Case notes

Patient C.P., a 57-year-old, left-handed (as determined by the Briggs and Nebes Handedness questionnaire—Reference Briggs and Nebes1975) man presented with right hemi-inattention, a non-fluent Broca's aphasia with oral and verbal dyspraxia, and a sublux right shoulder and flaccidity in his right upper limb consistent with a cortico-spinal lesion. His computed tomography (CT) scans revealed bilateral brain lesions (see Figure 1).

Fig. 1 Computed tomographic (CT) scan acquired 6 months post-onset. The selected sections depict bilateral middle cerebral artery distribution infarcts, with the left-sided infarct involving the basal ganglia and anterior limb of the internal capsule. CT images acquired 2 days after the patient was admitted for the acute left infarct indicated that the right lesion had been acquired some weeks beforehand.

Following institutional ethics approval and written informed consent, C.P. was recruited to the present study 6 weeks post-onset, at which time he still showed a rightward neglect during daily ward routine. However, his performance on subtests of the Behavioral Inattention Test (Wilson, Cockburn, & Halligan, Reference Wilson, Cockburn and Halligan1987) showed a more complex neglect that incorporated both contralesional and ipsilesional elements. He marked only 7 (all left-sided) of the 36 lines during the line-crossing task, cancelled only 10 (all left-sided) of all 54 stars during star cancellation, and omitted the rightmost figure during the shape copying task. On the picture scanning task, C.P. reported all details on the left-hand side of pictures, but failed to report at least two details from the right-side of each picture. In stark contrast, C.P. bisected 3 lines (each 20.4 cm long) 2.2 cm (SD = 1.4 cm) to the right of center.

Formal tests of visual extinction were not conducted, and visual field perimetry testing was reported as unreliable.

Experiment 1: Bisection in the Horizontal Plane

The aim of Experiment 1 was to identify a set of stimulus manipulations that reversed the direction of C.P.'s horizontal neglect. Based on studies reviewed above (Humphreys & Riddoch, Reference Humphreys and Riddoch1995; Na et al., Reference Na, Adair, Choi, Seo, Kang and Heilman2000; Riddoch et al., Reference Riddoch, Humphreys, Luckhurst, Burroughs and Bateman1995), we administered a bisection task in which the physical dimensions and configuration of stimuli were manipulated. Assuming that these stimulus changes reversed the direction of bias, it would then be possible (in a second experiment) to re-orient stimuli along the radial plane to determine if a similar reversal occurred during radial bisection.

Method

C.P. was asked to bisect three different types of stimuli; (1) “solid lines” (i.e., those conventionally used in line bisection studies), (2) “symbol lines” composed of horizontally aligned alphanumeric characters and geometric shapes resembling those used by Na et al. (Reference Na, Adair, Choi, Seo, Kang and Heilman2000), and (3) “lines with distracters” in which a solid line appeared amongst shorter, randomly oriented lines. The three types of to-be-bisected stimuli appeared at five different lengths ranging from 20.8 to 23.3 cm. Solid lines were 1 mm high while the symbol lines were 8 mm high. The distracters in the “lines with distracters” stimuli varied in length from 2.5 cm to 10 cm, were 0.5 mm wide, and varied in number from 12 to 24 with an equal number positioned in random orientation to the left/right of center, and above/below center.

Each stimulus appeared black on a white A4 sheet of paper, and was positioned on a desk approximately 30 cm from participants (i.e., within peri-personal space), aligned with the mid-sagittal plane.

Three blocks of trials were administered, one for each stimulus type. Solid lines were first bisected, followed by the symbol lines, and then the lines with distracters. The order of stimuli within each block was randomized, and each stimulus type was bisected 12 times. The experimenter sat directly opposite the patient throughout, and on handing the patient each sheet of paper asked him to “mark the midpoint,” using the pen provided. To provide a normative baseline, 6 female and 5 male, left-handed, neurologically healthy volunteers, mean age 51 (SD 12 years), were also tested.

Results

C.P. mis-bisected the solid lines 2.5 cm to the right, while the symbol lines and lines with distracters were mis-bisected 2.1 cm and 1.6 cm to the left, respectively (see Figure 2a). Bonferroni-corrected pair-wise t tests (α = 0.016) showed that the bisection scores for lines were significantly different to those for the symbol lines, (t(1,11) = 4.7; p = .01) and lines with distracters, (t(1,11) = 7.5; p < .01). The mean difference between symbol lines and lines with distracters did not reach significance, (t(1,11) = 0.8; p > .05).

Fig. 2 Mean bisection accuracy (with standard deviation bars) in (a) Experiment 1, and (b) Experiment 2. In the top panel, negative displacements represent leftward error and positive displacements represent rightward error. In the bottom panel, negative displacements represent proximal error and positive displacements represent distal error.

Ninety-five percent prediction intervals derived from the normative control sample indicated that all stimuli were bisected with an accuracy that fell outside normal limits, confirming the pathological nature of C.P.'s horizontal bisection performance. The controls bisected all stimuli relatively accurately (within 0.3 cm of mid-line) showing no significant pair-wise differences (all t scores < 1.5).

