c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
Available online at www.sciencedirect.com
Journal homepage: www.elsevier.com/locate/cortex
Research report
Paradoxical extension into the contralesional hemispace in spatial neglect Bigna Lenggenhager a,b,l,*, Tobias Loetscher a,c,d,1, Nicole Kavan a, Gianandrea Pallich a, Amy Brodtmann d, Michael E.R. Nicholls c and Peter Brugger a,e a
Neuropsychology Unit, Department of Neurology, University Hospital, Zurich, Switzerland Social and Cognitive Neuroscience Laboratory, Psychology Department of “Sapienza”, Rome, Italy c School of Psychology, Flinders University, Bedford Park, Australia d Florey Neuroscience Institutes, Melbourne, Australia e Zurich Center of Integrative Human Physiology (ZIHP), University of Zurich, Switzerland b
article info
abstract
Article history:
To explore the idea of a perceptual distortion of space in spatial neglect, neglect patients,
Received 18 June 2010
age-matched healthy controls and right hemisphere control patients judged the vanishing
Reviewed 3 August 2010
point of horizontally and vertically-moving stimuli. Hemifield of presentation and
Revised 1 November 2010
movement direction of the stimulus presentation was manipulated. The results suggest
Accepted 12 October 2011
that neglect patients show a stronger response bias in the direction of the moving stimuli
Action editor Giuseppe Vallar
(“representational momentum”) than healthy and right hemisphere controls. Furthermore,
Published online 22 October 2011
neglect patients, but not the control groups, showed a direction-specific response whereby the presence of neglect was associated with a larger representational momentum for
Keywords:
leftward-moving stimuli. The one left-hemisphere patient with right-sided neglect showed
Space distortion
the opposite effect. Thus, neglect patients showed a relative overextension into their
Representational momentum
neglected side of space. While these findings are in line with the idea of an extension in the
Spatial remapping
representation of contralesional space, other explanations such as deficient spatial
Smooth pursuit
remapping, impairments in smooth pursuit and distortions in memorized visuo-motor
Spatial memory
movements are considered. ª 2011 Elsevier Srl. All rights reserved.
1.
Introduction
Spatial neglect is a puzzling disorder, which often accompanies damage to the right parietal cortex. While neglect is primarily characterised by a failure to report, respond or orient to stimuli presented to the contralesional side (Heilman, 1979), there are occasions where excessive
perceptual/motor activity is directed towards the neglected hemispace. Such counterintuitive behaviour was demonstrated by Bisiach et al. (1994) when they presented a single point (corresponding to the midpoint of a previously presented line) to patients with left-sided neglect. When asked to mark the complementary endpoints of the virtual line, patients placed the left endpoint significantly further
* Corresponding author. Social and Cognitive Neuroscience Laboratory, Psychology Department of “Sapienza”, Via dei Marsi 78, 00185 Rome, Italy. E-mail address:
[email protected] (B. Lenggenhager). 1 Shared first authorship. 0010-9452/$ e see front matter ª 2011 Elsevier Srl. All rights reserved. doi:10.1016/j.cortex.2011.10.003
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
away from the centre than the right one. That is, there was an overextension instead of a neglect towards the left. This surprising finding was subsequently confirmed in a number of studies using the same or related techniques (e.g., Bisiach et al., 1996, 1998; Irving-Bell et al., 1999; Kerkhoff, 2000; Latini Corazzini et al., 2005). These studies also revealed that, although the leftward extension is frequently found, not all patients with neglect show this behaviour. Furthermore, a number of factors are known to affect the extent of overshooting, including hemianopia, stimulus position and scanning direction (e.g., Doricchi and Angelelli, 1999; Geminiani et al., 2002; Ishiai et al., 1994; Nico et al., 1999). According to Bisiach et al. (1994), leftward extension in the endpoint task was not compatible with many common theories of spatial neglect, which attribute the disorder to attentional impairments (Heilman and Van Den Abell, 1980; Mesulam, 1981), deficits in planning and the execution of movements (Heilman et al., 1985; Watson et al., 1978) and/or disruption of space representation (Bisiach and Luzzatti, 1978). Instead, Bisiach et al. (1996, 1998) speculated that a distortion of perceived space caused the leftward overextension. Specifically, they suggested that the representation of contralesional space would be logarithmically expanded, whereas the ipsilesional side would be compressed. Consequently, for two equal horizontal line segments in representational space, the left segment would map onto a larger physical extension than the right segment. Conversely, for two equal line segments presented in physical space, the left segment is perceived as shorter than the right counterpart (see Chatterjee, 2002). The assumption of an anisometric space representation provides a theoretical framework to understand why patients with neglect overshoot towards the left in the endpoint task and why they deviate rightwards in line bisection tasks. Related accounts of a perceptually shrunken left side of physical space have also been put forward by other investigators (Halligan and Marshall, 1991; Milner and Harvey, 1995). Others, however, have questioned this account and, if acknowledged at all, a distorted space representation is seen as a mere epiphenomenon of spatial neglect (Karnath and Ferber, 1998; Perri et al., 2000; Rode et al., 2006, 2008). The current study explored the hypothesis of a distorted space using the phenomenon of “representational momentum”. Representational momentum (RM) refers to the observation that, when people judge the vanishing point of a moving object that suddenly disappears, they misplace its final position in the direction of the movement (Freyd and Finke, 1984; Thornton and Hubbard, 2002). While RM has been extensively studied in healthy participants, it has only been studied once in neglect patients (McGeorge et al., 2006). In healthy subjects, variables such as velocity and direction of target motion and context of the display have been shown to modify RM. For example, a target that is believed to move faster (e.g., a rocket) results in a larger forward displacement than an object that is believed to move more slowly [e.g., a steeple (Thornton and Hubbard, 2002)]. For a review of the empirical literature on RM, see Hubbard (2005). The standard paradigm of RM requires participants to memorize the final position of a moving target that suddenly disappears. Since a distorted space representation in neglect
