This study adapted the method of partial lesions, combined with controlled fixation, to study the perceptual role of macaque inferotemporal (IT) cortex. Unilateral lesions were made in IT cortex of three monkeys, without section of the corpus callosum, and visual function was tested ipsilateral and contralateral to the lesion. The observed changes were compared to the effects of bilateral lesions of IT cortex in one monkey, the approach used in most previous studies. Unilateral lesions produced far less profound, although more selective, loss on the tested visual abilities than did bilateral lesions. All three monkeys with unilateral lesions showed decreased chromatic sensitivity, but sparing of achromatic sensitivity, and severely disrupted learning and performance of visual matching to sample, and in all cases, the visual loss was contralateral to the site of the lesion. Unexpectedly, the magnitude of the contralateral loss was not increased by later section of the corpus callosum and anterior commissure in one of the monkeys, a lesion that removes interhemisperic input to contralateral from ipsilateral temporal cortex neurons. These results support physiological findings that show that the response of IT cortex neurons is dominated by the contralateral visual field, despite the bilateral activation many IT neurons receive. Comparison to earlier studies of lesions of area V4, which provides input to IT cortex, shows that V4 and IT lesions produce qualitatively different effects.
In recent years, the function of retinotopic cortical visual areas, such as V1, V2 and V4, has been studied by making localized lesions, and using controlled fixation to place test stimuli within and outside the visual field region corresponding to the lesioned area. Because this approach involves a within-subject control for non-visual determinants of discrimination, such as motivation, and understanding of the task, it provides a direct comparison of visual function at affected and control regions of the visual field. Using this approach, it has been shown that lesions to area V1 cause devastating visual loss (Merigan et al., 1993), whereas those in areas V2, V4 and MT produce more moderate sensitivity loss, but complete loss of the ability to discriminate certain stimuli (Merigan et al., 1993; Schiller and Lee, 1994; Rudolph, 1997; Merigan, 2000). However, despite the advantages of this approach, with one exception (De Weerd et al., 1999), it has only been used for studies of cortical areas that have highly retinotopic visual organization (Merigan et al., 1993; Newsome and Pare, 1988; Rudolph and Pasternak, 1999).
Cortical areas in inferotemporal (IT) cortex (TEO and TE) do not show the precise retinotopy of earlier areas. The excitatory receptive fields of TEO neurons are confined largely to the contralateral hemifield (Boussaoud et al., 1991), and the single unilateral lesion study of TEO (De Weerd et al., 1999), also suggests unilateral organization. On the other hand, many receptive fields of TE neurons are large and bilateral, and one of their most dramatic features is the loose retinotopy they demonstrate (Tanaka et al., 1991). However, it is widely recognized that TE receptive fields extend to slightly greater eccentricities on their contralateral edge (Op De Beeck and Vogels, 2000), and that their response properties are dominated by stimuli placed in the contralateral field (Chelazzi et al., 1998). These physiological responses suggest that unilateral lesions of areas TEO or TE might result in greater contralesional than ipsilesional visual loss. If so, this method might be ideal for examining the effects of damage to IT, since emotional or other non-specific effects of bilateral IT lesions (Huxlin et al., 2000) could contribute to the disruption of visual abilities.
In this study, we tested the effects of unilateral IT lesions on those visual abilities we have found to be affected by bilateral IT lesions: luminance and chromatic contrast sensitivity, and the perception of illusory contours (Huxlin et al., 2000). In addition, we examined post-lesion learning and performance of visual matching of stimulus color and shape. Section of the corpus callosum and anterior commissure was added later in one monkey to determine if interhemispheric connections altered the effects of unilateral IT lesions.
Materials and Methods
Animals and Surgical Procedures
Subjects were four adult, female monkeys (Macaca nemestrina) weighing ∼5 kg. Following initial training and the achievement of stable visual thresholds, brain lesions were made using standard sterile neurosurgical procedures. Through a craniotomy, the dura was opened and the desired tissue removed by suction with a 20-gauge stainless steel tube. For two of the four monkeys lesions were made at a single time, but for the other two monkeys they were made in two stages, to help separate the effects of different components of the lesion. Visual testing was carried out until performance was stable, both before and after all lesions (although, as noted below, sensitivity often improved in subsequent test periods). Experiments were carried out in accordance with the NIH Guide for the Care and Use of Laboratory Animals (1987).
