- Email: [email protected]
Line bisection performances of 650 normal children
Neuropsychologia 38 (2000) 886±895
www.elsevier.com/locate/neuropsychologia
Line bisection performances of 650 normal children Peter van Vugt a,*, Ingrid Fransen a, Wouter Creten b, Philippe Paquier c, d a
Unit of Neurolinguistics, CBL-UIA, Universiteitsplein 1, B-2610, Wilrijk, Belgium Laboratory for Biomedical Physics and Statistics, University of Antwerp (RUCA), Belgium c Department of ENT-Surgery, School of Medicine, University of Antwerp (UIA), Belgium d Department of Neurology, University Hospital Erasme, Free University of Brussels (ULB), Belgium b
Received 17 September 1998; received in revised form 28 June 1999; accepted 25 July 1999
Abstract When bisecting lines, an important number of brain damaged patients tend to place their bisection marks in the hemispace ipsilateral to their lesion. Biases have also been reported in normal adults. In vertical bisection both patients and normal subjects present with upward shifts, although a downward displacement may occur eventually. Surprisingly, little is known on line bisection (LB) in normal or brain damaged children. A total of 650 subjects, aged 7±12 years, performed a horizontal and vertical LB task with their preferred hand. Asymmetry indices (AIs) were used to measure directional bias. Unsigned AIs served to evaluate accuracy and mastery of the LB skill. In vertical bisection a general and signi®cant upward bias was found, whereas in horizontal bisection subject (gender, handedness, utilized hand, age) and stimulus variables (orientation, length, position) yielded signi®cantly dierent AIs. Although with increasing age signi®cantly increasing accuracy was observed, none of the participating children mastered LB to mathematical precision. Dierences in IQ-level and attention test score did not yield signi®cantly dierent AIs. Impact from reading pro®ciency could not be demonstrated. It is suggested that stimulus length eect results are compatible with the Halligan and Marshall [Halligan, P., and Marshall, J. Toward a principled explanation of unilateral neglect. Cognitive Neuropsychology, 1994, 11, 167±206] model of hemispatial neglect. Moreover, data may support the hypothesis of greater hemispheric specialization of visuo-spatial skills in boys than in girls. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Hemispatial neglect; Children; Normed data; Hemispheric lateralization
1. Introduction Hemispatial neglect (HN) is an acquired cluster of symptoms characterized by a failure to orient, report or respond to stimuli located in one hemispace [21,49]. The phenomenon may or may not occur with confusion, motor weakness or visual ®eld defects. Left HN following right hemisphere stroke is associated with poor functional recovery [19,49]. Most often, head and body midline are used to divide space in two half-®elds. Traditionally, HN is detected and evaluated along the horizontal dimension, although it may also occur along the vertical (with the * Corresponding author. Tel.: +32-3-820-2960; fax: +32-3-8202957.
upper or lower part neglected) and radial (with underestimation of far or near parts) axes [41,47,57]. Degree and modality may vary: auditory, olfactory, tactile and visual HN have been reported [19,53]. It appears that HN is an attentional rather than a sensory disorder [25,36,42], but some evidence suggests that visuomotor [7,10] or representational [2,24] components might be implicated as well. Typically, left HN is reported in adult (righthanded) patients with right parietal brain damage, but cases of HN due to right frontal lobe [31,32], bilateral parieto-occipital [41,58], subcortical [12,16] or callosal [26] lesions have also been reported. Neglect behaviour due to damage of the language dominant hemisphere is thought to be masked by aphasia [29]. However, the higher incidence of left HN has also been explained as
0028-3932/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 9 9 ) 0 0 1 3 0 - X
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
887
Fig. 1. Demographic data of the examined population: L, left; R, right; A, ambidexterous.
the result of the orientational bias of each hemisphere interfering with global visuo-spatial specialization of the right hemisphere [19]. Left brain damaged patients dispose of the right-hemispheric systems that guarantee ``global perception'', especially in the left hemispace. During clinical examination their visuo-spatial de®cits might be less visible than those of right brain damaged patients, suering from disturbed ``detail perception'', predominantly in the left hemispace [19,53]. Over the years, the assessment of HN took the most various forms: daisy or clock drawing [14], line or letter cancellation tasks, geometric ®gure cancellation tasks with target and non-target conditions, sentence± picture mapping [29], Wundt±Jastrow illusion evaluation [34], graphomotor and tactile line bisection (LB) tasks. LB is considered a standard test and is used worldwide [37] as midpoint displacements have been documented quite in detail in a large number of brain damaged patients [6]. Moreover, performance of normal subjects on LB tasks has also been studied. Shelton et al.'s [47] normal subjects exhibited no biases. On the other hand, some authors [3,4] found a discrete but consistent rightward bias in normal adults, whereas others [17,33] observed an equal number of leftward and rightward displacements in their normal controls. Turnbull and McGeorge [52], in turn, reported a consistent leftward bias in their normal subjects. Some authors [9,18,33] reported a signi®cant de-
terioration of performance with increasing line length whereas others [20,38,46,59] did not. Although some authors failed to demonstrate the existence of HN in children with hemiplegic cerebral palsy [28], the occurrence of HN in children who suffered a unilateral cerebral lesion has nevertheless been documented [13,23,51]. Ferro et al. [13] have suggested that HN in children may show rapid recovery; HN may thus be overlooked if not searched for during the acute stage. However, the patient described by Johnston and Shapiro [23], who sustained a left hemisphere subdural haematoma after birth, still presented at the age of 14 with right sided visual HN interfering with daily life activities. Consequently, one may wonder whether the notion of rapid recovery in childhood HN should not be tempered on the analogy of the traditional views on the transient nature of acquired childhood aphasia [39,55]. In adults, data from clinical [1] and experimental studies in normal subjects [3,27,35,45] allow to increasingly document normal LB behaviour. However, little is known about LB in childhood: only scattered studies [13,23,40] incidentally report the use of LB tasks without giving details on test performance and normative studies have used but limited samples of subjects [10]. Well established norms would certainly help to detect and evaluate the up till now only infrequently reported NH in childhood.
