ‘‘allocentric’’ visual neglect. They suggest that in cases with left neglect of this kind, the patient should be better at recognizing objects when forced to allocate attention to the left side of each object. To test this, they attempted to hold attention at a constant egocentric position (where a central fixation point appeared) while manipulating