Illusion processing in hemispatial neglect

Neuropsychologia 39 (2001) 611– 625 www.elsevier.com/locate/neuropsychologia

Illusion processing in hemispatial neglect B. Olk a,*, M. Harvey a, L. Dow b, P.J.S. Murphy c a

Department of Experimental Psychology, Uni6ersity of Bristol, Bristol BS8 1TN, UK b Care of the Elderly, Blackberryhill Hospital, Bristol BS16 2EW, UK c Department of Medicine for the Elderly, Bristol Royal Infirmary/Bristol General Hospital, Bristol BS1 6SY, UK Received 25 February 2000; received in revised form 4 October 2000; accepted 1 November 2000

Abstract Twelve patients with hemispatial neglect and two control groups were tested to examine the effects of the Mu¨ller-Lyer and Judd illusions on bisection behaviour. The studies were designed to investigate whether neglect patients were indeed unaware of the left sides of the illusory figures. In Experiment 1, participants were asked to describe the illusory figures prior to bisection, whereas in Experiment 2, they compared two illusions whose fins, in the critical condition, differed on the left and then performed the bisection. It was found that the illusions worked equally well in all three groups. Interestingly, apart from one exception, almost all neglect patients explicitly reported the left-sided fins in Experiment 1. Only five patients failed to do so but only on an average of 16% of trials. In Experiment 2, six patients made errors in the comparison task but four of these patients did not neglect any left-sided fins in Experiment 1 (with the exception of three overall trials for LC and EdR). This finding seems a good indication that the two tasks differ in their requirements. The comparison task may be perceived as harder as it requires discrimination rather than detection and thus lead to more neglect type errors than the bisection task. In one neglect patient, the illusions consistently failed to work. This patient presented with an occipito-temporal and basal ganglia lesion and the mechanisms responsible for the processing of simple visual features might have possibly been impaired in her case. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: Line bisection; Implicit processing; Awareness; Attention

1. Introduction Dissociations between implicit and explicit processing have been observed in a range of neuropsychological syndromes such as amnesia, blindsight and aphasia [47], and it is clear that stimuli of which patients are not explicitly aware can nevertheless be used implicitly to influence the processing of other information. In relation to the clinical syndrome of hemispatial neglect, a study by Berti and Rizzolatti [7], in which primes were presented to the overtly neglected side, showed that these stimuli facilitated responses to the stimuli presented on the normal side. Additionally, patients who exhibit neglect dyslexia seem to process information on the neglected side as they substitute letters on the left side of words and form alternative words of about the * Corresponding author. Tel.: +44-117-9288581; fax: + 44-1179288588. E-mail address: [email protected] (B. Olk).

same length as the stimulus word [21]. In a study by Peru et al. [42] pre-attentive processing was also observed for the perception of chimeric figures. For overviews regarding this issue see Ellis et al. [16], Karnath and Hartje [31] and Driver and Mattingley [15]. Interestingly, the subject of implicit processing in hemispatial neglect can also be addressed with perceptual illusions such as the Mu¨ller-Lyer and Judd illusory figures. Two such studies, one single case [46] and one group study [34], have so far been carried out. Both studies showed that the rightward bisection error typically shown by neglect patients could be altered by the illusions, in that bisections were displaced away from the baseline bisection error and shifted towards the outward-projecting fin and away from the inward-projecting fin. However, although this effect was clearly shown by Ro and Rafal’s [46] patient, Mattingley et al.’s [34] group result proved less consistent: when presented with 100 mm lines neglect patients exhibited illusory effects in every condition except for unilateral

0028-3932/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 0 0 ) 0 0 1 5 1 - 2

612

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

left-sided inward-projecting fins; on 150 mm lines their bisection performance was influenced significantly by unilateral left-sided inward-projecting fins, unilateral right-sided outward-projecting fins and bilateral fins projecting to the right side, but for 200 mm lines no illusory effects were shown for unilateral left-sided fins, but only for unilateral right-sided outward-projecting fins and bilateral fins projecting to the left. Also, there were no significant differences between patients’ bisection performance in the unilateral when compared to the bilateral conditions. If fins on both sides influence bisection behaviour one would expect patients to make greater bisection errors in the bilateral conditions. Mattingley et al. [34] argue that this lack of bisection difference between unilateral and bilateral fins cannot be explained by the phenomenon of extinction since extinction involves competitive interactions between separate objects. In the illusory figures used, fins and shaft were joined and formed a single object. It is thus not clear why inconsistent effects were found in this group study as such illusions ‘‘…have long been acknowledged to reveal hard wired circuits in the human visual system which mediate early vision’’ [46]. Since hemispatial neglect is generally believed to be a higher level disorder [25,36,45] these inconsistent findings are surprising and worthy of further investigation. Indeed, Mattingley et al. [34] attributed the absence of significant illusory effects for left-sided outward-projecting fins at 200 mm lines to a reduction in the relative magnitude of fins as a function of increasing line length, as it had already been observed that the extent of illusory effects declines with increasing line length with outward-projecting but not inward-projecting fins [3]. When examining the individual bisection performance of their patients, they also found that some patients showed illusory effects of left-sided inward-projecting fins similar to those of controls, while others exhibited effects opposite to these. According to the authors these contrasting performances may reflect a qualitative difference in visual information processing of the left end of the line depending on type and location of lesion and time since injury, although they exclude directional hypokinesia as a possible cause, arguing that their data indicate significant differences in bisection errors as a function of stimulus type, a finding that would not be expected with such an impairment. In terms of implicit processing, the authors argue that the fins on the neglected left side did sometimes influence the patients’ bisection behaviour, despite the fact that they appeared not to perceive these. Unfortunately, Mattingley et al. [34] did not strictly test the lack of explicit awareness of left-sided fins in every patient: patients were simply asked to describe the figures on a pseudorandomized basis (only patient 1 and 2 had to describe all stimuli) after bisection. There is, thus, no hard evidence that the left side of the stimulus was indeed not explicitly processed.

This point was more stringently tested in Ro and Rafal’s [46] experiment. To ensure the lack of explicit awareness of features on the left, their patient was first asked to complete a comparison task in which fins where either the same on the right and left, different on the right, or different on the left. The patient correctly judged ‘same’ for 23 out of 24 same pairs and ‘different’ for 11 out of 12 pairs, which differed on the right. For those pairs of illusory figures that differed on the left side only, however, she answered that they were the same for all 12 pairs. She was then asked to bisect the Mu¨ller-Lyer and Judd illusions and it was found that bisection errors were significantly affected by both illusions and that fins on the left had as much influence on the bisection errors as features on the right. However, although Ro and Rafal [46] showed that their patient neglected the left-sided fins in the comparison task, it can not be ruled out that she was not explicitly aware of the left-sided fins in the bisection task. Two stimuli had to be attended to and discriminated in the comparison task, whereas only a single figure was present in the bisection task. Since task requirements are not the same in the two tasks, explicit attention may be distributed in different ways, possibly resulting in relatively larger neglect in the two stimuli as opposed to the single stimulus condition. The purpose of the present two experiments was thus threefold: (a) to test patients with hemispatial neglect for explicit awareness prior to bisection in every trial of the experiment and then (b) to study the impact of explicitly neglected features on the left on the perception of the illusory figures. As shown in the single case study by Ro and Rafal [46] we expected the illusions to work in neglect patients even if they proved unaware of the fins, and we wanted to show this in a group study, as the Mattingley et al. [34] experiment yielded inconsistent results. In our first study, we used the same illusions as Mattingley et al. [34] but only one shaft length with the fin length being 36% of the shaft length. This was done to ensure that the ratio of fin length to shaft length was optimal to induce very strong illusory effects [19]. In this study, we also included the original Mu¨llerLyer figures in order to replicate the effect of perceived line length observed in Ro and Rafal’s patient. In an additional manner, in the second investigation, we essentially replicated the Ro and Rafal experiment with a comparison and subsequent bisection task. This was done to investigate if the same group of patients may possibly show more neglect in the comparison task than in the task we gave them, i.e. the description of each illusion prior to bisecting it. Finally (c) we also included a patient control group in both experiments to be able to distinguish between the specific effects of neglect on illusion perception and the effects of right hemisphere lesions per se. Previous studies [2,28] have found similar illusory effects for right hemisphere-lesioned patients