888
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
This paper examines three aspects of LB performances in normal children. 1. As in normal adult subjects dierences in stimulus orientation, length, or position appear to result in important shifts of the subjective midpoint [3,4,9,33], we investigated whether these eects could also be observed in children. 2. As it has been suggested that biases observed in normal adult subjects may dier according to gender, handedness, hand used to perform the task, and of course also according to age [3,4,45], the in¯uence of these demographic variables on children's performances was also evaluated. 3. As a relationship between the magnitude of midpoint displacements and the mastering of attentional coping strategies has been described in adults [2,9,24,25,42,44], we wanted to know whether dierences in attention-test-score Section 3.3, or reading pro®ciency Section 3.4 have an impact on LB performances in children?
2. Method 2.1. Subjects and standard tests Three urban schools participated with an almost identical number of subjects in the present study. The schools were situated in socio-economically dierent districts. Thus data were obtained from an aselect representative sample of 716 subjects. Children known with a history of learning disability or presenting with academic underachievement due to medical problems (e.g. epilepsy, cardiovascular condition, diabetes, renal disease, cancer [43]) were excluded; 650 children (302 boys, 348 girls), with mean age 9.6 years (range: 6.11± 12.7 years), participated in the actual study. Detailed demographic data are shown in Fig. 1. Because of their small number (n = 18), the 12-year-olds were grouped with the 11-year-olds (11+). The children's individual assessment consisted of: an IQ test [8], two attention tests (the Symbol Digit Modalities Test [48] further abbreviated A1 and the Wechsler Digit Symbol Test [56] further abbreviated A2), two reading tests (the BVL-II [50] further abbreviated R1 and the Brus 1 Minuut Test [5] further abbreviated R2), and vocabulary, reading-comprehension, arithmetic and maths tests from School-vorderingentest BVL [11]. Since persisting slow reactions have been documented as an important component in HN [44], attention tests were chosen that are sensitive to visuo-perceptual impairments and general mental slowing [30]. In the Flemish-speaking part of Belgium both reading protocols (R1 and R2) are fre-
quently used tests that are reliably validated [54]. The Flemish school progress examination Schoolvorderingentest BVL [11] is daily used by school advisory services. 2.2. Experiment 2.2.1. Materials In this new experimental design 45 horizontal and 45 vertical black lines were presented on DIN-A4 sheets, each page containing 9±12 lines. Lines were 3 mm wide and of three dierent lengths (40, 80, and 120 mm). Within each length, category position and orientation were varied: 15 horizontal lines in leftsided position (HLP), 15 horizontal lines in centred position (HCP), 15 horizontal lines in right-sided position (HRP), 15 vertical lines in top position (VTP), 15 vertical lines in centred position (VCP), and 15 vertical lines in bottom position (VBP). The design was such that, by the end of the test, each quadrant of the visual ®eld had been exposed to the same amount of stimulus material. The capital letters A, B, F, H, O, S, X, and Z appeared in random order at the corners of the pages (Fig. 4). 2.2.2. Procedure Prior to the LB test, subjects were asked to point to or to read aloud the four letters on each page, in order to ensure the perception of the whole page. Then, subjects were instructed to mark the midpoint of each line with a ballpoint pen, using their preferred hand. The page orientation remained perpendicular upon the subjects' coronal plane during both the horizontal and vertical presentation. A ®ve item non-scored trial preceded the test. The distance from the right side of the horizontal lines (or the top of the vertical lines) to the subjects' mark was measured to 0.5 mm accuracy. In order to assess the subjects' directional bias an asymmetry index (AI) was calculated as: 2MD ÿ SL AI 100 SL where MD is measured distance from the right (top) end and SL is stimulus length. Positive AI values indicate an underestimation at the right (of horizontal lines) or on top (of vertical lines), negative AI values indicate an underestimation at the left or at the bottom. Individual mean AI's were calculated for all horizontal and vertical lines (horizontal asymmetry index (HAI) and vertical asymmetry index (VAI)), for the three length categories, as well as for the dierent stimulus positions. In order to avoid biases resulting from averaging positive and negative values, the unsigned AIs (vAIsv) were used to evaluate the subjects' accuracy. Moreover, mastery of the line
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
889
Fig. 2. Mean asymmetry indices and dierences triggered by stimulus-related characteristics: AI, asymmetry index; SD, standard deviation; t, tvalues; _ or ^, 0.001 < P < 0.05; = or ;, P < 0.001. w
890
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
Fig. 3. Mean asymmetry indices and dierences triggered by demographic variables: AI, asymmetry index; SD, standard deviation; vAIv, unsigned asymmetry index; t, t-values of the mastery analysis are bold typed when P < 0.05. Newman±Keuls multiple range tests were used (in the analysis of the directional bias and in the accuracy analysis) to separate the dierent groups into clusters, the mean AIs within a cluster not being signi®cantly dierent [15].
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
bisection skill was assessed by comparing the vAIsv with zero (i.e. the objective midpoint). To this end a null set of data was constructed. 2.2.3. Statistical analysis The impact on AIs of stimulus length, position and orientation was evaluated by means of Pearson correlation and single t-tests. Relationships between AI (or vAIv) and demographic or neuropsychological variables were examined using both parametric (MANOVA and Newman±Keuls test [15]) and non-parametric (Chisquare test) techniques. Additional single t-tests on vAIv were performed to determine whether the subject's marks dier signi®cantly from the mathematical midpoints [45,47].
3. Results Numerical data not included in the text are given in Figs. 2 and 3. 3.1. Stimulus characteristics The mean AIs yielded by dierent stimulus lengths, positions and orientation are shown in Fig. 2. The impact of stimulus characteristics manipulation on accuracy is illustrated by the interbox-connections and their labels. In the same Fig. 2, the within-box t-values result from the mastery analysis. One-line connections and regular underscoring re¯ect a signi®cance level with 0.05 > P > 0.001. Double connections and double underscoring correspond to a signi®cance level with P < 0.001. Fig. 2 clearly shows that by manipulating length, position and orientation signi®cantly dierent AIs are obtained, the only exception being the VTP and VCP which yield comparable AIs. Furthermore, the mastery analysis documents that a signi®cant shift to the left occurred in horizontal bisection, whereas in vertical bisection, an even more signi®cant upward displacement was seen. Displacements of the subjective midpoint became more important as stimulus length increased. In HLP and HCP lines a signi®cant leftward shift of the subjective midpoint occurred. On the other hand, HRP lines yielded a non-signi®cant displacement to the right. In all vertical lines, the subjective midpoint shifted signi®cantly toward the top. This shift was smaller in VBP lines than in VCP or VTP lines. Further analysis showed that there was no signi®cant correlation between HAI and VAI (r(650)=ÿ0.02, ns) and that there was no length position interaction in horizontal (F(4,27)=0.78, ns) nor in vertical bisection (F(4,27)=1.69, ns).