Discussion

As anticipated, the direction of bisection error in patient C.P. could be reversed by altering the configuration of stimuli, solid lines were mis-bisected to the right, while the other two stimulus forms were mis-bisected to the left. The change in directional bias could not be explained by a change in responding hand (Halligan, Manning, & Marshall, Reference Halligan, Manning and Marshall1991), start side (Humphreys & Riddoch, Reference Humphreys and Riddoch1995), line length (Monagan & Shillcock, Reference Monagan and Shillcock1998) or upper/lower visual field placement (Geldmacher & Heilman, Reference Geldmacher and Heilman1994), all of which were held constant. Although we cannot determine which particular stimulus property (e.g., number of perceptual items, physical dimensions) reversed the direction of bisection error, we have nevertheless identified conditions that capture C.P.'s bi-directional neglect within the horizontal plane. We could now test whether an analogous deficit occurred when the stimuli were oriented along the radial axis.

General Discussion

The presence of bi-directional neglect within the horizontal plane has been taken as a compelling endorsement of the view that neglect does not impose rigid left/right boundaries on the allocation of attention, but rather imposes boundaries that, although typically skewed toward one side of space, are malleable and dynamic in nature (Kinsbourne, Reference Kinsbourne1993) The data from patient C.P. indicate that this dynamic, and somewhat dramatic, interplay between directionally opposed vectors can also occur within radial peri-personal space.

Given that the individual had bilateral brain lesions, there is no fool-proof way to discern the specific contribution of either hemisphere. On one hand, the right (and by association distal) neglect could have stemmed from the left-sided lesion, while the left (and by association proximal) neglect could have stemmed from the right-sided lesion. On the other hand, the lesion distribution seen in the right hemisphere gives reason to believe that it was primarily responsible for the direction of horizontal, and by association radial, bias, with the left hemisphere lesion incidental.

We slightly favor the latter account because the frontal distribution of C.P.'s right-sided lesion is consistent with a disinhibition of the contralesional approach tendency (sometimes referred to as the “visual grasp”). Such disinhibition is deemed to release approach-related processes in occipital-parietal regions (Butters et al., 1988; Kwon and Heilman, Reference Kwon and Heilman1991) which in the present case could have invoked a leftward bisection bias. It is possible that the lines were less effective than the other stimuli in eliciting a right hemisphere visual grasp, which in turn allowed the inherently right (and by association distal) attentional bias of the left hemisphere to dominate that of the injured right hemisphere. The influence of the left hemisphere may have been further strengthened by the fact that that the line stimuli probably placed lighter demands on right hemisphere processes concerned with visuo-spatial search and scanning (see Corbetta & Shulman, Reference Corbetta and Shulman2002). It may also be relevant that, while the lines were most likely coded as a single perceptual element, the other stimuli were most likely treated as containing multiple perceptual items. The switch from single to multiple item displays elicited the same horizontal reversal in patient J.R., who was also left-handed and showed a very similar lesion distribution (Humphreys & Riddoch, Reference Humphreys and Riddoch1995). The authors attributed his left and right neglect to damaged stimulus encoding processes in right fronto-parietal and left occipital-parietal regions, respectively.

These links between left and proximal space on one hand, and right and distal space on the other, support the idea that representations of left and proximal space are mapped to one part of the brain (and accordingly neglected together), while representations of right and distal space are mapped to another (and likewise neglected together). To this end, we note that C.P.'s neglect of leftward and proximal space during line bisection has been reported in other individuals with bilateral brain injury (e.g., Mennemeier, Wertman, & Heilman, Reference Mennemeier, Wertman and Heilman1992), and contrasts with the rightward and distal neglect shown by those with different patterns of injury (e.g., Coslett et al., Reference Coslett, Schwartz, Goldberg, Hass and Perkins1993). C.P.'s bisection error is also consistent with the tendency of young, neurologically healthy adults to bisect lines in a leftward direction when placed relatively proximally, and in a more rightward direction when placed more distally (Varnava et al., Reference Varnava, McCarthy and Beaumont2002). According to Previc (Reference Previc1990), these spatial associations map to different cerebral hemispheres, such that the right hemisphere is specialized for both left hemi-space and near-vision functions (e.g., limb co-ordination, visual-vestibular integration, global perceptual analysis) while the left hemisphere is specialized for both right hemi-space and far-vision functions (e.g., recognition, local perceptual analysis). The behavior of patient C.P. provides particularly strong support for these associations because his neglect afflicts all four sectors of horizontal and radial space (i.e., left/proximal and right/distal). This contrasts with the “two sector” biases reported in all others who show co-incident horizontal and radial neglect (i.e., left/proximal or right/distal). These cases constitute a weaker form of evidence because they are also consistent with the contrary view that, in any given individual, only one region of horizontal space can be coupled with one region of radial space.

To conclude, we report that the direction of bisection error in horizontal and radial neglect can be co-reversed following a change in stimulus configuration. This co-reversal provides new evidence for the idea that common mechanisms guide the deployment of attention across orthogonal spatial planes, and endorses the view that left/proximal space is coded separately to right/distal (Previc, Reference Previc1990). Given that patient C.P. shows a predominantly right neglect, is left-handed, and has bilateral brain damage, studies should now seek to estimate the incidence of co-reversals in those with more common sub-types of neglect.