1321
necessarily leads to inaccurate task performance, RM seems particularly suited to uncover such distortions. We tested this idea in a group of patients with spatial neglect, in agematched healthy controls and in a further control group of patients with right hemisphere lesions, but no neglect symptoms (right hemisphere (RH) controls). Prior to the main task, we assessed how accurately participants point to stationary targets presented in a variety of locations on a touch screen. This test allowed us to demonstrate that the basic ability to detect targets and accurately point to them was preserved in neglect patients. The main task assessed RM as a function of movement direction and side of stimuli presentation. Participants pointed to the location where a moving target (moving either horizontally or vertically) suddenly vanished. For horizontal movements, a number of predictions were made. The account of distorted space representation in left neglect maintains that the left side of represented space is enlarged whereas the right side of space is compressed (Bisiach et al., 1996, 1998). If this account is correct, we expected neglect patients to show an enhanced overestimation (i.e., judging the final position further to the left than healthy participants or RH control patients) for targets moving leftwards and an underestimation for rightward movements. We also investigated whether the RM in neglect patients is modulated by the side on which the stimuli are presented (centre, left, and right hemispace). Given the inconsistent findings of position effects in the line extension task (e.g., Geminiani et al., 2002; Nico et al., 1999), no specific predictions were made for the side of stimulus presentation. Bisiach’s et al. (1996, 1998) framework only predicts distortions in the horizontal dimension. For vertical movements we therefore predicted no difference between healthy, control patients and neglect patients. The RM task was also applied to a single patient with the relatively rare condition of right-sided neglect after left brain damage. For this patient, we hypothesized the reversed pattern compared to left-sided neglect patients for the horizontal conditions, but not for the vertical conditions.
2.
Methods
2.1.
Subjects
Nine patients with left hemi-neglect after right-hemisphere brain damage (five women; mean age 64.5 years, standard error of mean (SEM) 5.0), nine sex- and age-matched healthy controls (five women; mean age 64.5 years, SEM 4.7) and nine age-matched RH controls (four women; mean age 59.5, SEM 5.2), as well as one patient with right hemi-neglect after lefthemisphere brain damage (a 58-years old man) participated in this study. As in our previous study (Loetscher and Brugger, 2009), neglect was defined by rather conservative criteria. That is, all neglect patients showed typical clinical symptoms of spontaneous neglect behaviour (Ferber and Karnath, 2001) and indications of neglect in at least two of the following paper-and-pencil tests: figure copying, cancellation and line bisection. The figure by Ogden (1985) was used for the copying task, any left-sided omission was taken as indication of neglect. In the cancellation task [“Bells test” (Gauthier et al.,
1322
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
1989)] the number of left minus right-sided omissions was calculated and scores above two were taken as a cut-off. Nine lines with lengths between 120 and 160 mm had to be bisected and the average deviation from the lines’ true midpoint was calculated. Rightward deviations larger than 6.5% were regarded as indicative of neglect. Cut-off scores for all these three neglect tests were as defined in Azouvi et al. (2002). All subjects were right-handed. Table 1 summarizes the demographic and clinical data of all neglect and RH control patients. The study protocol was performed in accordance with the ethical standards laid down in the Declaration of Helsinki and all subjects gave written consent.
2.2.
Stimuli and experimental set-up
Stimuli were presented on an Elo Entuitive 1700 LCD touch monitor (1280 1024 Pixel, 340 270 mm). The generation and sequencing of stimuli was controlled with Java Script programming. Participants were seated in a darkened room with their sagittal midplane aligned to the centre of the touch screen. Participants viewed the monitor from a distance of 600 mm and eye and limb movements were not constrained. The Pointing Accuracy (PA) Task was always completed first, followed by the RM Task.
2.2.1.
Pointing accuracy task
The PA task measured PA to static targets at different locations on the screen. Participants began each trial by depressing a centrally aligned button with their right index finger. When the button was pushed, a white circle (f 5 mm) appeared against a black background. Participants were told that their task was to move their finger from the central button and touch the target as accurately as possible. When the screen was touched, the target disappeared, and the horizontal and vertical coordinates of this touch were recorded. The next trial was initiated when participants returned
their finger to the central start button. This procedure ensured a central starting point for each trial. The targets were presented at three different horizontal (left, centre, right) and vertical (top or bottom) locations. The specific coordinates for the six locations were: Left/Top (85/þ45), Centre/Top (0/þ45), Right/Top (þ85/þ45), Left/ Bottom (85/45), Centre/Bottom (0/45) and Right/Bottom (þ85/45). The coordinates are given in mm with respect to the centre of the screen being 0/0 mm. The six targets were shown six times in a pseudo-randomized order, resulting in 36 trials.
2.2.2.