891
3.2. Demographic variables 3.2.1. Directional bias From Fig. 3 it is clear that, when bisecting horizontal lines boys tended to underestimate the left part (HAI=ÿ0.29) and girls the right part (HAI=0.92). The dierence is signi®cant (F(1,648)=16.75, P < 0.001). Therefore, boys and girls are represented as clearly separate groups. This eect was not due to averaging, since there were signi®cantly more boys with a negative HAI and girls with a positive HAI (w 2=4.40, P = 0.03). Handedness did not signi®cantly in¯uence VAIs, but yielded signi®cant dierences in HAIs. Right-handed subjects obtained signi®cantly smaller HAIs than nonright-handed subjects (i.e. the left-handed and ambidexterous subjects, in whom comparable HAIs were obtained). A further 2 3 MANOVA showed a signi®cant handedness stimulus position interaction (Wilks' lambda=0.95, F(3,646), P < 0.001). Nonright-handed subjects showed a signi®cant leftward bias in all three stimulus positions (HAIHLP=2.34, t(99)=ÿ5.98, P < 0.001; HAIHCP=2.92, t(99)=ÿ6.65, P < 0.001; HAIHRP=1.40, t(99)=ÿ2.52, P = 0.01). In right-handed subjects HLP lines yielded a signi®cant leftward bias (HAIHLP=0.77, t(549)=ÿ3.87, P < 0.001), HRP lines triggered a signi®cant rightward shift (HAIHRP=ÿ0.55, t(549)=3.07, P = 0.002), HCP lines yielding no signi®cant displacement (HAIHCP=0.01, t(549)=0.11, ns). The use of the right hand to perform the task resulted in signi®cantly smaller HAIs. In vertical bisection there was no hand eect. A 2 3 MANOVA showed a signi®cant hand stimulus position interaction (Wilks' lambda=0.95, F(3,646), P < 0.001). The pattern of results is identical to the one of the handedness stimulus position interaction. Dierences in age yielded signi®cant eects in HAI (F(4,645)=2.42, P < 0.05) and VAI (F(4,645)=12.19, P < 0.001). In general, younger subjects tended to underestimate more than the older ones. To verify which dierences were signi®cant among the ®ve mean AIs, a Newman±Keuls Multiple Range Test [15] was used. The result of this analysis is displayed in Fig. 3. In horizontal bisection none of the age groups dier signi®cantly with respect to mean AI, whereas in vertical bisection the mean AIs of the 7- and 8-year-olds diers signi®cantly from the mean AIs of the 9-, 10and 11+-year-olds. There was no signi®cant gender handedness hand age interaction in horizontal nor in vertical bisection. From the entire sample the within subject variability was estimated as SD=11.37. 3.2.2. Accuracy When bisecting horizontal lines, boys are signi®-
892
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
3.4. Reading Reading skills had no signi®cant impact on HAIs (R1: F(2,647)=0.59, ns; R2: F(2,647)=0.10, ns) nor on VAIs (R1: F(2,647)=0.52, ns; R2: F(2,647)=1.63, ns). Subjects exposed to bidirectional reading instruction (n = 11) did not present with signi®cantly dierent HAIs (F(1,648)=0.69, ns) nor VAIs (F(1,648)=0.16, ns) than their from left to right reading classmates. 4. Discussion
Fig. 4. Example of the stimulus design.
cantly more accurate than girls, but when they bisect vertical lines, both groups perform equally well (Fig. 3). In terms of accuracy, there is no handedness eect in horizontal nor in vertical bisection (Fig. 3). A signi®cant hand eect is only seen in vertical bisection: when the right hand has been used, the VAIs are smaller, and thus the performance more accurate (Fig. 3). Accuracy clearly increases with age: vHAIsv vVAIsv (F(4,645)=9.63, P < 0.001) and (F(4,645)=12.83, P < 0.001). However, the pace by which signi®cantly dierent levels of accuracy are reached is dierent in horizontal and in vertical bisection (Fig. 3). 3.2.3. Mastery Both boys' and girls' vHAIsv deviated signi®cantly from the objective midpoint (as can be deduced from the bold typed t-values in Fig. 3). And also when bisecting vertical lines both boys and girls signi®cantly underestimated the bottom part. In terms of mastery, no handedness or hand eect was demonstrated. All age groups displaced the marks signi®cantly, although, with increasing age a tendency towards increasing mastery was observed. 3.3. Attention Dierences in attention performance-level were not re¯ected in dierent HAIs (A1: F(2,647)=1.04, ns; A2: F(2,647)=0.49, ns) nor in VAIs (A1: F(2,647)=0.23, ns; A2: F(2,647)=0.04, ns). Furthermore, no signi®cant dierences were found when IQ-levels were compared with HAIs (F(3,646)=1.17, ns) and VAIs (F(3,646)=0.95, ns).