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

613

handed according to self-report and as assessed by administration of the 12-item ANNETT Handedness inventory [1]. No significant age differences were found between the groups (one-way analysis of variance (ANOVA)). The groups differed in years of education, however (one-way ANOVA, with controls spending longer in education than the patient groups which did not differ from each other). The two patient groups differed with regard to prevalence of hemianopia but not in relation to the presence of hemiparesis ( 2 tests) nor the time between stroke onset and testing (one-way ANOVA). Patient details are given in Table 1.

compared to healthy controls. Nevertheless, a study by Grabowska et al. [18] revealed stronger illusory effects in right hemisphere-lesioned patients compared to healthy controls.

2. Experiment 1

2.1. Method 2.1.1. Subjects Two groups of patients took part in the experiment: 12 patients with unilateral right hemisphere infarct (mean age=74.1, SD= 5.4; 7 male, 5 female) whose individual performance proved well outside normal limits on the formal subtests of the Behavioural Inattention Test (BIT) [52] and thus displayed hemispatial neglect, and 12 patients with unilateral right hemisphere infarct (mean age= 71.4, SD= 7.9; 6 male, 6 female) who did not show hemispatial neglect at the time of testing and who, according to the reports of the clinical staff, were reported never to have shown signs of neglect (RCVA control group). CT scans were available on all of the patients. Fifteen healthy control subjects (mean age=71.7, SD=3.6; 9 male, 6 female) with no appreciable medical, neurological, or psychiatric history were also tested. All subjects were right-

2.1.2. Neuropsychological tests The two patient groups were tested on three verbal subtests (Information, Vocabulary, and Digit Span) and three performance subtests (Picture Completion, Block Design, and Object Assembly) of the WAIS-R [51]; the New Adult Reading Test (NART) [39]; the Very Short Minnesota Aphasia Test [43]; the Benton Visual Form Discrimination Test (VFDT) [6] and the formal subtests of the BIT [52]. Healthy control subjects were assessed with the NART only. The groups differed with regard to the NART scores with healthy controls giving a higher score than the other two groups, who did not differ from each other (one-way ANOVA). In the neuropsychological tests,

Table 1 Clinical details and test performances of patients with and without (RCVA controls) hemispatial neglect Patients

Lesion location

Poststroke (in weeks)

Hemianopia

BIT

NART

Neglect KG BM HJ CP DS ER JC EdR KA CL JR LC

R R R R R R R R R R R R

parieto-occipital fronto-parietal fronto-parietal parietal fronto-parieto-temporal parietal fronto-parietal temporal MCA territory deep white matter MCA territory, R pons occipital, inf. temporal, basal ganglia

16 35 21 9 14 84 7 14 8 5 3 165

Yes Yes Yes No No No No Yes Yes No No Yes

126 67 108 98 86 116 19 110 101 43 112 115

114 95 114 114 119 122

RCVA IB BT DB LA ED MH MO MiH AW JP RD AQ

R R R R R R R R R R R R

medial thalamus, internal capsule parietal parietal centrum semiovale basal ganglia, centrum semiovale superior parietal basal ganglia temporo-parietal fronto-parietal thalamus, internal capsule parietal fronto-parietal

4 14 64 3 11 88 16 207 175 4 3 7

Yes Yes No No No No No No No No No No

133 136 134 140 143 140 133 139 135 145 143 144

98 104 117 111 109 114 110 100 105 96 114 107

a

Unassessable due to strong neglect.

a

101 112 110 95 101

614

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

the neglect patients attained significantly lower scores in the BIT, VFDT and the Object Assembly subtest of the WAIS-R when compared to the RCVA patients (one-way ANOVA). These are the tests most likely to reveal neglect type errors and the results are thus not unexpected. Patient groups did not differ in the Short Minnesota Aphasia Test and the other five WAIS-R subtests. The differences between the healthy controls and the other two groups in terms of NART score and years of education were not judged to be significant as the intended experiments were meant to tap into low level processes. This was clearly confirmed by the results. The hemianopia issue is dealt with in the discussion of Experiment 1. 2.1.3. Experimental test materials and procedures Stimuli consisted of horizontal lines with either unilateral inward- or outward-projecting fins on the left or right end of the line, and horizontal lines with bilateral inward- and/or outward-projecting fins (Judd and Mu¨ller-Lyer illusions). Shaft length was 200 mm. Fins were 72 mm long, i.e. 36% of the shaft length to ensure that the ratio of fin length to shaft length was optimal to induce strong illusory effects ([19]). The fins formed a 43° angle with the shaft in the inward-projecting fins condition and a 137° angle in the outward-projecting fins condition. Shaft and fins were 1 mm wide. Two hundred millimetres long lines without any fins were included to serve as a baseline condition. All stimuli were drawn with black ink and placed horizontally and centrally on a separate sheet of landscape oriented A3 paper (see Fig. 1 for example of stimuli used). The experimental session consisted of 72 trials, eight trials per stimulus type. The sequence of trials was pseudorandomized. A strict procedure was chosen to assess explicit awareness of the left-sided fins. At the beginning of the experiment, all participants were told that the stimuli consisted of a horizontal line and that an arrowhead might sometimes appear at one or both ends of the line. They were instructed to describe the stimulus and to tell the experimenter on a trial to trial basis if there was/ were arrowhead/s before they proceeded to bisect the shaft. It is worth noting that the subjects were not cued to look at the left side of the shaft, they were simply given this general instruction and then reminded on every trial to tell the experimenter what they saw. All subjects were seated at a table opposite to the experimenter, who ensured that the subject’s body position remained constant throughout the experiment. They were instructed to describe the illusion and then to centrally bisect the figure as accurately as possible. Subjects used their right hand proceeding through all the sheets. Head and eye movements were not restricted. The viewing distance was : 45 cm from the illusion that was to be bisected.

Fig. 1. Stimuli used in Experiment 1: (a) baseline; (b) unilateral left-sided inward-projecting fin; (c) unilateral right-sided inward-projecting fin; (d) unilateral left-sided outward-projecting fin; (e) unilateral right-sided outward-projecting fin; (f) bilateral inward- and outward-projecting fins to the right (Judd-illusion); (g) bilateral inward- and outward-projecting fins to the left (Judd-illusion); (h) bilateral inward-projecting fins (Mu¨ller-Lyer illusion); (i) bilateral outward-projecting fins (Mu¨ller-Lyer illusion).