Representational momentum task
Like the PA task, each trial was initiated by pressing a centrally aligned button with the right index finger. After pressing the button, a target appeared (same dot as in the PA task), which started to move smoothly across the display and then suddenly disappeared. Participants were required to move their finger without any delay from the central button and touch the position of the screen where the target had vanished. The corresponding horizontal and vertical coordinates were recorded. To ensure that the movement was initiated after the disappearance of the target, the participants had to keep the central button pressed until the target vanished. In cases where the button was released too early, the trial was rejected and repeated. The targets either moved horizontally or vertically. For horizontal trials, the targets were all presented in the same elevation along the vertical centre of the screen. The horizontal location of the targets was varied in three steps across the screen. The start/vanish coordinates were: Left (145/0 and 25/0), Centre (60/0 and þ60/0) and Right (þ25/ 0 and þ145/0) with respect to the centre of the screen being 0/ 0 mm. The start/vanish positions were alternated within each of the three steps so that sometimes the target appeared on the left and vanished on the right (and vice versa). Thus, targets moved left and right along three paths located to the left, right and centre of the screen (see Fig. 1). This condition allowed us
Table 1 e Demographic and clinical data of all patients. Patient 1 2 3 4 5 6 7 8 9 10 C1 C2 C3 C4 C5 C6 C7 C8 C9
Age
Sex
Lesion type
Lesion side
Lesion site
Days from onset
Visual field deficit
Paresis
Showing neglect in
85 57 60 74 60 80 62 64 38 58 37 75 34 56 53 69 77 67 67
f f m f m m m f m m m w m m w w m w m
Vascular Tumour(Glioma) Vascular Vascular Tumour (Glioblastoma) Vascular Vascular Vascular Tumour (Sarcoma) Tumour (Glioblastoma) Vascular Vascular Tumour (Gliom) Vascular Tumour (Meningeom) Tumour (Glioblastoma) Vascular Vascular Vascular
Right Right Right Right Right Right Right Right Right Left Right Right Right Right Right Right Right Right Right
F, P, O F, T F, P F, SC F, T, O F, SC F F, T, P T, P, O F, O, SC F, P F, P F, P F, T, P, SC P F T, F, SC P, O F, P
2 Post-op 3 6 Pre-op 21 2 120 Pre-op Pre-op 8 Post-op Pre-op 14 Post-op Post-op 5 NA 5
Yes Yes No No Yes Yes Yes No Yes NA No No No Yes No No Yes No No
No No Yes Yes No Yes Yes No NA No Yes No No Yes No No No Yes Yes
CT, LB C, CT, LB C, CT, LB C, CT, LB C, LB C, LB C, LB C, LB C, LB C, LB e e e e e e e e e
Site of lesion: F ¼ Frontal, T ¼ Temporal, P ¼ Parietal, O ¼ Occipital, SC ¼ Subcortical; Days from onset: number indicates days after stroke, postop ¼ postoperative, pre-op ¼ preoperative; NA ¼ not available; Showing neglect in: C ¼ Copying, CT ¼ Cancellation task, LB ¼ Line bisection.
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
Fig. 1 e Schematic presentations of starting and vanishing locations as a function of presentation side (left, centre, right) and the target’s movement. The lines within the screen depict the six pathways along which the target could move. In each case, the target could move in either direction.
to evaluate RM for horizontally-moving targets located in the left and right hemispaces as well as along the centre. For vertical trials, the starting position of the target corresponded to the six locations in the PA Task (see above). For half of the trials, the target started in one of the positions in the lower half of the screen and moved upwards. In the other half of trials, the target appeared in the upper half of the screen and moved downwards (see Fig. 1). Therefore, targets moved up and down along three paths located in the left, right and centre of the screen. This condition allowed us to evaluate RM for vertically-moving targets located in the left and right hemispaces as well as along the centre. While targets differed in respect to the direction of their trajectory (leftwards, rightwards, upwards, and downwards), speed (120 mm/sec) and trajectory length (120 mm) were kept constant. There were six repetitions for each of the six vertical and six horizontal conditions. The corresponding 72 trials were presented in a pseudo-randomized order.
2.3.
Data acquisition and processing
For both tasks, deviations from the true target location were calculated to the nearest millimetre. In the PA, task, positive values indicate rightward and upward deviations, respectively. In the RM task, deviation was only measured in respect to the trajectory of the target. Positive RM values denote an extension in the direction of the moving stimuli whereas negative values indicate under-extension. The deviations were analysed using repeated-measures analyses of variance (ANOVAs). Bonferroni correction was used post-hoc analyses. The single right-neglect patient was excluded from all statistical analysis and is reported using descriptive statistics only.
3.
Results
3.1.
Pointing accuracy task
Horizontal Deviations. The data were analysed with a 3 3 ANOVA with the factors Group (neglect, healthy, RH controls)
1323
and Side (left, centre, right). There was a significant effect for Side [F(2,48) ¼ 7.5, p ¼ .003, h2p ¼ :40]. Post-hoc tests demonstrated that pointing to right-sided targets (mean .4 mm, p ¼ .003) was more accurate than to central targets (mean 1.7 mm). Furthermore there was a main effect of Group [F(2,24) ¼ 4.0, p < .03, h2p ¼ :25] with a slightly better PA in healthy (mean .2 mm) than in RH controls (mean 1.7 mm; p ¼ .03). Importantly for the current study, there was no difference in PA between the neglect patients and either control group (healthy p ¼ .30, RH controls p ¼ .84). Vertical Deviations. The data were analysed with a 3 3 ANOVA with the factors Group (neglect, healthy, RH controls) and Side (left, centre, right). There were no effects for Side [F(2,48) ¼ 1.4, p ¼ .3] and Group (F(2,24) ¼ 3.2, p ¼ .1) and no interaction between these two factors [F(4,48) ¼ 2.0, p ¼ .1].