Results obtained with our protocol show that normal children's LB performances are subject to biases. In horizontal lines left- and rightward shifts of the subjective midpoint occur, along the vertical axis an upward shift seems to be the rule. HAIs and VAIs are independent: performances in the horizontal dimension do not allow predictions as to the vertical bisection and vice versa. Moreover, our childrens' VAIs are signi®cantly larger than their HAIs, a ®nding which corroborates previously reported data in normal adults [47]. Thus, horizontal and vertical LBs seem to require separate skills in both young and adult normal subjects. The eects discussed below, seem to consolidate this distinction, especially with respect to children. Indeed, performances become more accurate with age, and LB may perhaps be mastered much later in life, but certainly not before the age of 11. But strikingly, AIs and accuracy measures do not evolve at the same pace in horizontal and in vertical LB. Moreover, in horizontal bisection only girls and non-right-handers present with an important directional bias, whereas a sex or handedness eect is absent in vertical bisection. The importance of individual variations in LB has repeatedly been stressed [2,33]. Therefore, it is important to note that the sex linked pattern found here is not due to an averaging eect: there are indeed more girls with a rightward bias. With respect to the impact of handness and hand performance, one can not but observe that in our study both eects coincide. This may be due to the aselect nature of our population (and thus to the not very impressive number of non-right-handed performers). Moreover, subjects completed the forms only once, using one hand, which makes our data far from ideal in view of the assessment of the hand eect. Nevertheless, some observations can be made. Along the vertical axis, our right-handed, left-handed and ambidexterous subjects obtain comparable AIs and in terms of accuracy the three groups perfom equally well. This is in contrast with Scarisbrick et al.'s [45] normal left-handed adults, who bisected vertical lines far more accurately than their right-handed participat-
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
ing colleagues. Moreover, Scarisbrick et al. [45] failed to demonstrate a hand eect in vertical LB, whereas our subjects who used their right hand are signi®cantly more accurate than those who held the pencil in their left hand. Both these dierences may be linked to the age of our subjects and to the fact that vertical bisection assumes a skill acquired apparently after the age of 11. Moreover, a lot of variability in that skill has been described in adults: Rapcsak et al.'s [41] normal subjects, when bisecting vertical lines, tended to underestimate slightly the top part, whereas Shelton et al.'s [47] normal subjects tended to underestimate the bottom part. In accordance with the authors who predicted (on the basis of Weber's law) and reported signi®cant length eects in normal adults [9,18], the signi®cance of the misbisection increases together with line length in our population. In addition to length, stimulus position is another source of variation. Nichelli et al. [38] found that, in horizontal bisection, normal adults displace the subjective midpoint toward the end opposite to the side of hemispace presentation (i.e. body midline attraction). Our right-handed subjects (as well as all right-hand bisectors) display the opposite phenomenon: stimuli presented in the left hemi-space yielded a signi®cant leftward bias, lines presented to the right resulted in a signi®cant rightward displacement of the marks, whereas only the centred stimuli yielded no signi®cant bias. Our non-right-handed subjects (as well as those who performed the task with their left hand) misbisected signi®cantly to the left in all three conditions. The ``excentric'' shift of the right-handers' marks is compatible with the long standing view that each hemisphere orients preferentially toward the contralateral hemispace [19,45]. Stimuli presented in the left hemi-space cause a more important selective activation of the right hemisphere resulting in an enhancement of the left visual ®eld, and thus a leftward deviation of the midpoint, and vice versa. Our nonright-handed subjects showed a stronger leftward bias than our right-handed children, as was the case in Turnbull and McGeorge's [52] normal adults. This might well be the result of a ``dominant'' right hemisphere extending its leftward bias over its usual scan®eld; i.e. both hemispaces [19]. Subjects with lower scores on substitution tests performed equally well on LB as subjects with higher scores. This may indeed be seen as a remarkable ®nding since in patients with HN, neglect seems to be associated with a general de®cit in the intensity aspects of attention, i.e. with de®cient generation or maintenance of alertness [44]. Moreover, HN phenomena seem to improve with sustained attention training [42]. One might thus have expected decreasing AI with increasing substitution test results. The reason why this trend was not observed might well be that substitution tests
893
are not ``pure'' measures of attention [30] and that they are not suitable measures for the vigilance and spatial attention implicated in LB [19,42]. On the other hand, this ®nding suggests not to overestimate the possibly ``distracting nature'' of our complex protocol in which several lines are presented on one sheet. Another way of exploring the relationship between LB and attentional processes is based upon Chokron and Imbert's [9] ®ndings that LB performance and reading behaviour seem to be related, because reading instruction directs (single or multiple component) spatial attention toward one hemispace [9]. Therefore, we expected to ®nd signi®cantly dierent HAIs in poor and strong readers. However, in our population, the in¯uence of reading on LB was not as prominent as in previous reports [9]. This may be explained by the fact that we compared reading pro®ciency with LB performance in predominantly left to right readers, whereas Chokron and Imbert [9] studied the in¯uence of reading direction upon LB results in left to right readers as well as in right to left readers. Moreover, their 60 French subjects were older (range: 13.9±44.3 years) [9]. However, the relevance of this last argument depends upon whether or not future research is able to discover a signi®cant change in horizontal LB performance in subjects older than 12 years. In summary, horizontal and vertical LB appear to be distinct tasks appealing on dierent skills. This may incite to look for double dissociations in HN patients. It was found that LB is a skill which seems not to be mastered before the age of 11, and as even normal adults display signi®cant biases, certainly along the vertical axis [27], it remains doubtful whether LB ever may attain some degree of ``mathematical precision''. Furthermore, length and position seem to act as intensi®ers of the asymmetry phenomenon. If it is accepted that tasks which involve the appreciation of visual space selectively activate the right hemisphere, the overall right underestimation we found may be explained as resulting from an enhancement of the left visual ®eld and consequently a displacement of the subjective midpoint to the left. As in adult non-righthanders [45,52], this right hemisphere activation eect is also greater in non-dextral children. The fact that boys attain signi®cantly more accuracy in horizontal bisection than girls might possibly support the hypothesis of greater hemispheric specialization for visuospatial tasks in males than in females [22]. Acknowledgements This research was supported by the Provincie Antwerpen, the Belgische Stichting Roeping and the Antwerp University (UIA). We gratefully acknowledge the help of P. Wackenier and P. Bal. We wish to
894
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
thank Y. De Smet and J. Turnbull for reviewing the manuscript and M. Isacson for his kind bibliographical support. We also gratefully acknowledge the helpful suggestions of two anonymous reviewers.