2.2. Results 2.2.1. Illusions 2.2.1.1. Statistical analyses. We investigated whether the illusory figures were effective and worked in a similar way in all groups by measuring the deviation of the bisection mark from the true midpoint of the shaft for each subject and each condition to the nearest millimetre across the eight instances. Errors to the right of the midpoint were given a positive value and those to the left a negative value. As there is now evidence that neglect patients with and without hemianopia may perceive line length and the subjective straight ahead differently [14,17] an initial two-way ANOVA with group (neglect patients with hemianopia/neglect patients without hemianopia) as a between factor and stimulus type as a within factor was run. There was, however, no main effect of group (F(1,10)= 0.56, P\0.05) nor an interaction with group (F(8,80)= 1.07, P\ 0.05) showing that neglect patients with and without hemianopia did not differ in the way they bisected the illusions. Therefore subsequently, these data where combined to allow a more powerful analysis: the mean error scores were subjected to a two-way ANOVA with group (healthy controls, RCVA controls and neglect

B. Olk et al. / Neuropsychologia 39 (2001) 611–625 Table 2 Results of paired comparisons used for Experiments 1 and 2

patients) as a between factor and stimulus type (baseline, bilateral inward, bilateral outward, bilateral left, bilateral right, unilateral right inward, unilateral right

615

outward, unilateral left inward, unilateral left outward) as a within factor. Analysis of the mean error scores showed a significant main effect of group (F(2,36)= 5.69, P=0.007) and stimulus type (F(4.947,177.74)= 162.25, PB0.001) as well as a highly significant interaction (F(9.87,177.74)= 2.8, P= 0.003). To further examine these effects, separate ANOVAs were calculated for each group. The means of the six illusory conditions (Fig. 1b–g) were compared with the mean of the baseline condition. Comparisons were also made between the critical unilateral and bilateral trials and between the two Mu¨ller-Lyer figures. This restricted set of planned comparisons was conducted by means of paired-samples t-tests, which are listed in the form of effect tables in Table 2. All paired t-tests run in this and subsequent analyses were Bonferroni corrected. The ANOVA for the healthy controls revealed a significant main effect of stimulus type (F(3.5,49.61)= 145.76, PB 0.001). Paired sample t-tests showed significant deviations from the baseline (− 0.88 mm) for all unilateral figures and the bilateral Judd illusions (Fig. 2). There was no difference between the performances in the Mu¨ller-Lyer illusion (Fig. 2). The comparison of unilateral and bilateral figures showed that the illusory effect was significantly stronger for the bilateral figures (Fig. 2). Compared to the healthy controls, RCVA patients without neglect exhibited a weak bisection bias to the right (3.42 mm) but the effects of the illusory figures showed the same pattern as for the healthy subjects: There was a main effect of stimulus type (F(4.31,47.43)=83.77, PB 0.001) and paired samples t-tests showed that bisection errors differed significantly from the baseline performance in all conditions (Fig. 3). The difference between the Mu¨ller-Lyer figures was not

Fig. 2. Bisection errors (in mm) of the healthy control group, separately for each stimulus type. Positive numbers indicate rightward and negative numbers leftward bisection errors. Ref. Table 2 for specific comparisons made.

616

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

Fig. 3. Bisection errors (in mm) of the RCVA patient group, separately for each stimulus type. Positive numbers indicate rightward and negative numbers leftward bisection errors. Ref. Table 2 for specific comparisons made.

significant and when the Bonferroni correction was applied, the illusory effect of the unilateral outward pointing fin on the left also no longer remained significant. However, effects were significantly larger in bilateral as compared to unilateral figures (Fig. 3). In the neglect group, as expected, bisection judgements were shifted more towards the right than in the RCVA control group (16.76 mm). However, there was again a significant effect of stimulus type (F(8,88)= 23.11, PB 0.001) with bisection performance differing from the baseline condition in all illusory figures (Fig. 4, and see also Table 3 for individual break down of each neglect patient’s performance). This time, however, bisection performance between the two Mu¨llerLyer figures proved significant: smaller errors were made in the inward-projecting compared to the outward-projecting figure (Fig. 4). Of the two unilateral/bilateral pairs, the comparison between the unilateral right outward-projecting figure and the Judd illusion with two rigthward-projecting fins (Fig. 4) proved significant. However, when the effects were Bonferroni corrected, of the unilateral fins, only the right inward pointing fin still produced a significant effect and none of the unilateral/bilateral pairs remained significant. This could indicate that the left fins worked less well in the neglect group. To determine more sensitively whether left-sided fins were in fact as effective as right-sided fins in all groups and whether bilateral fins were more effective than unilateral ones, especially in the neglect group, the difference in bisection performance between the baseline condition and each stimulus condition was calculated for each condition, separately for each group. These difference scores were then subjected to a two-way ANOVA with the factors group and stimulus type (with this analysis the necessary planned comparisons are significantly reduced). Both factors reached significance (group: F(2,36)=4.97, P =0.012; stimulus type:

F(4.4,158)= 34, PB 0.001) but the interaction between these factors proved insignificant, indicating that the illusions worked in the same way in all the three groups. Separate ANOVAs for each group confirmed the effectiveness of the illusory figures with significant effects of this factor (healthy controls: F(2.7,37.3)= 20.3., PB 0.001; RCVA group: F(2.8,21.6)= 23.1, PB 0.001; neglect group: F(7,77)= 5.3, PB 0.001). Selected paired t-tests (h-adjusted) were then performed individually for each group comparing unilateral inward-projecting fins on the left and right and also unilateral outward-projecting figures on the left and right (Table 4). There were no differences between these for either group, indicating that fins on the left side were as effective as fins on the right side. Paired t-tests were also carried out to compare the effectiveness of bilateral and corresponding unilateral figures (Table 4). The illusion with the bilateral leftward-projecting fins did not yield significantly greater effects than the unilateral figure with the right inward-projecting fin in neither the healthy control group nor the neglect group although it did for the RCVA controls. The illusory effect was, however, greater for all groups with the bilateral figure with the rightward-projecting fins compared to the unilateral figure with the right outward-projecting fin, proving the additional effectiveness of the left fin of the bilateral figure. All effect sizes are listed in Table 4. In summary, although neglect patients showed a strong rightward bias, their bisection performance was altered by the illusions in that judgements were shifted away from the inward-projecting fins and towards the outward-projecting fins. Although the initial analysis gave an indication that left fins might possibly be less effective in the neglect group, analyses of the difference scores showed that the effects of unilateral leftward fins were as strong as the effect of unilateral rightward fins and that illusions worked in the same way in all the

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

617

Fig. 4. Bisection errors (mm) of the neglect patient group, separately for each stimulus type. Positive numbers indicate rightward and negative numbers leftward bisection errors. Ref. Table 2 for specific comparisons made.

three groups. Only the bilateral figure with the rightward-projecting fins compared to the unilateral figure with the right outward-projecting fin, was shown to have additional effectiveness of the left fin of the bilateral figure for all three groups. The other unilateral/bilateral comparison only proved more effective for the RCVA control group. The effect of perceived line length induced by the Mu¨ller-Lyer illusion was observed for the neglect patient group only.

