THE INTERACTION OF SPATIAL AND OBJECT PATHWAYS - EVIDENCE FROM BALINTS-SYNDROME

An earlier report described a patient (RM) with bilateral parietal dam age who showed severe binding problems bet tween shape and color and s hape and size (Friedman-Will, Robertson, & Treisman, 1995). When shown two different-colored letters, RM reported a large number of illusory conjunctions (ICs) combining the shape of one letter with the color o f the other, even when he was looking directly at one of them and had as long as 10 sec to respond. The lesions also produced severe deficit s in locating and reaching for objects, and difficulty in seeing more than one object at a time, resulting in a neuropsychological diagnosis of Balint's syndrome or dorsal simultanagnosia. The pattern of defici ts supported predictions of Treisman's Feature integration Theory (FIT ) that the loss of spatial information would lead to binding errors. T hey further suggested that the spatial information used in binding dep ends on intact parietal function. In the present paper we extend these findings and examine other deficits in RM that would be predicted by FIT. We show that: (1) Object individuation is impaired, making it imp ossible for him correctly to count more than one or two objects, even when he is aware that more are present. (2) Visual search for a target defined by a conjunction of features (requiring binding) is impaired, while the detection of a target defined by a unique feature is not. S earch for the absence of a feature (O among Qs) Is also severely impai red, while search for the presence (Q among Os) is not. Feature absenc e can only be detected when all the present features are bound to the nontarget items. (3) RM's deficits cannot be attributed to a general b inding problem: binding errors were far more likely with simultaneous presentation where spatial information was required than with sequenti al presentation where time could be used as the medium for binding. (4 ) Selection for attention was severely impaired, whether it was based on the position of a marker or on some other feature (color). (5) Spat ial information seems to exist that RM cannot access, suggesting that feature binding relies on a relatively late stage where implicit spati al information is made explicitly accessible. The data converge to sup port our conclusions that explicit spatial knowledge is necessary for the perception of accurately bound features, for accurate attentional selection, and for accurate and rapid search for a conjunction of feat ures in a multiitem display. It is obviously necessary for directing a ttention to spatial locations, but the consequences of impairments in this ability seem also to affect object selection, object individuatio n, and feature integration. Thus, the functional effects of parietal d amage are not limited to the spatial and attentional problems that hav e long been described in patients with Balint's syndrome. Damage to pa rietal areas also affects object perception through damage to spatial representations that are fundamental for spatial awareness.