[23] [24] [25]
References [1] Barton J, Black S. Line bisection in hemianopia. Journal of Neurology, Neurosurgery, and Psychiatry 1998;64:660±2. [2] 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. [3] Bowers D, Heilman K. Pseudoneglect: eects of hemispace on a tactile line bisection task. Neuropsychologia 1980;18:491±8. [4] Bradshaw J, Nettleton N, Hathan G, Wilson L. Bisecting rods and lines: eects of horizontal and vertical posture on left-side underestimation by normal subjects. Neuropsychologia 1985;23:421±5. [5] Brus B, Voeten H. EeÂn minuut test. Nijmegen: Berkhout BV, 1980. [6] Bruyer R. Neglects in hemineglect: a comment on the study of Bisiach et al. (1983). Brain and Cognition 1984;3:231±7. [7] Cantagallo A, Ferraresi G, Basaglia N. Visual attentive and premotor factors of visual neglect in the line bisection test. Journal of Clinical and Experimental Neuropsychology 1993;15:413. [8] Cattell RB. Measuring intelligence with the culture fair tests. Illinois: Institute for Personality and Ability Testing, 1973. [9] Chokron S, Imbert M. In¯uence of reading habits on line bisection. Cognitive Brain Research 1993;1:219±22. [10] Dellatolas G, Coutin T, DeAgostini M. Bisection and perception of horizontal lines in normal children. Cortex 1996;32:705± 15. [11] De Vos L, Lievens S, Vallaeys A. Schoolvorderingentest BVL, voor de leerjaren 2, 3, 4, 5, en 6 van het gewoon lager onderwijs en voor het buitengewoon onderwijs. Tessenderlo: Boekenprokuur Broeders van Liefde, 1986. [12] Ferro JM, Kertesz A, Black SE. Subcortical neglect: quantitation, anatomy, and recovery. Neurology 1987;37:1487±92. [13] Ferro JM, Martins IP, Tavora L. Neglect in children. Annals of Neurology 1984;15:281±4. [14] Freedman M, Leach L, Kaplan E, Winocur G, Shulman K, Delis D. Clock drawing: A neuropsychological analysis. MultiHealth Systems Inc, 1994. [15] Godfrey K. Statistics in practice: comparing the means of several groups. New England Journal of Medicine 1985;313:1450± 6. [16] Graveleau Ph, Viader F, Masson M, Cambier J. NeÂgligence thalamique. Revue Neurologique (Paris) 1986;142:425±30. [17] Halligan P, Manning L, Marshall J. Individual variation in line bisection: a study of four patients with right hemisphere damage and normal controls. Neuropsychologia 1990;28:1043±51. [18] Halligan P, Marshall J. Line bisection in visuo-spatial neglect: disproof of a conjecture. Cortex 1989;25:517±22. [19] Halligan P, Marshall J. Toward a principled explanation of unilateral neglect. Cognitive Neuropsychology 1994;11:167±206. [20] Heilman K, Bowers D, Watson R. Pseudo-neglect in a patient with partial callosal disconnection. Brain 1984;107:519±32. [21] Heilman K, Watson R, Valenstein E. Neglect and related disorders. In: Heilman K, Valenstein E, editors. Clinical neuropsychology. New York: Oxford University Press, 1985. p. 243±93. [22] Hiscock M, Israelian M, Inch R, Jacek C, Hiscock-Kalil C. Is there a sex dierence in human laterality? II: an exhaustive survey of visual laterality studies from six neuropsychology jour-
[26]
[27] [28] [29] [30] [31]
[32]
[33] [34] [35] [36] [37] [38] [39] [40]
[41] [42]
[43] [44]
nals. Journal of Clinical and Experimental Neuropsychology 1995;17:590±610. Johnston C, Shapiro E. Hemi-inattention resulting from left hemisphere brain damage during infancy. Cortex 1986;22:279± 87. Jeerakathil T, Kirk A. A representational vertical bias. Neurology 1994;44:703±6. Kaplan RF, Verfaellie M, Meadows M-E, Caplan LR, Pessin MS, DeWitt LD. Changing attentional demands in left hemispatial neglect. Archives of Neurology 1991;48:1263±6. Kashiwagi T, Kashiwagi A, Nishikawa T, Tanabe H, Okuda JI. Hemispatial neglect in a patient with callosal infarction. Journal of Clinical and Experimental Neuropsychology 1989;11:377. Kashmere J, Kirk A. The complex interaction of normal biases in line bisection. Neurology 1997;49:887±9. Katz N, Cermak S, Shamir Y. Unilateral neglect in children with hemiplegic cerebral palsy. Perceptual and Motor Skills 1998;86:539±50. Lecours AR, Lupien S, Parente MA. Visual attention in left sylvian strokes. Journal of Neurolinguistics 1989;4:255±71. Lezak M. Neuropsychological assessment. New York: Oxford University Press, 1995. Maeshima S, Funahashi K, Ogura M, Itakura T, Komai N. Unilateral spatial neglect due to right frontal haematoma. Journal of Neurology, Neurosurgery, and Psychiatry 1994;57:89±93. Maeshima S, Terada T, Nakai K, Nishibayashi H, Ozaki F, Itakura T, Komai N. Unilateral spatial neglect due to haemorrhagic contusion in the right frontal lobe. Journal of Neurology 1995;242:613±7. Manning L, Halligan P, Marshall J. Individual variation in line bisection: a study of normal subjects with application to the interpretation of visual neglect. Neuropsychologia 1990;28:647±55. Massironi M, Antonucci G, Pizzamiglio L, Vitale MV, Zoccolotti P. The Wundt±Jastrow illusion in the study of spatial hemi-inattention. Neuropsychologia 1988;26:161±6. McCourt M, Olafson C. Cognitive and perceptual in¯uences on visual line bisection: psychophysical and chronometric analyses of pseudoneglect. Neuropsychologia 1997;35:369±80. Mennemeier M, Wertman E, Heiman K. Neglect of near peripersonal space: evidence for multidirectional attentional systems in humans. Brain 1992;115:37±50. Monaghan P, Shillcock R. The cross-over eect in unilateral neglect, modelling detailed data in the line-bisection task. Brain 1998;121:907±21. 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. Paquier PF, Van Dongen HR. Review of research on the clinical presentation of acquired childhood aphasia. Acta Neurologica Scandinavica 1996;93:428±36. Paquier P, Van Mourik M, Van Dongen H, CatsmanBerrevoets C, Creten W, Stronks D. Clinical utility of the judgment of line orientation test and facial recognition test in children with acquired cerebral lesions. Journal of Child Neurology 1999;14:243±8. Rapcsak SZ, Cimino CR, Heilman KM. Altitudinal neglect. Neurology 1988;38:277±81. Robertson IH, TegneÂr R, Tham K, Lo A, Nimmo-Smith I. Sustained attention training for unilateral neglect: theoretical and rehabilitation implications. Journal of Clinical and Experimental Neuropsychology 1995;17:416±30. Rourke B, Fisk J, Strang J. Neuropsychological assessment of children. New York: Guilford Press, 1986. Samuelsson H, Hjelmquist E, Jensen C, Ekholm S, Blomstrand C. Nonlateralized attentional de®cits: an important component