3.2.
Representational momentum
Two separate, but analogous, analyses were performed for horizontal and vertical movements:
3.2.1.
Horizontal movement
The horizontal RM deviations were submitted to a 3 3 2 repeated-measure ANOVA with the between-subject factor Group (neglect, healthy, RH controls) and the within-subject factors Side (left, centre, right) and Direction (leftward, rightward). The analysis revealed a main effect of Group [F(2,24) ¼ 11., p < .001, h2p ¼ :50] with left neglect patients (mean 11.8 mm 2.2 SEM) showing a larger forward displacement (i.e., RM) than healthy controls (mean 2.2 mm .6 SEM, p ¼ .001) and RH controls (1.8 mm 1.7 SEM, p ¼ .001). There was also a main effect of Side [F(2,48) ¼ 3.8, p ¼ .04, h2p ¼ :25] with a slightly larger RM in the centre than in the right condition. The interaction between Direction and Group was significant [F(2,24) ¼ 3.7, p ¼ .04, h2p ¼ :24] with the forward displacement being significantly larger for leftwards than for rightwards movements in the neglect group only ( p ¼ .01). Finally, there were significant two-way interactions between Side and Group [F(4,48) ¼ 2.6, p ¼ .048, h2p ¼ :18] and Side and Direction [F(2,24) ¼ 15.6, p < .001, h2p ¼ :58] as well as a three-way interaction between Group, Direction and Side [F(4,48) ¼ 4.2, p ¼ .006, h2p ¼ :26; Fig. 2].
Fig. 2 e Forward displacement in mm (mean ± SE) as a function of movement direction, side of stimulus presentation and group.
1324
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
A closer inspection of the interaction between Direction and Side revealed that overall there was no significant RM for the two conditions with the most lateral vanishing positions. That is, there was neither an RM for leftward moving targets presented on the left side [t(26) ¼ 1.1, p ¼ .23] nor for rightward moving targets presented on the right side [t(26) ¼ .01, p ¼ .99], whereas the expected significant forward displacement was found in all the other horizontal conditions (each with p .003). We therefore focused our analysis on central trials (see Discussion section for an interpretation of the absence of an RM in the other conditions). Accordingly, a 3 2 repeated-measures ANOVA with the factors Group (neglect, healthy, RH controls) and Direction (leftward, rightward) was calculated. This analysis again revealed a main effect of Group [F(2,24) ¼ 9.9, p < .001, h2p ¼ :45] with neglect patients showing a larger RM (mean 15.5 mm 3.0 SEM) than the healthy controls (mean 3.4 mm .8 SEM, p ¼ .004) and RH controls (mean 2.3 mm 1.7 SEM, p < .001). There was also a significant effect of Direction [F(1,24) ¼ 6.7 p ¼ .016, h2p ¼ :22] indicating a larger RM for leftward moving targets (mean 8.7 mm 1.7 SEM) compared to rightward movements (mean 5.4 mm 1.3 SEM). Finally there was an interaction effect between Group and Direction [F(2,24) ¼ 12.1, p < .001, h2p ¼ :50], which showed that the direction effect was solely caused by the neglect group (see Fig. 3). Bonferroni post-hoc tests confirmed a significant effect of movement direction on RM in neglect patients ( p < .001), but not in healthy controls ( p ¼ .99) nor in RH controls ( p ¼ .33). That is, the RM effect in patients with left-sided neglect was significantly larger for leftward movements (mean 21.5 mm 4.6 SEM) than for rightward movements (mean 9.4 mm 3.0 SEM, see Fig. 3). To exclude the possible effect of visual field defects, we compared neglect patients with (n ¼ 6) and without (n ¼ 3) visual field deficits (compare Table 1). The results show that visual field defects did not modulate the horizontal RM in patients with neglect (Wilcoxon W ¼ 28, p > .6, r ¼ .17).
Fig. 3 e Forward displacement in mm (mean ± SEM) for targets presented in the centre of the screen as a function of movement direction. Data are provided for healthy participants, the RH controls, the left-sided neglect patients and a single patient with right-sided neglect. Note that, in the case of a rightward motion, the target’s starting position was on the left and it disappeared on the right (and vice versa for leftward motion).
3.2.2.
Vertical movement
The RM deviations for the vertical condition were submitted to a 2 3 2 ANOVA with the factors Group (neglect, controls), Side (left, centre, right) and Direction (upwards, downwards). There was a highly significant main effect of Group [F(2,24) ¼ 8.4, p < .002, h2p ¼ :41], again with a larger RM for neglect patients compared to the controls. There were no other significant main or interaction effects and therefore no further tests were performed for the vertical RM.
3.2.3.
Data for the single right-neglect patient
Descriptive statistics are provided for the single right neglect only. He deviated slightly to the right (.5 mm) and downwards (4.4 mm) when pointing to stationary targets (PA task). For the RM task, leftward moving targets produced deviations of 2.4 mm, 17.5 mm and 23.7 mm for trials to the left, centre and right, respectively. Rightward moving targets produced deviations of 20.5 mm, 33.5 mm and 3.7 mm for trials to the left, centre and right, respectively. The single right-neglect patient thus showed a similar framing effect as the other participants and, most importantly, larger forward displacements for rightward moving targets compared to leftward moving targets in the centre condition. That is, he showed a reversed pattern of RM compared to patients with left neglect (see also Fig. 3).
4.