Monkey A (TEO + TE) received a one-stage unilateral lesion of right IT cortex that included both TEO and TE (von Bonin and Bailey, 1947; Iwai and Mishkin, 1969). The areas removed extended caudally to include the lateral bank of the inferior occipital sulcus (IOS), and rostrally to the tip of the temporal pole. Dorsally, the ablations included the inferior bank, the fundus and the upper bank of the superior temporal sulcus (STS). Ventrally, they extended to the lateral bank of the occipito-temporal sulcus (OTS).
Monkey B (TEO + TE) received a two-stage unilateral lesion: (i) of right area TEO, and (ii) of right area TE. Monkey C (TE + CC) received a two-stage lesion: (i) of left area TE, and (ii) of the posterior corpus callosum and anterior commissure. Monkey D (bil. TEO + TE) (bilateral lesion) served as a comparison for the effects of bilateral IT lesions, and received a one-stage bilateral lesion of areas TEO and TE.
Following each lesion, monkeys were tested until performance stabilized, which required ∼4 months.
Behavioral Testing Apparatus and Procedures
The monkeys were seated, facing a high-resolution (1152 × 870) 17″ Nanao video display at a distance of 84 cm. They were tested for 200–250 trials/day, 5 days per week, until performance was stable. All testing was carried out binocularly, and none of the monkeys had more than 0.5 D of refractive error in either eye. All animals had free access to monkey chow, supplemented regularly with fresh fruit, but they had limited access to fluids, which served as rewards for correct responses during behavioral testing.
Visual Tasks and Stimuli
Monkeys A (TEO + TE), B (TEO + TE), and C (TE + CC), were tested, with controlled fixation, on the three tasks illustrated in Figure 1, in their right and left visual field, before and after each stage of their unilateral IT lesions. Monkey D (bil. TEO + TE) was tested on the same tasks without fixation control, before and after receiving a bilateral IT lesion.
Luminance Contrast Sensitivity
The monkeys were trained to discriminate the vertical versus horizontal orientation of small patches of grating stimuli (Gabor) (vertical orientation shown in Fig. 1A) by pressing one of two response buttons on the test panel in front of them. Contrast thresholds were then tested, using this procedure, by varying grating contrast under a staircase procedure that increased contrast with each error, and decreased contrast with each three correct responses. Grating patches were 3° by 3° in size, of moderate spatial frequency (2 c/deg.) and their contrast was windowed by a two-dimensional gaussian function with sigma 1.2°. For monkeys A (TEO + TE), B (TEO + TE) and C (TE + CC), the Gabor stimuli were presented at an eccentricity of 4° in the lower left or lower right visual fields. For monkey D (bil. TEO + TE), the stimuli were presented at the center of the display.
Chromatic Contrast Sensitivity
The stimuli (not shown) were Gabor patches of isoluminant grating, identical in size and spatial frequency to the achromatic stimuli described above. Two types of isoluminant chromatic contrast thresholds were measured, with red–green modulated gratings, and with tritanopic (approximately yellow–blue) gratings. As above, the stimuli were 3° by 3° in size, and were shown at an eccentricity of 4° from the fixation target in the lower right or lower left visual field for monkeys A (TEO + TE), B (TEO + TE), and C (TE + CC), and at the center of the display for monkey D (bil. TEO + TE).
Illusory Contour Orientation
The monkeys discriminated the horizontal or vertical orientation of illusory contours made from a set of six concentric circles. Figure 1B shows a horizontal illusory contour stimulus. The three center circles were black, and the outer three circles were of progressively lower contrast. Stimuli were 3° in diameter, and were presented at the same locations as described above for grating stimuli.
After completion of the above testing of visual capacities, the three monkeys with unilateral IT lesions were trained to match shape and color stimuli in a match-to-sample task, in which sample and test stimuli were presented for 0.5 s each, and onset of the test stimulus occurred at offset of the sample stimulus. In alternating sessions the monkeys matched shape stimuli, and color stimuli. As above, the monkeys maintained fixation during testing, and the test stimuli were presented at 4° eccentricity in either the right or left visual field during each test session. Color stimuli were circular patches of highly saturated yellow, green, blue and red, and the shape stimuli tested are illustrated in Figure 8. Stimuli were 3° in diameter and sample and test stimuli were presented at the same location.