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
[45] [46] [47] [48] [49] [50] [51] [52]
behind persisting visuospatial neglect? Journal of Clinical and Experimental Neuropsychology 1998;20:73±88. Scarisbrick D, Tweedy J, Kulanski G. Hand preference and performance eects on line bisection. Neuropsychologia 1987;25:695±9. Schenkenberg T, Bradford D, Ajax E. Line bisection and unilateral visual neglect in patients with neurological impairment. Neuropsychologia 1987;25:695±9. Shelton PA, Bowers D, Heilman KM. Peripersonal and vertical neglect. Brain 1990;113:191±205. Smith A. Symbol digit modalities test. Los Angeles: Western Psychological Services, 1973. Stone S, Halligan P, Marshall J, Greenwood R. Unilateral neglect: a common but heterogeneous syndrome. Neurology 1998;50:1902±5. Tanghe M. Schoolvorderingentest BVL. Tessenderlo: Boekenprokuur Broeders van Liefde, 1971. Thompson N, Ewing-Cobbs L, Fletcher J, Miner M, Levin H. Left unilateral neglect in a preschool child. Developmental Medicine and Child Neurology 1991;33:636±44. Turnbull O, McGeorge P. Lateral bumping: a normal-subject
[53] [54] [55] [56] [57] [58]
[59]
895
analog to the behaviour of patients with hemispatial neglect? Brain and Cognition 1998;37:31±3. UmiltaÁ C. Domain-speci®c forms of neglect. Journal of Clinical and Experimental Neuropsychology 1995;17:209±19. Van den Bos K, lutje Spelberg H, Scheepstra A, De Vries J. De klepel, vorm A en B: een test voor de leesvaardigheid van pseudowoorden. Nijmegen: Berkhout, 1994. Van Dongen H. Clinical aspects of acquired aphasia and dysarthria in childhood. Delft: Eburon, 1988. Van Haasen PP, et al. De WISC-R, Nederlandse uitgave. Lisse: Swets and Zeitlinger, 1986. Vuilleumier P, Valenza N, Mayer E, Reverdin A, Landis T. Near and far visual space in unilateral neglect. Annals of Neurology 1998;43:406±10. Weintraub S, Daner K, Ahern G, Price B, Mesulam M-M. Right sided hemispatial neglect and bilateral cerebral lesions. Journal of Neurology, Neurosurgery, and Psychiatry 1996;60:342±4. Werth R, PoÈppel E. Compression and lateral shift of mental coordinate systems in a line bisection task. Neuropsychologia 1988;26:741±5.
www.elsevier.com/locate/neuropsychologia
Line bisection performances of 650 normal children Peter van Vugt a,*, Ingrid Fransen a, Wouter Creten b, Philippe Paquier c, d a
Unit of Neurolinguistics, CBL-UIA, Universiteitsplein 1, B-2610, Wilrijk, Belgium Laboratory for Biomedical Physics and Statistics, University of Antwerp (RUCA), Belgium c Department of ENT-Surgery, School of Medicine, University of Antwerp (UIA), Belgium d Department of Neurology, University Hospital Erasme, Free University of Brussels (ULB), Belgium b
Received 17 September 1998; received in revised form 28 June 1999; accepted 25 July 1999
Abstract When bisecting lines, an important number of brain damaged patients tend to place their bisection marks in the hemispace ipsilateral to their lesion. Biases have also been reported in normal adults. In vertical bisection both patients and normal subjects present with upward shifts, although a downward displacement may occur eventually. Surprisingly, little is known on line bisection (LB) in normal or brain damaged children. A total of 650 subjects, aged 7±12 years, performed a horizontal and vertical LB task with their preferred hand. Asymmetry indices (AIs) were used to measure directional bias. Unsigned AIs served to evaluate accuracy and mastery of the LB skill. In vertical bisection a general and signi®cant upward bias was found, whereas in horizontal bisection subject (gender, handedness, utilized hand, age) and stimulus variables (orientation, length, position) yielded signi®cantly dierent AIs. Although with increasing age signi®cantly increasing accuracy was observed, none of the participating children mastered LB to mathematical precision. Dierences in IQ-level and attention test score did not yield signi®cantly dierent AIs. Impact from reading pro®ciency could not be demonstrated. It is suggested that stimulus length eect results are compatible with the Halligan and Marshall [Halligan, P., and Marshall, J. Toward a principled explanation of unilateral neglect. Cognitive Neuropsychology, 1994, 11, 167±206] model of hemispatial neglect. Moreover, data may support the hypothesis of greater hemispheric specialization of visuo-spatial skills in boys than in girls. # 2000 Elsevier Science Ltd. All rights reserved. Keywords: Hemispatial neglect; Children; Normed data; Hemispheric lateralization
1. Introduction Hemispatial neglect (HN) is an acquired cluster of symptoms characterized by a failure to orient, report or respond to stimuli located in one hemispace [21,49]. The phenomenon may or may not occur with confusion, motor weakness or visual ®eld defects. Left HN following right hemisphere stroke is associated with poor functional recovery [19,49]. Most often, head and body midline are used to divide space in two half-®elds. Traditionally, HN is detected and evaluated along the horizontal dimension, although it may also occur along the vertical (with the * Corresponding author. Tel.: +32-3-820-2960; fax: +32-3-8202957.