Discussion
In the PA task, we established that the neglect patients was just as accurate as the controls when pointing to visible stationary targets. Importantly, we demonstrated that the equivalence between the neglect and control populations was the same irrespective of where the target appeared. This finding is in line with previous research with neglect patients, which demonstrates an absence of pointing/reaching impairments in terms of accuracy or deviation (Himmelbach and Karnath, 2003; Karnath et al., 1997; Rossit et al., 2009b, but see Coulthard et al., 2006 for conflicting results). The possibility that the findings in the subsequent RM task were confounded by impairments of goal-directed pointing movements can therefore be excluded. For the RM task, healthy controls displayed a typical RM, i.e., they misplaced the subjective final position of the target too far along its trajectory (Freyd and Finke, 1984; Thornton and Hubbard, 2002). Interestingly, patients with neglect showed a much larger RM than healthy age-matched controls, an effect that was independent of the trajectory (left, right, up, down). This effect might suggest that the larger RM across these conditions is rather due to impairments of non-lateralized attentional functions. In support of this account, spatial neglect is often accompanied by non-spatial deficits such as impairments in selective and sustained attention (Husain and Rorden, 2003). The idea that a general increase in RM for neglect patients is the result of a non-lateralized attentional deficit ties in nicely with research in healthy controls, which shows that diminished attention leads to an increase in RM (Hayes and Freyd, 2002; but see Kerzel, 2003 for conflicting results). It is noteworthy, however, that RH controls did not
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
show a comparable general increase in RM. More testing would have been needed to establish whether general attentional functions were less impaired in RH controls compared to neglect patients. While our study shows a clear elevation of RM in neglect patients, the one other study that has investigated RM in this population has not observed this effect (McGeorge et al., 2006). In that study, RM was investigated as a function of the trajectory’s direction, speed, and distance. The authors found no overall differences in the RM effect between patients with neglect and healthy age-matched controls. They reported, however, that RM in neglect patients decreased with increases in the distance of the trajectory. For the shortest trajectory, the RM seemed to be significantly larger in neglect patients than in controls (see Fig. 3A in McGeorge et al., 2006). For trajectories, whose length approximated the length used in the current study, no elevation in RM was observed for neglect patients. The discrepancy between the current study and the one carried out by McGeorge et al. (2006) may be due to methodological differences. In our study, participants indicated the position of the vanishing target by directly pointing to the touch screen, whereas McGeorge and co-authors asked participants to navigate a mouse cursor, which was, rather sub-optimally, always placed in the top lefthand corner. Furthermore, they varied the speed of the target, which is known to modulate the RM, whereas the velocity was kept constant in the current study. Clearly, given the fundamental methodological differences between the only two studies that investigated RM in neglect patients, it is mute to further compare these two experiments. Unexpectedly, no RM was observed when the target disappeared close to the left or right edge of the screen. The two conditions that led to this effect were: (a) rightward movements for trials contained in the right side of the screen and (b) leftward movement for trials contained in the left side of the screen (see also Fig. 1). In these conditions, the target disappeared only 25 mm from the screen’s edge while in the other horizontal conditions, distance from the vanishing point to the screen’s edge was at least 110 mm. The fact that the disruption of RM was found independent of the direction of the target’s movement in both healthy controls and patients suggest that this finding is the result of an unexpected “framing effect”. It seems likely that the screen’s frame acted as a reference point, which increased the accuracy of judgements and prevented an RM from occurring. Furthermore, the finding that topedown knowledge and contextual surroundings can modulate RM ties in with research in this area (Hubbard, 2005). Trials where targets disappeared close to the screen’s edge have clearly contaminated the results and reduced the effect of RM. To avoid these framing effects, we limited further analyses to the ‘centre’ trials where the target crossed the centre into the contralateral hemispace. Analysis of these data revealed that healthy participants as well as RH controls did not show any direction-specific RM effects. In stark contrast, patients with left neglect showed marked forward displacements of the perceived vanishing position, which depended on the target’s direction. The effect of RM was more pronounced when the target moved into the neglected side of space (i.e., right-to-left) than when it moved towards the
1325
ipsilesional side (i.e., left-to-right). The one patient with right neglect showed exactly the opposite pattern and showed more RM for left-to-right movement (see Fig. 3). Asymmetries in RM between leftwards and rightwards movement mimic findings related to the overextension of lines in neglect (e.g., Bisiach et al., 1996, 1998). A larger RM for targets moving into the left hemispace is in line with Bisiach’s theory of anisometric mental coordinates. That is, due to a relative expansion of contralesional representational space, the memorized spatial locations map further towards the left side in physical space. The contralesional overextension cannot be accounted for by purely spatial-attentional theories such as, hyperattention to the right, hypoattention to the left or a disengagement deficit (Heilman and Van Den Abell, 1980; Mesulam, 1981; Posner et al., 1984). Furthermore, the overextension cannot be attributed to more peripheral explanations such as visual field defects. While such defects have been shown to be critically associated with a contralesional overshooting in static line extension tasks (e.g., Doricchi and Angelelli, 1999; Doricchi et al., 2002a), we found no indication for such an association in the RM task with moving targets. It cannot be excluded that, in our patients with hemianopia and neglect, the visual field defects may have been confined to stationary targets (no dynamic perimetry was undertaken). While the RM was markedly stronger for leftward moving targets, a significant effect of RM was still observed for rightward movements. This overestimation into the right hemispace might challenge Bisiach’s idea (Bisiach et al., 1996, 1998) of a logarithmically compressed representation of