At the conclusion of testing, the four monkeys were overdosed with Pentobarbital and perfused through the heart with saline, followed by fixative (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4). Their brains were removed, and 50 µm thick frozen coronal sections were obtained. Alternate sections were stained with Cresyl Violet for Nissl substance to map the extent of the lesion. The stained coronal sections were matched with those from a standard normal brain at 0.5 mm levels throughout the antero-posterior and medio-lateral extent of the lesions, and the lesions were then mapped onto these standardized sections. The dorsal lateral geniculate (LGN) and medial pulvinar nuclei were also examined for cell loss and degeneration.
The ablations were largely as intended. Coronal reconstructions for each animal are shown in Figure 2. Slight damage to the superior temporal gyrus was seen in monkey B (TEO + TE) (arrow in Fig. 2), likely due to reopening the dura to make the second stage lesion.
In monkey C (TE + CC), the corpus callosum was transected in its entirety, caudally from the splenium through the rostral part of the genu. The anterior commissure was also cut completely. There was no observable extra damage to neighboring structures, except for possible slight damage to the fornix, where it was transected to access the anterior commissure. Degeneration was found in the ventral lateral part of the medial pulvinar as described previously (Huxlin et al., 2000; Britten et al., 1992; Dean and Weiskrantz, 1974). In all four cases, no degeneration was observed in the LGN.
Initial Post-operative Behavior
No disruptions of behavioral control were seen in any monkeys after unilateral IT lesions. Following the bilateral lesion, monkey D (bil. TEO + TE) showed severely disrupted visual behavior, which suggested that it did not recognize the technician, other monkeys or familiar food objects (e.g. vitamin tablets or bananas). By 1 week after surgery, its behavior had largely returned to normal, although competitive social interactions with other monkeys did not return to pre-surgery level.
Effects of Unilateral and Bilateral IT Lesions on Visual Capacities
Figures 3 and 4 show that, unlike bilateral IT lesions, which reduced both luminance and chromatic contrast sensitivity, unilateral IT lesions reduced only chromatic contrast sensitivity. Figure 3 shows the effect of the lesions on luminance contrast sensitivity. The contrast sensitivity of monkey D (bil. TEO + TE) was severely affected by the bilateral IT lesion, dropping by more than a factor of five. However, no significant loss of sensitivity was seen in any of the three monkeys with unilateral lesions in comparison of contra- to ipsi-lesional visual fields (Fig. 3). Furthermore, in no case did a unilateral lesion cause a decrease in sensitivity. Sensitivity either increased when tested after a lesion, or remained stable in a few cases where post-lesion testing followed closely after pre-lesion testing (e.g. monkey A in Fig. 5). Subsequent addition of a callosal and anterior commissure section in monkey C (TE + CC) had no effect on contrast sensitivity.
Figure 4 shows results for blue–yellow chromatic contrast sensitivity. The sensitivity of monkeys with unilateral lesions was reduced 0.2–0.3 log units in the contralesional visual field relative to the ipsilesional field, and that of monkey D (bil. TEO + TE) was reduced ∼0.4 log units by the bilateral lesion. Addition of section of the callosum and anterior commissure in Monkey C caused no significant change in the relative contralateral versus ipsilateral sensitivity. Figure 5 shows the time course of sensitivity changes in monkey A (TEO + TE) from before placement of a one stage TEO + TE lesion (Pre), across daily measures of sensitivity immediately after the lesion (Session post-lesion), and with final stable thresholds (Post). Both luminance and chromatic sensitivity were initially decreased in the contralateral field, but showed rapid improvement. Luminance sensitivity recovered to values that were not significantly changed from before the lesion. On the other hand, the yellow–blue sensitivity of this monkey was decreased sharply after the lesion on the contralateral side, and then showed recovery. Because sensitivity increased in the ipsilateral field, final sensitivity was substantially greater in the ipsilateral than contralateral field.
Examination of the separate effects of TEO and TE lesions in monkey B (TEO + TE), Figure 6, revealed that no color sensitivity loss followed the TEO lesion, thus the full loss resulted from the subsequent, TE lesion. That the TE lesion was the basis of the chromatic sensitivity loss was supported by results in monkeys A (TEO + TE) and C (TE), both of whom showed a substantial loss after a lesion of TE alone (monkey C) or after a combined lesion of TEO and TE (monkey A). The contralateral to ipsilateral sensitivity ration was very close to 100% in all three monkeys before the TE lesion. The lesions caused a 0.2–0.3 log unit loss of chromatic sensitivity.