2.2.2. Indi6idual results As demonstrated in Table 3, illusions worked in the expected directions for most neglect patients. There were occasional conditions in which the left-sided fins did not have an effect, but this proved the case for individuals in the control groups as well. For patient CL, the unilateral inward-projecting fin on the right shifted the bisection to the right whereas the unilateral inward projecting fin on the left biased bisection slightly to the left. (Note that in the unilateral figures with outward-projecting fins the predicted illusory effect is in the same direction as a possible cueing effect.) The results for patient LC (Table 3), however, differed markedly from those of the other patients in that the illusion effects varied quite considerably and clearly did not work: although some of the illusions biased behaviour in the expected direction, some had no effect and others worked in the opposite direction. 2.2.3. Explicit awareness Apart from studying the illusionary effects, the second aim of this study was to investigate if the neglect patients were indeed unaware of the left-sided fins in this task and if so, were nevertheless influenced by them. All of the healthy controls and the RCVA controls completed the stimuli description perfectly,

whereas the neglect patients made errors in that they failed to report the left fins (no incorrect descriptions were given). Five out of the twelve patients reliably reported the left-sided fins on every trial. Two further patients failed to report a left fin once and twice, respectively (Table 5). The individual results of the remaining five patients who did occasionally not report the left-sided fin, were studied in more detail to investigate if the left fins had an effect in those trials: Patient HJ failed to report the left fins 11 times (22.92%) but the bisection deviations in those trials were always within the range of the illusions with congruent right unilateral fins. For example, the leftsided fins were not reported in a figure with bilateral Table 3 Mean bisection errors (in mm) of all 12 neglect patients for each stimulus typea

a Mean bisection errors and standard deviations (in parenthesis) for each group, separately for each stimuli type are also given.

618

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

Table 4 Results of selected paired comparisons of strengths of effects and strengths of illusory effects (mm) in Experiment 1

Table 5 Number of errors listed separately for each patient and stimulus type. All errors were failures to report the left fin. No incorrect descriptions of the stimulus were given

In all of the five instances (10.42%) where KG did not mention the left fins his bisection errors where within the range of the right fins. In general, left fins were neglected more often in figures with outward-projecting fins on the left than with inward-projecting fins on the left and there was also a tendency to neglect left fins of bilateral rather than unilateral figures. The number of trials with neglected left fins is given in Table 5 separately for each patient and illusion condition. Patient JC was the only patient who never reported the left-sided fins (Table 5). JC showed very severe neglect with bisection performance in the baseline condition proving to be 89 mm to the right of the true centre (Table 3). A numerical look at JC’s data (Table 3) seems to imply that the illusory effects worked in the expected direction in all conditions apart from one illusion which had unilateral fins on the left side projecting leftwards. However, paired t-tests on the nine selected comparisons listed in Table 2 showed that none of these effects were significant neither for the illusions on the right (which he reported) nor for those on the left (which he missed). To summarise, it can be said that the majority of patients consciously perceived the left-sided fins on every trial and said so, whereas four patients were unaware of left-sided fins on a small minority of trials (mean 16%). Only on a small subset of these trials did the illusions prove effective nevertheless. However, the number of trials for which this occurred were far too small to lead to any definite conclusions. One patient (JC) failed to see the left-sided illusions on all trials and although they seemed to have worked nevertheless, the reported effects did not prove significant.

2.3. Discussion

outward projecting fins and the bisection error was 4.5 mm to the right, which is within the range of his deviations in figures with unilateral outward-projecting fins on the right. Patient CL failed to report left-sided fins only nine times (18.75%). For seven of these trials the bisection error was within the range he produced for the congruent unilateral right figure. Twice this was not the case and the bisection judgement seems to have been enhanced by the left fins. BM failed to report the left-sided fins only seven times (14.58%) but in three of these instances her bisection performance was outside the range of her congruent unilateral right illusions and the left fins seemed to have contributed to the bisection error.

2.3.1. Illusions In contrast to the inconsistent illusion effects shown for the neglect group in the Mattingley et al. [34] study, the results of Experiment 1 illustrate that the stimuli effectively induced illusory effects in all groups and that they worked in the same directions in the neglect group as in the two other control groups: For all subjects, the bisection mark was shifted towards outward-projecting fins and away from inward-projecting fins. Although in the initial analyses there was an indication that the left fins might work less effectively than the right fins in the neglect group, this finding was not upheld when difference scores were analysed. However, only for one of the two critical illusions were the effects enhanced when bilateral were compared to unilateral fin conditions. In this case fins on both ends had an impact on the bisection performance in all groups. However, the other bilateral illusion (leftward-projecting fins compared to the unilateral figure with right inward-projecting fin)

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

not only failed to have an enhancing effect in the neglect group but also in the healthy control group. This could indicate that the inward pointing fin alone already produces a very powerful illusion that might not be enhanced any further by adding the second fin. It is generally assumed that the Mu¨ller-Lyer illusion arises from mechanisms at several different levels of processing [13] and that some of these processes operate very early in vision. Support for this argument comes from patient LC. She was the only subject for which only very few illusions were effective while others did not seem to affect bisection performance and some even worked in the wrong direction. The reason for this performance may well lie in the location of the brain lesion of this patient: fairly atypical for a neglect patient, her lesion involved the occipital as well as the inferior temporal lobe and the basal ganglia. It seems possible that the mechanisms which are responsible for the processing of simple visual features were indeed impaired in this patient due to the occipital lesion and that the illusions were thus not processed correctly.

2.3.2. Explicit awareness A good demonstration that low level processes are indeed intact in neglect patients would be clear illusory effects in the absence of explicit processing of the left-sided fins. However, to our great surprise, the neglect patients were, with very few exceptions, always aware of the left fins. They clearly perceived the stimulus figure as a whole and processed it accordingly, despite the fact that all of them had acute and fairly strong neglect (BIT scores in Table 1). This makes it impossible to assess whether early processes are indeed largely preserved in the majority of neglect patients and that the main disorder of the syndrome arises from a disturbance at higher levels of processing [45]. On some trials left-sided fins were not reported. However, since this was the case in a small minority of trials only, the analysis, if in such instances the left fins influenced the bisection behaviour, was difficult: in most of these cases the bisection errors were within the range of the congruent right figures thus not showing the specific effect of left-sided fins. There were only five trials in which the left fins were not within that range and had an effect in the expected direction. Also, JC the only patient who never reported the left fins, showed no significant illusory effects (although they were in the expected direction), so even the limited trials in which the fins were not explicitly processed are inconclusive. In some respects awareness of the left-sided fins is not that surprising as a number of studies have now shown that patients with hemispatial neglect do not so much neglect single stimuli but rather misrepresent them. In a series of studies Milner and colleagues [26,37,38] showed that the typical rightward bisection error of these patients may result from a distortion of the

619

subjective space in that objects on the left are perceived as relatively smaller than objects on the right, possibly as the result of an attentional/representational deficit. Similar findings have also been reported by others [9,29,41]. It is also quite likely that none of the patients tested here (apart from possibly JC) experienced objectbased neglect, which is indeed rare [50] but would almost certainly lead to a neglect of the left-sided fins. To our knowledge nobody has yet tested the effects of the Mu¨ller-Lyer and Judd illusions on patients with identified object-based neglect. It was also found that left fins were neglected more frequently in figures with outward-projecting rather than inward-projecting fins on the left. This may have been due to the further extension of the outward-projecting fins into left space. There was also a slight tendency to neglect left fins of bilateral illusions more than left fins of unilateral figures. Although one might argue that this effect cannot be explained by the phenomenon of extinction since extinction involves competitive interactions between separate objects [34], our instruction to describe the object required the subject to individuate the features. In fact the ends of the shaft could even be judged to be similar either in terms of the direction of the arrows (inward/outward) or in terms of the left or right opening, so extinction may have been a contributor to the (albeit small) effect of greater non-reports of bilateral over unilateral left fins. Since the patients never gave an incorrect description of the figure but only failed to report a fin altogether, there is no way of telling which dimension they used but it is likely one was consistently chosen. This makes an extinction effect even more likely as the phenomenon is believed to be stronger when stimuli are the same on a relevant dimension [4]. It is worth noting that six out of the twelve neglect patients had hemianopia. The initial ANOVA and paired t-tests clearly showed that the presence of hemianopia did not prevent the effective processing of the illusions in the neglect patients but five of these six hemianopic neglect patients belonged to the group of patients who did occasionally neglect the left fins. Although it is expected that hemianopic patients compensate for their hemianopia [5,27,30] and it has been stated that hemianopia does not exacerbate neglect [24] it is still possible that neglect interferes with the compensation which normally occurs in hemianopic patients (in fact the two RCVA control patients with hemianopia never failed to report the left fins). Although head and eye movements were not restricted in this study, the hemianopic neglect patients might have occasionally failed to compensate for their hemianopia and, therefore, not reported the left-sided fins. However, as the illusions clearly worked in these patients and the number of unreported fins was 6ery small, this possible effect must have been very small.