upper or lower part neglected) and radial (with underestimation of far or near parts) axes [41,47,57]. Degree and modality may vary: auditory, olfactory, tactile and visual HN have been reported [19,53]. It appears that HN is an attentional rather than a sensory disorder [25,36,42], but some evidence suggests that visuomotor [7,10] or representational [2,24] components might be implicated as well. Typically, left HN is reported in adult (righthanded) patients with right parietal brain damage, but cases of HN due to right frontal lobe [31,32], bilateral parieto-occipital [41,58], subcortical [12,16] or callosal [26] lesions have also been reported. Neglect behaviour due to damage of the language dominant hemisphere is thought to be masked by aphasia [29]. However, the higher incidence of left HN has also been explained as
0028-3932/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 9 9 ) 0 0 1 3 0 - X
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
887
Fig. 1. Demographic data of the examined population: L, left; R, right; A, ambidexterous.
the result of the orientational bias of each hemisphere interfering with global visuo-spatial specialization of the right hemisphere [19]. Left brain damaged patients dispose of the right-hemispheric systems that guarantee ``global perception'', especially in the left hemispace. During clinical examination their visuo-spatial de®cits might be less visible than those of right brain damaged patients, suering from disturbed ``detail perception'', predominantly in the left hemispace [19,53]. Over the years, the assessment of HN took the most various forms: daisy or clock drawing [14], line or letter cancellation tasks, geometric ®gure cancellation tasks with target and non-target conditions, sentence± picture mapping [29], Wundt±Jastrow illusion evaluation [34], graphomotor and tactile line bisection (LB) tasks. LB is considered a standard test and is used worldwide [37] as midpoint displacements have been documented quite in detail in a large number of brain damaged patients [6]. Moreover, performance of normal subjects on LB tasks has also been studied. Shelton et al.'s [47] normal subjects exhibited no biases. On the other hand, some authors [3,4] found a discrete but consistent rightward bias in normal adults, whereas others [17,33] observed an equal number of leftward and rightward displacements in their normal controls. Turnbull and McGeorge [52], in turn, reported a consistent leftward bias in their normal subjects. Some authors [9,18,33] reported a signi®cant de-
terioration of performance with increasing line length whereas others [20,38,46,59] did not. Although some authors failed to demonstrate the existence of HN in children with hemiplegic cerebral palsy [28], the occurrence of HN in children who suffered a unilateral cerebral lesion has nevertheless been documented [13,23,51]. Ferro et al. [13] have suggested that HN in children may show rapid recovery; HN may thus be overlooked if not searched for during the acute stage. However, the patient described by Johnston and Shapiro [23], who sustained a left hemisphere subdural haematoma after birth, still presented at the age of 14 with right sided visual HN interfering with daily life activities. Consequently, one may wonder whether the notion of rapid recovery in childhood HN should not be tempered on the analogy of the traditional views on the transient nature of acquired childhood aphasia [39,55]. In adults, data from clinical [1] and experimental studies in normal subjects [3,27,35,45] allow to increasingly document normal LB behaviour. However, little is known about LB in childhood: only scattered studies [13,23,40] incidentally report the use of LB tasks without giving details on test performance and normative studies have used but limited samples of subjects [10]. Well established norms would certainly help to detect and evaluate the up till now only infrequently reported NH in childhood.
888
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
This paper examines three aspects of LB performances in normal children. 1. As in normal adult subjects dierences in stimulus orientation, length, or position appear to result in important shifts of the subjective midpoint [3,4,9,33], we investigated whether these eects could also be observed in children. 2. As it has been suggested that biases observed in normal adult subjects may dier according to gender, handedness, hand used to perform the task, and of course also according to age [3,4,45], the in¯uence of these demographic variables on children's performances was also evaluated. 3. As a relationship between the magnitude of midpoint displacements and the mastering of attentional coping strategies has been described in adults [2,9,24,25,42,44], we wanted to know whether dierences in attention-test-score Section 3.3, or reading pro®ciency Section 3.4 have an impact on LB performances in children?
2. Method 2.1. Subjects and standard tests Three urban schools participated with an almost identical number of subjects in the present study. The schools were situated in socio-economically dierent districts. Thus data were obtained from an aselect representative sample of 716 subjects. Children known with a history of learning disability or presenting with academic underachievement due to medical problems (e.g. epilepsy, cardiovascular condition, diabetes, renal disease, cancer [43]) were excluded; 650 children (302 boys, 348 girls), with mean age 9.6 years (range: 6.11± 12.7 years), participated in the actual study. Detailed demographic data are shown in Fig. 1. Because of their small number (n = 18), the 12-year-olds were grouped with the 11-year-olds (11+). The children's individual assessment consisted of: an IQ test [8], two attention tests (the Symbol Digit Modalities Test [48] further abbreviated A1 and the Wechsler Digit Symbol Test [56] further abbreviated A2), two reading tests (the BVL-II [50] further abbreviated R1 and the Brus 1 Minuut Test [5] further abbreviated R2), and vocabulary, reading-comprehension, arithmetic and maths tests from School-vorderingentest BVL [11]. Since persisting slow reactions have been documented as an important component in HN [44], attention tests were chosen that are sensitive to visuo-perceptual impairments and general mental slowing [30]. In the Flemish-speaking part of Belgium both reading protocols (R1 and R2) are fre-
quently used tests that are reliably validated [54]. The Flemish school progress examination Schoolvorderingentest BVL [11] is daily used by school advisory services. 2.2. Experiment 2.2.1. Materials In this new experimental design 45 horizontal and 45 vertical black lines were presented on DIN-A4 sheets, each page containing 9±12 lines. Lines were 3 mm wide and of three dierent lengths (40, 80, and 120 mm). Within each length, category position and orientation were varied: 15 horizontal lines in leftsided position (HLP), 15 horizontal lines in centred position (HCP), 15 horizontal lines in right-sided position (HRP), 15 vertical lines in top position (VTP), 15 vertical lines in centred position (VCP), and 15 vertical lines in bottom position (VBP). The design was such that, by the end of the test, each quadrant of the visual ®eld had been exposed to the same amount of stimulus material. The capital letters A, B, F, H, O, S, X, and Z appeared in random order at the corners of the pages (Fig. 4). 2.2.2. Procedure Prior to the LB test, subjects were asked to point to or to read aloud the four letters on each page, in order to ensure the perception of the whole page. Then, subjects were instructed to mark the midpoint of each line with a ballpoint pen, using their preferred hand. The page orientation remained perpendicular upon the subjects' coronal plane during both the horizontal and vertical presentation. A ®ve item non-scored trial preceded the test. The distance from the right side of the horizontal lines (or the top of the vertical lines) to the subjects' mark was measured to 0.5 mm accuracy. In order to assess the subjects' directional bias an asymmetry index (AI) was calculated as: 2MD ÿ SL AI 100 SL where MD is measured distance from the right (top) end and SL is stimulus length. Positive AI values indicate an underestimation at the right (of horizontal lines) or on top (of vertical lines), negative AI values indicate an underestimation at the left or at the bottom. Individual mean AI's were calculated for all horizontal and vertical lines (horizontal asymmetry index (HAI) and vertical asymmetry index (VAI)), for the three length categories, as well as for the dierent stimulus positions. In order to avoid biases resulting from averaging positive and negative values, the unsigned AIs (vAIsv) were used to evaluate the subjects' accuracy. Moreover, mastery of the line
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
889
Fig. 2. Mean asymmetry indices and dierences triggered by stimulus-related characteristics: AI, asymmetry index; SD, standard deviation; t, tvalues; _ or ^, 0.001 < P < 0.05; = or ;, P < 0.001. w
890
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
Fig. 3. Mean asymmetry indices and dierences triggered by demographic variables: AI, asymmetry index; SD, standard deviation; vAIv, unsigned asymmetry index; t, t-values of the mastery analysis are bold typed when P < 0.05. Newman±Keuls multiple range tests were used (in the analysis of the directional bias and in the accuracy analysis) to separate the dierent groups into clusters, the mean AIs within a cluster not being signi®cantly dierent [15].