Similar results were obtained for red–green chromatic sensitivity (not shown). Unilaterally lesioned monkeys showed losses of ∼0.1 log units and the bilaterally lesioned monkey D showed ∼0.4 log unit loss.
Performance on the illusory contour discrimination is shown in Figure 7. The performance of the unilaterally lesioned monkeys remained near perfect in both ipsi- and contra-lesional visual fields. Additional section of the corpus callosum and anterior commissure in monkey C (TE + CC) was without effect on this performance. However, the performance of monkey D (bil. TEO + TE), was devastated, dropping from 95% correct to near chance.
Effects of Unilateral IT Lesions on Visual Matching-to-sample Performance
All three monkeys learned to perform shape (Fig. 8A) and color (Fig. 8B) matching in their ipsilesional fields, with percent correct performance near 90%. Their rate of learning and final performance was similar to that of three other control (non-lesioned) monkeys recently trained on this task (not shown). However, in their contralesional field, they were unable to perform much above 70% correct for the matching of color stimuli, and above chance for the shape–shape comparisons. Section of the anterior corpus callosum and anterior commissure in monkey C (TE + CC) did not further exacerbate this deficit in performance.
The most striking result of this study was that unilateral IT cortex lesions produced contralateral visual and perceptual loss, with no loss seen ipsilateral to the lesions. Furthermore, the contralateral effects of unilateral lesions were not exacerbated by sectioning the callosal and anterior commissure fibers that provide input to IT neurons from the ipsilateral visual field. Whereas bilateral IT cortex lesions disrupted all tested visual abilities, unilateral lesions caused no change in luminance contrast sensitivity and the discrimination of illusory contours. The unilateral deficits observed in chromatic vision appear to parallel the contralateral color vision loss that has been often observed in humans with localized lesions in the lingual and fusiform gyri. Finally, we found that unilateral IT lesions did not disrupt the perception of illusory contours that is reliably disrupted by lesions of cortical area V4 (unpublished), even though V4 provides much of the input to IT cortex.
Unilateral IT Lesions Produce Contralesional Visual Loss
That unilateral IT lesions caused contralateral loss of visual thresholds and visual matching performance seems at odds with the bilateral extent of some IT receptive fields in the macaque. The portion of the visual field affected by cortical lesions at earlier stages of the macaque visual pathway is always contralateral to the lesion and consistent with the excitatory receptive fields of the damaged neurons. A small lesion in macaque area V1 (Merigan et al., 1993) caused severe loss of vision in a region of the visual field that is likely to encompass only the excitatory centers of receptive fields, and not the much larger receptive field surrounds. Similarly, lesions in macaque area V4 (Merigan, 1996) produced visual loss in a quadrant of the visual field, approaching within 1–2° of the horizontal and vertical meridia (Merigan, 1996; Merigan and Pham, 1998), a region which covers the excitatory receptive field centers of the damaged V4 neurons (Gatass et al., 1988). This loss does not include the surrounds of V4 receptive fields, which extend across the vertical and horizontal meridia of the visual field (Desimone et al., 1993). This failure to find visual impairment ipsilateral to V4 lesions suggests that the visual loss that results from such cortical damage is due primarily to inactivation of receptive field centers.
The receptive fields of TEO neurons are larger than those of V4 neurons, and although they show coarse retinotopic organization, with upper and lower field representations, they show little extension across the vertical meridian of the visual field (Boussaoud et al., 1991). In the present study, in agreement with our earlier findings with bilateral TEO lesions (Huxlin et al., 2000), we did not find any visual loss after unilateral TEO lesions, presumably because we did not test the learning of simple pattern discriminations (Iwai and Mishkin, 1969; Blake et al., 1977), or the effects of distracting stimuli (De Weerd et al., 1999), which are affected by lesions of TEO.