620

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

Fig. 5. Stimuli used for the bisection task of Experiment 2.

Compared to the healthy controls, RCVA patients without neglect showed a weak bisection bias to the right. As reported by Machado and Rafal [32], RCVA patients without neglect also produced overall rightward bisection errors albeit of smaller magnitude than those shown by the neglect patients. The authors conclude that chronic parietal lesions produce an ipsilesional bias even in the absence of clinical signs of neglect and these data seem to confirm that argument although a previous study by Harvey et al. [26] showed that RCVA patients without neglect made small leftward errors. However, in that study the lesions were not confined to the parietal lobe whereas the majority of patients here had lesions with parietal involvement.

3. Experiment 2 In this study we essentially replicated the Ro and Rafal [46] experiment with a comparison and subsequent bisection task to investigate if the same group of patients may possibly show more neglect in the comparison task than in the task we gave them, i.e. the description of each illusion prior to bisecting it. As we have already shown in Experiment 1 that our group of patients shows illusory effects these findings will be presented in a brief manner only (with no figures given) and we will concentrate on the assessment of implicit and explicit processing within and between the two studies instead.

3.1. Method 3.1.1. Subjects Subjects were identical to Experiment 1. 3.1.2. Experimental test materials and procedures The experiment consisted of two tasks: a comparison task to test for explicit awareness and a bisection task. Stimuli consisted of horizontal lines with bilateral inward- and/or outward-projecting fins (Judd and Mu¨ller-Lyer illusions, see Fig. 5, four illusions). Dimensions were the same as in Ro and Rafal’s [46] experi-

ment with a shaft length of 64 mm and fin length of 24 mm (that is 37.5% of the shaft length). The fins formed a 45° angle with the shaft in the inward-projecting fins condition and a 135° angle in the outward-projecting fins condition. As in Experiment 1, shaft and fins were 1 mm wide. Pairs of these stimuli were used for the comparison task (Fig. 6). Each pair was printed centrally on a sheet of landscape oriented A4 paper. Half of the pairs were the same (24) and half were either different on the left (12) or on the right (12). Pairs were presented sequentially in a pseudorandomized order. All subjects were seated at a table opposite the experimenter, who ensured that the subject’s body position remained constant throughout the experiment. Subjects were instructed to report whether the figures in the pair were the same or different. Only half of the trials (same, 12; different on the left, 6; different on the right, 6) were given to the healthy controls since it was assumed that they would show perfect performance on this task. Head and eye movements were not restricted. The viewing distance was : 45 cm from the illusions that were to be judged. For the bisection task each stimulus was placed horizontally and centrally on a separate sheet of landscape oriented A4 paper. The task comprised 20 trials, five trials per stimulus type and the sequence of trials was pseudorandomized. In order to replicate Ro and Rafal’s experiment as closely as possible subjects were simply asked to bisect the figure but not to describe it beforehand. Again, all subjects were seated at a table opposite to the experimenter, who ensured that the subject’s body position remained constant throughout the experiment. Subjects used their right hand proceeding through all the sheets. Head and eye movements were not restricted. The viewing distance was :45 cm from the illusion that was to be bisected.

3.2. Results 3.2.1. Illusions Errors in shaft bisection for each subject and for each condition were measured to the nearest millimetre and averaged across the five instances. Errors to the right of the midpoint were given a positive value and those to the left a negative value. As in the Ro and Rafal [46] study, the crucial comparisons were those between the

Fig. 6. Stimuli used for the comparison task in Experiment 2.

B. Olk et al. / Neuropsychologia 39 (2001) 611–625 Table 6 Performance on the same/different comparison task; listed separately for each patient

illusions differing on the left side. Thus, the performances between the Judd illusion with rightward-projecting fins and the Mu¨ller-Lyer illusion with outward-projecting fins, and between the Judd illusion with leftward-projecting fins and the Mu¨ller-Lyer figure with inward-projecting fins and also between the two Mu¨ller-Lyer figures were investigated (see Table 2 for these comparisons). These comparisons were carried out separately for each group, using paired t-tests for each comparison. In a similar vein to Experiment 1, initially we performed the tests on the neglect patients with hemianopia only. All three comparisons proved significant (t(5)= − 3.11, P =0.044; t(5) =5.47, P= 0.026; t(5)= 2.67, P = 0.003 respective to order above) indicating that the illusions worked in this subgroup. All the neglect patients were then combined in the subsequent t-tests, again to allow more powerful comparisons and all t-tests were Bonferroni corrected. For the healthy controls we found significant differences between the Judd and Mu¨ller-Lyer figures. In the Judd figure with rightward-projecting fins bisection marks were placed 3.6 mm to the right of the true middle as opposed to 0.6 mm to the right in the Mu¨ller-Lyer illusion with outward-projecting fins (Table 2). In the Judd illusion with leftward-projecting fins the mark was placed 3.2 mm to the left of the true middle as opposed to 0.2 to the right in the Mu¨ller-Lyer figure with inward-projecting fins. There was no significant difference between the bisection performance in the Mu¨ller-Lyer figures. For the RCVA control group, we found significant differences in all comparisons. The bisection mark in the Judd figure with rightward-projecting fins was placed 5 mm to the right of the middle and in the Mu¨ller-Lyer figure with outward-projecting fins 1.1 mm to the right (Table 2). In the Judd figure with leftwardprojecting fins the bisection mark was placed 3.8 mm to the left and in the Mu¨ller-Lyer figure with inward-projecting fins 0.3 mm to the right. The difference between the two Mu¨ller-Lyer figures was highly significant. The results for the neglect patients showed a significant difference in all three comparisons (Table 2). In

621

the Judd figure with rightward-projecting fins, the bisection mark was made 5.3 mm to the right as compared to 1.7 mm in the Mu¨ller-Lyer figure with outward-projecting fins. In the Judd figure with leftward-projecting fins, the bisection mark was placed 4.7 mm to the left and in the Mu¨ller-Lyer figure with inward-projecting fins 1.3 mm to the left. The difference between the Mu¨ller-Lyer figures was significant. As expected, the results clearly show that all illusions worked in the predicted directions in all groups although there was no significant difference between the Mu¨ller-Lyer figures for the healthy control group.