ipsilesional space. If the representation of ipsilesional space was compressed, an underestimation for memorized locations on that side would have been expected. This was, however, not the case (see Fig. 3), since the magnitude of the RM for rightward movements (9.4 mm) was roughly equivalent to the RM for vertical movements (7.7 mm downwards and 8.3 mm upwards). Advocates of the anisometry hypothesis might object, however, that the use of dynamic stimuli could have affected the structure of represented space. Rejecting the idea of a static left-right anisometry of space, Bisiach et al. (1999) proposed that a patient’s motor action in space leads to a deformation of the represented structure of space. In line with this idea, neglect patients are known to exhibit lateral asymmetries in the slow ocular pursuit of visual stimuli (Doricchi et al., 2002b, 2007; Incoccia et al., 1995; see also Baloh et al., 1980; Lynch and McLaren, 1983) and a critical role of eye movements has been implicated in the generation of the RM effect in healthy subjects (Kerzel, 2000; Kerzel et al., 2001). Thus, asymmetries in lateral eye movements are a key candidate in contributing to the reported RM imbalances in neglect. Further research, however, will be necessary to elucidate how lateral oculomotor abnormalities are linked to asymmetrical misrepresentation of horizontal space in RM. Studies using static stimuli provide alternative explanations for the observed RM imbalances. A recent study on the impact of neglect on spatial memory has yielded results remarkably similar to our own findings (Rossit et al., 2009b). In that study, neglect patients reached towards visible and memorized positions of stationary targets. For visible targets,
1326
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
a high level of accuracy was observed for the left and right sides. For memorized targets, however, neglect patients overshot to the left for targets on the left, whereas no bias was observed for targets on the right. The authors interpreted their findings in the context of Goodale and Milner’s seminal model of vision (Goodale and Milner, 1992). In a nutshell, the model distinguishes between vision for action and vision for perception. It is proposed that visuo-motor control for goaldirected “on-line” movements is mediated by a dorsal visual stream. A ventral visual system, on the other hand, is thought to mediate visual input for object identification and for controlling actions to objects that are processed out of view e or “off-line” (Rossit et al., 2009b). With this in mind Rossit et al. (2009b) proposed that a damaged ventral, but relatively intact dorsal visual stream, could cause the accurate “on-line” and distorted “off-line” movements in their study. This claim was substantiated in a neuro-anatomical follow-up study, where it was shown that the overshooting to memorized left-sided locations was associated with damage to the ventral visual stream (Rossit et al., 2009b).2 While the studies by Rossit et al. (2009a, 2009b) looked specifically at immediate and delayed movements to memorized target locations, a similar temporal distinction has been made in a purely perceptual memory task (Pisella et al., 2004). In this task, patients were shown a matrix containing four objects in different positions. After a short delay, a second matrix was shown and patients detected changes in location, colour or shape of the objects. Neglect patients with parietal lesions were impaired in their ability to detect changes in object locations, whereas the detection of changes in colour or shape was relatively preserved. Importantly, these impairments were only found when there was a delay between the two matrices (off-line condition), but not when the two matrices were presented immediately after each other (on-line condition). The authors attributed their results to disrupted spatial remapping mechanisms, which prevented accurate representation of spatial locations across saccades. Vuilleumier et al. (2007) empirically tested the idea of impaired remapping in patients with neglect and demonstrated that perceptual memory of locations is indeed impaired across gaze shifts in these patients. Further research showed that spatial working memory is particularly impaired across saccades in neglect (Husain et al., 2001). It is tempting to speculate that our findings are also related to spatial remapping. The current results suggest that the reported asymmetries in RM are linked to the presence of neglect, but whether distortions in spatial representation are specific to neglect is a rather controversial issue. As noted in the introduction, there is both support in favour and against the idea that space distortions are neglect specific. From a theoretical point of view, it seems hard to dispute that a distorted representation of horizontal space should manifest itself in some form of 2 One reviewer suggested an interesting alternative interpretation. He pointed out that in Rossit et al., 2009a, 2009b the major anatomical correlate of leftward overextension in the positioning of remembered static visual location was the lesion of the optic radiations (see corresponding panel C in figure 1, page 2151). This type of lesion usually engenders hemianopia. The leftward overextension might thus have been mediated by such a visual field defect, which is implicated to cause incorrect estimates of horizontal distances (Doricchi and Angelelli, 1999).
asymmetrical behaviour in space. This raises a fundamental question: At what point are asymmetries in spatial behaviour considered to constitute “neglect”? The question is whether neglect is indicated by any asymmetric space-related behaviour or whether it can only be specified by classical neglect tasks such as line bisection and cancellation tasks. A related issue contributing to the controversy is the possibility that an observation of a distorted space representation depends on the specific task requirements. There have been suggestions, for example, that distortions in neglect patients become evident in tasks where access to a more global representation of space is required, while tasks requiring more local levels of processing appear not be affected by putative distortions of representational space (Gallace et al., 2008). Furthermore, given the variety of neglect manifestations, distortions in space representation may vary across different subtypes of spatial neglect. While these points need to be investigated further, they do not detract from the general finding that neglect asymmetrically affects RM. A logical next step would be to monitor eye movements during the RM task to determine the potential impact of oculomotor deficits on the observed effects (Doricchi and Angelelli, 1999). In addition, deviations of pointing to vanished stationary targets should be directly compared to the deviations in pointing to vanished moving targets (i.e., the RM). Such comparisons would determine the contribution of smooth-pursuit deficits to the observed RM effects.
5.