We did find dramatic visual loss in this study following unilateral TE or TEO plus TE lesions in three monkeys. Unlike the receptive fields of TEO neurons, many TE receptive fields extend well across the vertical meridian, and some are bilaterally symmetric (Kobatake and Tanaka, 1994; Op De Beeck and Vogels, 2000). TE receptive fields do show the same falloff in the spatial scale of processing with visual field eccentricity from the fovea that is found in primate visual perception (Cowey and Rolls, 1974; Popovic and Sjostrand, 2001). However, certain stimulus selectivities and response properties (latency, amplitude, etc.) of TE neurons are relatively consistent throughout the receptive field (Sary et al., 1993; Ito et al., 1995). For example, Tovee et al. (1994) found that the response of TE neurons to face stimuli was strongly translation invariant, with no significant change in the neural information about face identity even with fixation more than 7° left or right of the fovea. However, it is well established that even at the level of area TE, cortical neurons have a bias towards the contralateral visual field in terms of center location and receptive field extent (Boussaoud et al., 1991; Op De Beeck and Vogels, 2000). The most dramatic evidence of contralateral domination of response is seen when two stimuli are presented simultaneously on either side of the vertical meridian, the response of TE neurons is dominated by the stimulus presented in the contralateral visual field (Chelazzi et al., 1998; Sato, 1989). These observations, coupled with the findings of the present study, suggest that the contribution of TE neurons to vision is much more substantial in the contralateral than ipsilateral visual hemifields.
Bilateral IT Lesions Cause Substantially More Visual Loss than Unilateral Lesions
The one monkey that received bilateral IT lesions in the present study, as well as monkeys with bilateral IT lesions we have studied previously (Huxlin et al., 2000), showed more substantial and widespread visual loss than monkeys with unilateral lesions. In the present study, bilateral lesions affected all tested visual abilities: luminance contrast sensitivity, red–green and yellow–blue chromatic contrast sensitivity and the perception of illusory contours, whereas unilateral lesion affected only chromatic sensitivity. Similarly non-selective loss of vision has been found previously after bilateral IT lesions (Huxlin et al., 2000). Some visual abilities recover completely (Dean, 1979) or partially (Huxlin et al., 2000) after bilateral IT lesions, but the range of residual visual disruption can remain substantial (Heywood et al., 1995).
The relatively preserved sensitivity after unilateral lesions cannot be due to the maintenance of visual function by the interhemispheric pathways that provide the bilateral receptive fields of IT neurons, since section of the corpus callosum and anterior commissure in monkey C did not exacerbate the visual loss. The greater loss after bilateral IT lesions could in part reflect a generalized disturbance of behavior, such as the Kluver–Bucy syndrome that is often found after bilateral, but never after unilateral IT lesions (Kluver and Bucy, 1997). A second more speculative possibility is that IT lesions cause a disruption in the appearance of visual stimuli, which makes discriminations difficult. Monkeys may eventually learn to compensate for the changed appearance after unilateral IT lesions by comparing stimulus appearance in the affected and intact visual field. Such a mechanism would depend on the post-lesion visual experience of the monkey. Thus, deficits after unilateral lesions would initially be as great as those following bilateral lesions, but would decline with the use of vision.
TEO Lesions Do Not Mimic the Effects of V4 Lesions
Curiously, as reported previously for bilateral lesions (Huxlin et al., 2000), lesions of area TEO, which receives input from areas V2 and V4, did not produce reliable effects of V2 and V4 lesions, the disruption of illusory contour discrimination or grouping performance (Merigan et al., 1993; Merigan, 1996). In addition to providing most of the input to TEO (Boussaoud et al., 1991; Rockland et al., 1994), V4 receives a large feedback connection from TEO (Rockland et al., 1994). V4 also provides input to, and receives feedback directly from, TE (Baizer et al., 1991; Felleman and Van Essen, 1991). If the visual effects of V4 lesions depended on the interruption of top-down influences in the ventral pathway, TEO plus TE lesions should cause the same disruption of performance as V4 lesions. Indeed, both V4 and TEO lesions have been reported (De Weerd et al., 1999) to enhance the disruptive effects of visual distractors on visual discrimination, an effect attributed to impaired attention. However, the disruption of grouping by V4 lesions does not appear to be due to altered attention (Merigan, 2000). Nor is visual grouping disrupted by IT lesions (Huxlin et al., 2000), suggesting instead that this ability involves feed-forward processing that is not sensitive to loss of feedback from IT cortex.