3.2.2. Explicit awareness All of the healthy and RCVA controls completed the comparison task perfectly. There were six patients in the neglect group who made mistakes in the comparison task (Table 6) although only two of them showed the large neglect type errors of Ro and Rafal’s patient. We then examined whether the left-sided fins had an effect in these patients: patient JR made one mistake in the comparison task. In her case the illusions worked in the expected direction and were thus in line with the behaviour shown by the neglect patients as a group. Patient KA made three mistakes. Again the illusions worked in the expected way with exception of the Mu¨ller-Lyer figure with outward-projecting fins, which he bisected slightly towards the left of the middle. Patient EdR made five mistakes in the comparison task and illusions worked in the expected directions, apart from the Mu¨ller-Lyer figure with outward-projecting fins which shifted the bisection mark to the left of the true middle. Patient BM made 13 mistakes but all illusions worked in the expected direction. Patient JC made 17 mistakes. All his bisection marks where made to the left of the middle, behaviour possibly caused by the relatively short shaft. In the figure with the inwardprojecting fins his bisection was shifted more to the left than in the figure with leftward-projecting fins, a result in the wrong direction. The other two comparisons, however, produced the correct shift in bisection. Patient LC made four mistakes. In her case, the Judd illusion with the rightward-projecting fins hardly shifted the bisection mark towards the right. The Mu¨ller-Lyer figure with outward-projecting fins also worked in the wrong direction. Contrary to Experiment 1 where, in the limited cases in which the patients failed to report a fin, the illusion induced effects were generally not found; in this experiment, the illusions worked in the expected direction even when the patients failed to perceive them in the comparison task. The direct comparison between the procedures will be addressed in Section 3.3.

622

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

3.3. Discussion 3.3.1. Illusions The comparison of the Judd and Mu¨ller-Lyer illusions showed that the illusions worked in the expected way in all three groups. The bisection judgement in the Judd figure with the rightward-projecting fins was placed to the right of the mark of that of the Mu¨llerLyer figure with outward projecting fins. The bisection mark in the Judd figure with leftward-projecting fins was placed to the left of the mark of that of the Mu¨ller-Lyer figure with inward-projecting fins. These observations show that the fins on the left side influenced the bisection behaviour. As in Experiment 1, patient LC proved an exception to this pattern. The comparison between the two Mu¨ller-Lyer figures reached significance in the two patient groups. Here the bisection mark was placed further to the right in the figure with outward-projecting fins as opposed to the figure with inward-projecting fins. This effect of perceived line length had been observed by Ro and Rafal [46] in their neglect patient and also in Experiment 1 with larger illusory figures but only for the neglect group. The effect of line length on the bisection error is an established result in neglect patients, but less consistently observed in right hemisphere lesioned patients without neglect. It is not clear why this difference proved significant for the RCVA controls in Experiment 2 but not in Experiment 1.

3.3.2. Explicit awareness Six of the twelve patients made errors in the comparison task. Apart from LC, who also showed no illusory effects in Experiment 1, the illusions worked predominantly well in the remaining five patients. The comparison of the results of the assessment of awareness of both experiments shows that there were two patients (HJ and CL) who occasionally failed to report the left fins in Experiment 1 but had no problems in the comparison task in Experiment 2. A possible reason for this may be the different size of the stimulus figures. Stimuli in Experiment 1 were about three times larger than the illusions in Experiment 2 and extended far more into left space, thus rendering a possible neglect of the left-sided fins more likely. Patient BM, who made 13 errors in the comparison task, did not report the left-sided fins in Experiment 1 in 15% of the trials. Patient JC, who made 17 mistakes in the comparison task, never reported the left-sided fins in Experiment 1. These two patients are interesting because they both showed a lack of explicit awareness in Experiment 2 but only JC also showed this in Experiment 1. If both these patients, however, are influenced by left-sided fins in the same way then this would provide strong corroboration for the conclusion

that length illusions are generated just as robustly when individuals have no explicit awareness of the features inducing the illusion then when they do. However, this proved not the case: as already described in Experiment 1, JC’s illusory effects did not reach significance whereas in BM’s case left-sided fins significantly biased her bisection performance for Experiment 1 (4 out of the 6 paired t-tests on the illusions containing a left fin proved significant at the 0.01 level, no paired-comparisons on individual patients performance could be run in Experiment 2 as there were not enough trials per illusion). Interestingly, the remaining four patients who made mistakes in the comparison task, did not neglect any left-sided fins in Experiment 1 (these include LC who made two mistakes and EdR who made one mistake), despite the fact that the stimuli were much smaller in the second experiment, extended less into left hemispace and should therefore have led to less neglect. It is clear that the two tasks differ in their requirements (detection in the bisection and discrimination in the comparison task) and our data indicate that the latter leads to more neglect type errors. It, therefore, does not necessarily follow that patients who neglect the left side in the comparison task will also do so in the bisection task. More indirect support for this argument comes from the finding that when our patients failed to report the left fin in Experiment 1, the illusions also tended not to work (admittedly these were very few trials overall), whereas they did work in Experiment 2 even for the patients who failed the comparison task. The reason for this difference may well lie in the fact that the left side was indeed explicitly processed in the bisection task in Experiment 2. This might also mean that Ro and Rafal’s patient, despite having performed poorly in the comparison task, may well have been aware of the left fins in the bisection task. This interpretation does seem unlikely, however, in view of the errors she made on the comparison task, which were indeed a 100% in the critical comparison. Only two of our patients (BM and JC) produced error rates comparable to this patient and JC did indeed neglect all left-sided fins in Experiment 1 as well, although BM neglected only 15% of those. To really address this issue a larger number of patients with a 100% error rate on the comparison task would have to be identified and then asked to describe the ‘to be bisected’ illusions. Finally, four out of the six patients who occasionally made mistakes had hemianopia. As discussed above this could have contributed to the neglect of left-sided fins; however, it has to be borne in mind that JC who never reported any left-sided fins in either experiment did not have a field deficit.

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

4. General discussion The two experiments described were designed to address whether visual illusions are effectively processed in patients with hemispatial neglect and if so, if this processing does indeed occur implicitly. We also investigated if the effect of changes of line length produced by the Mu¨ller-Lyer illusion would alter the bisection behaviour in the neglect patients, as such errors are known to be sensitive to line length [8,33,35]. These three issues shall be dealt with consecutively.

4.1. Illusions The results of both experiments showed clearly that the stimuli effectively induced illusory effects in all groups and that these illusions worked in the same directions for the neglect group (whether hemianopia was present or not) as well as the two other control groups: For all subjects, the bisection mark was shifted towards outward-projecting fins and away from inward-projecting fins. Fins on the left side proved as effective in their illusionary effects as the fins on the right side. For one of the two critical illusions the effects were also enhanced in the bilateral when compared to the unilateral fin conditions. In the same vein in Experiment 2, the bisection judgement in the Judd figure, with the rightward-projecting fins, was placed to the right of the mark of the Mu¨ller-Lyer figure, with outward-projecting fins. The bisection mark in the Judd figure, with leftward-projecting fins, was placed to the left of the mark of the Mu¨ller-Lyer figure, with inwardprojecting fins. All these observations together confirm that the fins on both ends had an impact on bisection performance. These findings are in full agreement with Ro and Rafal’s [46] single case study and have now been replicated and extended in a group study. Vallar and colleagues [49] who tested six neglect patients with the Brentano form of the Mu¨ller-Lyer illusion also found comparable illusory effects between patients and healthy controls. These data are also in agreement with other studies that have investigated other illusions in hemispatial neglect [48] and more generally in brain lesioned subjects studies [2,18,28]. The inconsistent findings of Mattingley et al.’s [34] study may simply have been due to the fact that the stimuli chosen did not give rise to the strongest possible illusions. This was certainly the case for the longest stimuli for which even healthy subjects failed to show an effect. Another explanation may be derived from the lesion location of these patients: one had an occipital lesion like our patient LC for whom none of the illusions worked (Section 2.3), another patient was simply described as having an MCA lesion and three more patients had subcortical as well as cortical damage. It may well be the case that the damage in these patients affected areas