Conclusions
While the current study reports an apparently anomalous contralesional overshooting effect in patients with spatial neglect, the study cannot unravel the underlying processes. The reported findings are in part compatible with the idea of a distorted space representation (Bisiach et al., 1996, 1998). However, the findings can also be explained with reference to impairments in smooth pursuit of a visual target (Doricchi et al., 2007) or the maintenance of memorized target locations (Pisella et al., 2004; Rossit et al., 2009a, 2009b; Vuilleumier et al., 2007). Future studies are needed to disentangle these potential explanations. Ultimately, shedding light on the underlying mechanisms is likely to reveal fundamental insights into how motion and space are represented in the human brain.
Acknowledgement We thank Enrique Wintsch (Zurich) for programming assistance. We are further indebted to Professor Fabrizio Doricchi and two anonymous reviewers for their valuable comments on the manuscript. BL and TL were supported by the Swiss National Science Foundation.
references
Azouvi P, Samuel C, Louis-Dreyfus A, Bernati T, Bartolomeo P, Beis Jm, et al. Sensitivity of clinical and behavioural tests of
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
spatial neglect after right hemisphere stroke. Journal of Neurology, Neurosurgery and Psychiatry, 73(2): 160e166, 2002. Baloh RW, Yee RD, and Honrubia V. Optokinetic nystagmus and parietal lobe lesions. Annals of Neurology, 7(3): 269e276, 1980. Bisiach E and Luzzatti C. Unilateral neglect of representational space. Cortex, 14(1): 129e133, 1978. Bisiach E, Pizzamiglio L, Nico D, and Antonucci G. Beyond unilateral neglect. Brain, 119(3): 851e857, 1996. Bisiach E, Ricci R, Berruti G, Genero R, Pepi R, and Fumelli T. Two-dimensional distortion of space representation in unilateral neglect: Perceptual and response-related factors. Neuropsychologia, 37(13): 1491e1498, 1999. Bisiach E, Ricci R, and Modona M. Visual awareness and anisometry of space representation in unilateral neglect: A panoramic investigation by means of a line extension task. Consciousness and Cognition, 7(3): 327e355, 1998. Bisiach E, Rusconi ML, Peretti VA, and Vallar G. Challenging current accounts of unilateral neglect. Neuropsychologia, 32(11): 1431e1434, 1994. Chatterjee A. Spatial anisometry and representational release in neglect. In Karnath HO, Milner AD, and Vallar G (Eds), The Cognitive and Neural Bases of Spatial Neglect. Oxford: University Press, 2002: 167e180. Coulthard E, Parton A, and Husain M. Action control in visual neglect. Neuropsychologia, 44(13): 2717e2733, 2006. Doricchi F and Angelelli P. Misrepresentation of horizontal space in left unilateral neglect: Role of hemianopia. Neurology, 52(9): 1845e1852, 1999. Doricchi F, Iaria G, Silvetti M, Figliozzi F, and Siegler I. The “Ways” we look at dreams: Evidence from unilateral spatial neglect (with an evolutionary account of dream bizarreness). Experimental Brain Research, 178(4): 450e461, 2007. Doricchi F, Galati G, Deluca L, Nico D, and D’olimpio F. Horizontal space misrepresentation in unilateral brain damage. I. Visual and proprioceptive-motor influences in left unilateral neglect. Neuropsychologia, 40(8): 1107e1117, 2002a. Doricchi F, Siegler I, Ilaria G, and Berthoz A. Vestibulo-ocular and optokinetic impairments in left unilateral neglect. Neuropsychologia, 40(12): 2084e2099, 2002b. Ferber S and Karnath HO. How to assess spatial neglect e Line bisection or cancellation tasks? Journal of Clinical and Experimental Neuropsychology, 23(5): 599e607, 2001. Freyd JJ and Finke RA. Facilitation of length discrimination using real and imaged context frames. American Journal of Psychology, 97(3): 323e341, 1984. Gallace A, Imbornone E, and Vallar G. When the whole is more than the sum of the parts: Evidence from visuospatial neglect. Journal of Neuropsychology, 2(2): 387e413, 2008. Gauthier L, Dehaut F, and Joanette Y. The bells test e A quantitative and qualitative test for visual neglect. International Journal of Clinical Neuropsychology, 11(2): 49e54, 1989. Geminiani G, Camaschella E, Mariani C, Alberoni M, and Farina E. Direction and position factors in performance of line extension task by unilateral neglect subjects. Neuropsychologia, 40(11): 1834e1840, 2002. Goodale MA and Milner AD. Separate visual pathways for perception and action. Trends in Neurosciences, 15(1): 20e25, 1992. Halligan PW and Marshall JC. Spatial compression in visual neglect: A case study. Cortex, 27(4): 623e629, 1991. Hayes Ae and Freyd JJ. Representational momentum when attention is divided. Visual Cognition, 9(1/2): 8e27, 2002. Heilman KM. Neglect and related disorders. In Heilman KM and Valenstein E (Eds), Clinical Neuropsychology. Oxford University Press, 1979: 268e307. Heilman KM, Bowers D, Coslett HB, Whelan H, and Watson RT. Directional hypokinesia: Prolonged reaction times for leftward movements in patients with right hemisphere lesions and neglect. Neurology, 35(6): 855e859, 1985.