Chromatic Sensitivity Loss with Sparing of Achromatic Vision
In the present study, unilateral IT lesions affected only chromatic sensitivity, whereas bilateral lesions affected all tested visual abilities. Color vision loss has previously been reported following bilateral IT lesions in macaques (Heywood et al., 1988) and to persist even following extensive retraining, given to insure maximal recovery from the lesion effects (Huxlin et al., 2000). The severity of this color vision loss is variable from study to study, ranging from apparently complete color blindness (Heywood et al., 1995), to moderate reductions in hue discrimination (Huxlin et al., 2000). Because we found no post-lesion degeneration of neurons in the lateral geniculate nucleus, this loss cannot be due to damage to the optic radiations, which can be damaged by IT lesions (Huxlin et al., 2000).
Although severe unilateral loss of human color vision, with partial sparing of luminance vision, is commonly reported to follow unilateral cortical lesions, it is not clear if these lesions are comparable to the IT lesions studied in this experiment. Plant (1991) reviewed published cases of unilateral color vision loss, including the studies of Kolmel (1988), and described frequent contralateral, upper quadrant loss of color detection in patients, with substantial sparing of luminance vision. Rizzo and colleagues (Rizzo et al., 1993) tested two patients with color vision loss secondary to cortical lesions and found, like the present study, decreased sensitivity along both red–green and blue–yellow axes, with relative sparing of achromatic vision. More recently, two patients were described with upper left quadrant color vision loss, as well as loss of form vision, following a localized lesion in extrastriate cortex (Merigan et al., 1997; Gallant et al., 2000). It is still not clear how the locus of the cortical lesion in these cases corresponds to the somewhat better demarcated boundaries of V4, TEO and TE in the macaque. The human lesions are typically found in the lingual and/or fusiform gyri on the ventral surface of the brain. This region appears to be anterior to the location identified as V4 in functional magnetic resonance imaging mapping studies (Sereno et al., 1995; DeYoe et al., 1996; Engel et al., 1997), and could correspond to the color responsive area identified in such studies as V8 (Hadjikhani et al., 1998). If human V8 does indeed correspond to macaque TEO, one might expect hemifield losses after lesions of this area, since TEO contains a hemifield representation (Boussaoud et al., 1991). However, human color loss often involves only the upper visual field quadrant (Plant, 1991). Moreover, unilateral TEO lesions in the present study caused no loss of color vision, this was found only after unilateral TE or TEO plus TE lesions. Thus, at present it appears that the unilateral color vision loss in humans is more profound than that seen in macaques in the present study, and may involve a cortical area anterior to V4, but with quadrant representation of the visual field. Thus, the area damaged in humans who show quadrant color loss may not correspond to that damaged in this study.
Learning/Performance of Visual Matching to Sample Was Severely Affected in the Contralesional Visual Field
Visual matching performance was severely affected contralateral to the lesion, especially the matching of shapes, on which the performance of all three monkeys was reduced to near chance. In contrast, performance ipsilateral to the lesion was normal. This unilateral loss stands in marked contrast to the purely bilateral visual deficits found in humans after unilateral temporal lobe lesions (Huxlin and Merigan, 1998). Although this species difference could reflect differences in the organization of temporal lobe areas in macaques and humans, it is more likely that the dramatic unilateral loss observed here was due to the type of testing performed. One important difference between the severely disrupted matching and the minimally disrupted orientation discriminations in this study, was that the latter discriminations were all trained and tested until stability before the lesions were made. Several studies have shown that discriminations learned after IT lesions can be more disrupted than those very thoroughly learned (often termed ‘overlearned’) before the lesions (Gross, 1973). In addition, it may be that visual matching itself is particularly susceptible to disruption by IT cortex lesions (Fuster et al., 1985), as suggested by reports that have emphasized the role of IT in short-term memory (Iversen and Weiskrantz, 1970; Wilson et al., 1972; Dean, 1974). Furthermore, a previous study that combined unilateral IT lesions with callosal section (Vogels et al., 1997), found impaired orientation matching contralesionally, despite prior training on this task, but almost no disruption of orientation discrimination. Our results suggest that future explorations of unilateral temporal lobe damage in humans should also examine visual matching performance. In addition, it may be possible to amplify the severity of unilateral deficits in humans, for diagnostic purposes, by presenting irrelevant distractor stimuli in the unaffected hemifield, in order to maximize cross-midline stimulus competition (Chelazzi et al., 1998).
The authors thank Dr Krystel Huxlin for valuable comments on the manuscript. This work was supported in part by grants EY00898 and P30 EY01319 from NIH and by an unrestricted grant from Research to Prevent Blindness.