623

mediating early visual processing and that this led to the inconsistency in the findings. The only behavioural difference existent between the groups proved to be the effect of perceived line length induced by the Mu¨ller-Lyer illusion. In Experiment 1 this effect was observed for the neglect patients only. Bisection judgements were shifted more to the right in the Mu¨ller-Lyer figure with outward-projecting fins, where the shaft was seen as being longer than in the Mu¨ller-Lyer figure with inward-projecting fins, where the shaft was seen as being shorter. This finding is in full agreement with previous studies that have repeatedly shown that the line bisection errors neglect patients produce increase linearly with increasing line length [8,33,35]. This finding obviously generalises to perceived as well as physical line length and Ro and Rafal report the same effect in their single case study. In a very recent study in which Ricci et al. [44] investigated bisection behaviour in the Oppel Kundt illusion, in which physical line length is identical but one line perceived as shorter, this line length effect was also found, again indicating that the effect is based on perceived line length rather than angular distance on the retina. As expected, this effect did not prove significant for the healthy control subjects as a large number of studies have shown that such subjects are very variable in the bisection errors they make, ranging from leftward [11,12] to rightward [20,40] to non-significant errors [22,23] and that systematic line length effects thus prove very unlikely and were indeed not found in study by Harvey et al. [26]. Surprisingly though, the comparison between the two Mu¨ller-Lyer figures also reached significance for the RCVA control group but only for the second experiment. So far the effect has only been observed for neglect patients, in fact hardly any studies have looked at line length effects in RCVA patients without neglect. This was done by Harvey et al. [26] but just as for healthy subjects, they found that the bisection errors did not change with increasing line length. We have no real explanation for this effect and assume that, since it was only found in Experiment 2, it is a weak and inconsistent effect. Another explanation might be that some of the RCVA patients still had mild neglect and that this contributed to the effect. However, this seems unlikely as great care was taken with patient selection, most patients were tested as soon as they proved to be in a stable condition after their stroke onset and were also given a range of neglect tests. Moreover, this could not explain the absence of the effect in Experiment 1.

4.2. Explicit awareness To our surprise, the results of Experiment 1 showed that neglect patients were, with very few exceptions, always aware of the left fins. They clearly perceived the

624

B. Olk et al. / Neuropsychologia 39 (2001) 611–625

stimulus figure as a whole and processed it accordingly, despite the fact that all of them had acute and fairly strong neglect. It is therefore likely that the patients tested by Mattingley et al. [34] also processed the left-sided fins explicitly rather than implicitly and as argued previously, in some respects this result is not that surprising as a number of studies have now shown that patients with hemispatial neglect do not so much neglect single stimuli but rather misrepresent them. However, it has to be granted that in Mattingley et al.’s [34] study the patients were simply asked to bisect the shaft and only occasionally made to report what they had seen after the transection was made. This may have resulted in more unconscious omissions than our more stringent criteria where patients were asked to describe the stimulus every time before they bisected it. Our instruction may have resulted in longer and more conscious scanning behaviour and thus more complete reports. We then assessed the same patients on a comparison task where they had to indicate whether two illusions were identical or not: two patients who occasionally failed to report the left fins in Experiment 1 had no problems in the comparison task in Experiment 2 and four patients, who made mistakes in the comparison task, did not neglect any left-sided fins in Experiment 1. It, therefore, does not necessarily follow that patients who neglect the left side in the comparison task will also do so in the bisection task and vice versa. It rather seems that the requirements of these tasks are sufficiently different to yield these inconsistent results. This finding, that the task requirements are not the same, also poses difficulties for the question as to whether illusions are indeed implicitly processed. As most patients explicitly reported the left-sided fins, these data cannot address the question at all. In Experiment 1, in the few trials in which a left fin was indeed neglected, the illusions tended not to work. In Experiment 2 in which six subjects made mistakes on some comparisons the illusion biases were in the right direction. However, these data are confounded by the fact that these patients may well have been aware of the left-sided fins as here the task requirements, that assessed the explicit processing, were not the same as those involved in the bisection task. Also, as described in more detail in Section 3.3, JC who never reported any left-sided fins in Experiment 1 and made a lot of errors in Experiment 2 showed no significant illusory biases although they were generally in the expected direction (inspection of JC’s data across the two tasks may give the impression that this patient was suffering from directional hypokinesia. However, when tested with both the Milner and Bisiach Landmark tests [10,38] his disorder was clearly identified as perceptual). We can, therefore, not claim that our data support the idea of implicit illusion processing in neglect patients as

(a) most patients perceived the illusions explicitly and (b) in the few cases in which they did not, the results tended to point in the opposite direction. However, we would argue that it is very likely that Ro and Rafal’s patient was indeed unaware of the left fins in the bisection task, despite the fact that the comparison task had different task requirements. Like JC, she produced 100% error rates in the critical comparison task and showed significant biases in her illusion processing. In JC the illusory effects were not significant, despite going in the right direction, but then the power of the test was almost certainly too low to produce a significant result. So despite our attempts the Ro and Rafal [46] result is so far still the only one arguing for implicit illusion processing in hemispatial neglect and this finding has to be interpreted with caution as the procedures that were adopted to insure the lack of explicit processing were not stringent enough.

Acknowledgements We would like to thank all participants for their co-operation and two anonymous reviewers for most helpful comments on an earlier version of this manuscript, which have now been incorporated in the text. M. Harvey is supported by the Wellcome Trust and B. Olk is solely funded by the Wellcome Trust (No: 050184/Z).

References [1] Annett M. The binomial distribution of right, mixed and left handedness. Quarterly Journal of Experimental Psychology 1967;19:327– 33. [2] Basso A, Bisiach E, Faglioni P. The Mu¨ller-Lyer illusion in patients with unilateral brain damage. Cortex 1974;10:26–35. [3] Bayer CA, Pressey AW. Geometric illusions as a function of pigmentation of the fundus oculi and target size. Psychonomic Science 1972;26:77– 9. [4] Baylis GC, Driver J, Rafal RD. Visual extinction and stimulus repetition. Journal of Cognitive Neuroscience 1993;5:453–66. [5] Behrmann M, Watt S, Black SE, Barton JJS. Impaired visual search in patients with unilateral neglect: an oculographic analysis. Neuropsychologia 1997;35:1445– 8. [6] Benton AL, De Hamsher S, Varney NR, Spreen O. Contributions to neuropsychological assessment. New York: Oxford University Press, 1978. [7] Berti A, Rizzolatti G. Visual processing without awareness: evidence from unilateral neglect. Journal of Cognitive Neuroscience 1992;4:345– 51. [8] Bisiach E, Bulgarelli C, Sterzi R, Vallar G. Line bisection and cognitive plasticity of unilateral neglect of space. Brain and Cognition 1983;2:32– 8. [9] Bisiach E, Pizzamiglio L, Nico D, Antonucci G. Beyond unilateral neglect. Brain 1996;119:851– 7. [10] Bisiach E, Ricci R, Lualdi M, Colombo MR. Perceptual and response bias in unilateral neglect: two modified versions of the Milner Landmark Task. Brain and Cognition 1998;37:369–86.