1327
Heilman KM and Van Den Abell T. Right hemisphere dominance for attention: The mechanism underlying hemispheric asymmetries of inattention (neglect). Neurology, 30(3): 327e330, 1980. Himmelbach M and Karnath Ho. Goal-directed hand movements are not affected by the biased space representation in spatial neglect. Journal of Cognitive Neuroscience, 15(7): 972e980, 2003. Hubbard TL. Representational momentum and related displacements in spatial memory: A review of the findings. Psychonomic Bulletin and Review, 12(5): 822e851, 2005. Husain M, Mannan S, Hodgson T, Wojciulik E, Driver J, and Kennard C. Impaired spatial working memory across saccades contributes to abnormal search in parietal neglect. Brain, 124(5): 941e952, 2001. Husain M and Rorden C. Non-spatially lateralized mechanisms in hemispatial neglect. Nature Reviews Neuroscience, 4(1): 26e36, 2003. Incoccia C, Doricchi F, Galati G, and Pizzamiglio L. Amplitude and speed change of the optokinetic response in patients with and without neglect. Neuroreport, 6(16): 2137e2140, 1995. Irving-Bell L, Small M, and Cowey A. A distortion of perceived space in patients with right-hemisphere lesions and visual hemineglect. Neuropsychologia, 37(8): 919e925, 1999. Ishiai S, Sugishita M, Watabiki S, Nakayama T, Kotera M, and Gono S. Improvement of left unilateral spatial neglect in a line extension task. Neurology, 44(2): 294e298, 1994. Karnath HO, Dick H, and Konczak J. Kinematics of goal-directed arm movements in neglect: Control of hand in space. Neuropsychologia, 35(4): 435e444, 1997. Karnath HO and Ferber S. Is space representation distorted in neglect? Neuropsychologia, 37(1): 7e15, 1998. Kerkhoff G. Multiple perceptual distortions and their modulation in leftsided visual neglect. Neuropsychologia, 38(7): 1073e1086, 2000. Kerzel D. Eye movements and visible persistence explain the mislocalization of the final position of a moving target. Vision Research, 40(27): 3703e3715, 2000. Kerzel D. Attention maintains mental extrapolation of target position: Irrelevant distractors eliminate forward displacement after implied motion. Cognition, 88(1): 109e131, 2003. Kerzel D, Jordan JS, and Musseler J. The role of perception in the mislocalization of the final position of a moving target. Journal of Experimental Psychology: Human Perception and Performance, 27(4): 829e840, 2001. Latini Corazzini L, Geminiani G, Stucchi N, Gindri P, and Cremasco L. Visual acceleration and spatial distortion in right brain-damaged patients. Experimental Brain Research, 161(3): 276e284, 2005. Loetscher T and Brugger P. Random number generation in neglect patients reveals enhanced response stereotypy, but no neglect in number space. Neuropsychologia, 47(1): 276e279, 2009. Lynch JC and McLaren JW. Optokinetic nystagmus deficits following parieto-occipital cortex lesions in monkeys. Experimental Brain Research, 49(1): 125e130, 1983. McGeorge P, Beschin N, and Della Sala S. Representing target motion: The role of the right hemisphere in the forward displacement bias. Neuropsychology, 20(6): 708e715, 2006. Mesulam M. A cortical network for directed attention and unilateral neglect. Annals of Neurology, 10(4): 309e325, 1981. Milner Ad and Harvey M. Distortion of size perception in visuospatial neglect. Current Biology, 5(1): 85e89, 1995. Nico D, Galati G, and Incoccia C. The endpoints’ task: An analysis of length reproduction in unilateral neglect. Neuropsychologia, 37(10): 1181e1188, 1999. Ogden JA. Anterior-posterior interhemispheric differences in the loci of lesions producing visual hemineglect. Brain and Cognition, 4(1): 59e75, 1985. Perri R, Bartolomeo P, and Gainotti G. Lack of impairments on leftward and rightward line extension tasks in neglect patients. International Journal of Neuroscience, 103(1): 101e113, 2000.
1328
c o r t e x 4 8 ( 2 0 1 2 ) 1 3 2 0 e1 3 2 8
Pisella L, Berberovic N, and Mattingley JB. Impaired working memory for location but not for colour or shape in visual neglect: A comparison of parietal and non-parietal lesions. Cortex, 40(2): 379e390, 2004. Posner MI, Walker JA, Friedrich FJ, and Rafal RD. Effects of parietal injury on covert orienting of attention. Journal of Neuroscience, 4(7): 1863e1874, 1984. Rode G, Michel C, Rossetti Y, Boisson D, and Vallar G. Left size distortion (hyperschematia) after right brain damage. Neurology, 67(10): 1801e1808, 2006. Rode G, Revol P, Rossetti Y, and Vallar G. 3D left hyperschematia after right brain damage. Neurocase, 14(5): 369e377, 2008. Rossit S, Malhotra P, Muir K, Reeves I, Duncan G, Birschel P, and Harvey M. The neural basis of visuomotor deficits in
hemispatial neglect. Neuropsychologia, 47(10): 2149e2153, 2009a. Rossit S, Muir K, Reeves I, Duncan G, Birschel P, and Harvey M. Immediate and delayed reaching in hemispatial neglect. Neuropsychologia, 47(10): 1563e1572, 2009b. Thornton Im and Hubbard Tl (Eds), Representational Momentum: New Findings, New Directions. Hove, U.K: Psychology Press, 2002. Vuilleumier P, Sergent C, Schwartz S, Valenza N, Girardi M, Husain M, and Driver J. Impaired perceptual memory of locations across gaze-shifts in patients with unilateral spatial neglect. Journal of Cognitive Neuroscience, 19(8): 1388e1406, 2007. Watson RT, Miller BD, and Heilman KM. Nonsensory neglect. Annals of Neurology, 3(6): 505e508, 1978.