B. Olk et al. / Neuropsychologia 39 (2001) 611–625 [11] Bradshaw JL, Nathan G, Nettleton NC, Wilson L, Pierson J. Why is there a left-side underestimation in rod bisection? Neuropsychologia 1987;25:735–8. [12] Bradshaw JL, Nettleton NC, Nathan G, Wilson L. Bisecting rods and lines: effects of horizontal and vertical posture on left-side underestimation by normal subjects. Neuropsychologia 1985;23:421– 5. [13] Coren S, Girgus JS. Seeing is deceiving: the psychology of visual illusions. Hillsdale, NJ: Lawrence Erlbaum Associates, 1978. [14] Doricchi F, Angelelli P. Misrepresentation of horizontal space in left unilateral neglect. Neurology 1999;52:1845–52. [15] Driver J, Mattingley JB. Parietal neglect and visual awareness. Nature Neuroscience 1998;1:17–22. [16] Ellis AX, Della Sala S, Logie RH. The Bailiwick of visuo-spatial working memory: evidence from unilateral spatial neglect. Cognitive Brain Research 1996;3:71–8. [17] Ferber S, Karnath H-O. Parietal and occipital lobe contributions to perception of straight ahead orientation. Journal of Neurology, Neurosurgery and Psychiatry 1999;67:572–8. [18] Grabowska A, Nowicka A, Szymanska O. Sex related effect of unilateral brain lesions on the perception of the Mu¨ller-Lyer illusion. Cortex 1999;35:231–41. [19] Greene E, Nelson B. Evaluating Mu¨ller-Lyer effects using single fin-set configurations. Perception and Psychophysics 1997;59:293– 312. [20] Halligan PW, Marshall JC. Line bisection in visuospatial neglect: disproof of a conjecture. Cortex 1989;25:503–15. [21] Halligan PW, Marshall JC. The history and clincial presentation of neglect. In: Robertson IH, Marshall JC, editors. Unilateral neglect: clinical and experimental studies. Hillsdale: Lawrence Erlbaum Associates, 1993:3–25. [22] Halligan PW, Manning L, Marshall JC. Individual variation in line bisection: a study of four patients with right hemisphere damage and normal controls. Neuropsychologia 1990;28:1043– 51. [23] Halligan PW, Manning L, Marshall JC. Hemispheric activation vs spatio-motor cueing in visual neglect: a case study. Neuropsychologia 1991;29:165–76. [24] Halligan PW, Marshall JC, Wade DT. Do visual field deficits exacerbate visuo-spatial neglect? Journal of Neurology, Neurosurgery and Psychiatry 1990;53:487–91. [25] Harvey M. Perspectives of visuospatial neglect. In: Milner AD, editor. Comparative neuropsychology. Oxford: Oxford University Press, 1998:213–27. [26] Harvey M, Milner AD, Roberts RC. Differential effects of line length on bisection judgements in hemispatial neglect. Cortex 1995;31:711– 22. [27] Hornak J. Ocular exploration in the dark by patients with visual neglect. Neuropsychologia 1992;30:547–52. [28] Houlard N, Fraisse P, He´caen H. Effects of unilateral hemispheric lesions on two types of optico-geometric illusions. Cortex 1976;12:232– 40. [29] Irving-Bell L, Small M, Cowey A. A distortion of perceived space in patients with right hemisphere-lesions and visual hemineglect. Neuropsychologia 1999;37:919–25. [30] Ishiai S, Furukawa T, Tsukagoshi H. Visuospatial processes of line bisection and the mechanisms underlying unilateral spatial neglect. Brain 1989;112:1485–502. [31] Karnath H-O, Hartje W. Neglect – Symptomatik, Diagnostik, Anatomie und Erkla¨rungskonzepte. In: Markowitsch HJ, editor. Enzyklopa¨die der Psychologie, Band 2, Klinische Neuropsychologie. Go¨ttingen: Hogrefe, 1997:91–154.

.

.

625

[32] Machado L, Rafal RD. Ipsilesional line bisection bias in patients with chronic parietal lesions. NeuroReport 1999;10:3143–8. [33] Marshall JC, Halligan PW. Line bisection in a case of visual neglect: psychophysical studies with implications for theory. Cognitive Neuropsychology 1990;7:107– 30. [34] Mattingley JB, Bradshaw JL, Bradshaw JA. The effects of unilateral visuospatial neglect on perception of Mu¨ller-Lyer illusory figures. Perception 1995;24:415– 33. [35] Mattingley JB, Pierson JM, Bradshaw JL, Phillips JG, Bradshaw JA. To see or not to see: the effects of visible and invisible cues on line bisection judgements in unilateral neglect. Neuropsychologia 1993;31:1201– 15. [36] Milner AD, Goodale MA. The visual brain in action. Oxford: Oxford University Press, 1995. [37] Milner AD, Harvey M. Distortion of size perception in visuospatial neglect. Current Biology 1995;5:85– 9. [38] Milner AD, Harvey M, Roberts RC, Forster SV. Line bisection errors in visual neglect: misguided action or size distortion? Neuropsychologia 1993;31:39– 49. [39] Nelson HE, O’Connell A. Dementia: the estimation of premorbid intelligence levels using the new adult reading test. Cortex 1978;14:234– 44. [40] Nichelli P, Rinaldi M, Cubelli R. Selective spatial attention and length representation in normal subjects and in patients with unilateral spatial neglect. Brain and Cognition 1989;9:57–70. [41] Nico D, Galati G, Incoccia C. The endpoints’ task: an analysis of length reproduction in unilateral neglect. Neuropsychologia 1999;37:1181– 8. [42] Peru A, Moro V, Avesani R, Aglioti S. Influence of perceptual and semantic conflicts between the two halves of chimeric stimuli on the expression of visuo-spatial neglect. Neuropsychologia 1997;35:583– 9. [43] Powell GE, Bailey S, Clark E. A very short version of the Minnesota Aphasia Test. British Journal of Social and Clinical Psychology 1980;19:189– 94. [44] Ricci R, Calhoun J, Chatterjee A. Orientation bias in unilateral neglect: representational contributions. Cortex 2000;36:671–7. [45] Riddoch MJ, Humphreys GW. Perceptual and action systems in unilateral visual neglect. In: Jeannerod M, editor. Neurophysiological and neuropsychological aspects of spatial neglect. Amsterdam: Elsevier, 1987:151– 81. [46] Ro T, Rafal RD. Perception of geometric illusions in hemispatial neglect. Neuropsychologia 1996;34:973– 8. [47] Schacter DL, McAndrews MP, Moscovitch M. Access to consciousness: dissociations between implicit and explicit knowledge in neuropsychological syndromes. In: Weiskrantz L, editor. Thought without language. Oxford: Claredon Press, 1988:242– 78. [48] Soroker N, Calamaro N, Myslobodsky M. ‘‘McGurk Illusion’’ to bilateral administration of sensory stimuli in patients with hemispatial neglect. Neuropsychologia 1995;33:461– 70. [49] Vallar G, Daini R, Antonucci G. Processing of illusion of length in spatial hemineglect: a study of line bisection. Neuropsychologia 2000;38:1087– 97. [50] Walker R, Findlay JM, Young AW, Lincoln NB. Saccadic eye movements in object-based neglect. Cognitive Neuropsychology 1996;13:569– 615. [51] Wechsler D. Wechsler adult intelligence scale – revised. New York: The Psychological Corporation, 1981. [52] Wilson B, Cockburn J, Halligan PW. Behavioural inattention test. Titchfield, Hampshire: Thames Valley Test Company, 1987.