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
bisection skill was assessed by comparing the vAIsv with zero (i.e. the objective midpoint). To this end a null set of data was constructed. 2.2.3. Statistical analysis The impact on AIs of stimulus length, position and orientation was evaluated by means of Pearson correlation and single t-tests. Relationships between AI (or vAIv) and demographic or neuropsychological variables were examined using both parametric (MANOVA and Newman±Keuls test [15]) and non-parametric (Chisquare test) techniques. Additional single t-tests on vAIv were performed to determine whether the subject's marks dier signi®cantly from the mathematical midpoints [45,47].
3. Results Numerical data not included in the text are given in Figs. 2 and 3. 3.1. Stimulus characteristics The mean AIs yielded by dierent stimulus lengths, positions and orientation are shown in Fig. 2. The impact of stimulus characteristics manipulation on accuracy is illustrated by the interbox-connections and their labels. In the same Fig. 2, the within-box t-values result from the mastery analysis. One-line connections and regular underscoring re¯ect a signi®cance level with 0.05 > P > 0.001. Double connections and double underscoring correspond to a signi®cance level with P < 0.001. Fig. 2 clearly shows that by manipulating length, position and orientation signi®cantly dierent AIs are obtained, the only exception being the VTP and VCP which yield comparable AIs. Furthermore, the mastery analysis documents that a signi®cant shift to the left occurred in horizontal bisection, whereas in vertical bisection, an even more signi®cant upward displacement was seen. Displacements of the subjective midpoint became more important as stimulus length increased. In HLP and HCP lines a signi®cant leftward shift of the subjective midpoint occurred. On the other hand, HRP lines yielded a non-signi®cant displacement to the right. In all vertical lines, the subjective midpoint shifted signi®cantly toward the top. This shift was smaller in VBP lines than in VCP or VTP lines. Further analysis showed that there was no signi®cant correlation between HAI and VAI (r(650)=ÿ0.02, ns) and that there was no length position interaction in horizontal (F(4,27)=0.78, ns) nor in vertical bisection (F(4,27)=1.69, ns).
891
3.2. Demographic variables 3.2.1. Directional bias From Fig. 3 it is clear that, when bisecting horizontal lines boys tended to underestimate the left part (HAI=ÿ0.29) and girls the right part (HAI=0.92). The dierence is signi®cant (F(1,648)=16.75, P < 0.001). Therefore, boys and girls are represented as clearly separate groups. This eect was not due to averaging, since there were signi®cantly more boys with a negative HAI and girls with a positive HAI (w 2=4.40, P = 0.03). Handedness did not signi®cantly in¯uence VAIs, but yielded signi®cant dierences in HAIs. Right-handed subjects obtained signi®cantly smaller HAIs than nonright-handed subjects (i.e. the left-handed and ambidexterous subjects, in whom comparable HAIs were obtained). A further 2 3 MANOVA showed a signi®cant handedness stimulus position interaction (Wilks' lambda=0.95, F(3,646), P < 0.001). Nonright-handed subjects showed a signi®cant leftward bias in all three stimulus positions (HAIHLP=2.34, t(99)=ÿ5.98, P < 0.001; HAIHCP=2.92, t(99)=ÿ6.65, P < 0.001; HAIHRP=1.40, t(99)=ÿ2.52, P = 0.01). In right-handed subjects HLP lines yielded a signi®cant leftward bias (HAIHLP=0.77, t(549)=ÿ3.87, P < 0.001), HRP lines triggered a signi®cant rightward shift (HAIHRP=ÿ0.55, t(549)=3.07, P = 0.002), HCP lines yielding no signi®cant displacement (HAIHCP=0.01, t(549)=0.11, ns). The use of the right hand to perform the task resulted in signi®cantly smaller HAIs. In vertical bisection there was no hand eect. A 2 3 MANOVA showed a signi®cant hand stimulus position interaction (Wilks' lambda=0.95, F(3,646), P < 0.001). The pattern of results is identical to the one of the handedness stimulus position interaction. Dierences in age yielded signi®cant eects in HAI (F(4,645)=2.42, P < 0.05) and VAI (F(4,645)=12.19, P < 0.001). In general, younger subjects tended to underestimate more than the older ones. To verify which dierences were signi®cant among the ®ve mean AIs, a Newman±Keuls Multiple Range Test [15] was used. The result of this analysis is displayed in Fig. 3. In horizontal bisection none of the age groups dier signi®cantly with respect to mean AI, whereas in vertical bisection the mean AIs of the 7- and 8-year-olds diers signi®cantly from the mean AIs of the 9-, 10and 11+-year-olds. There was no signi®cant gender handedness hand age interaction in horizontal nor in vertical bisection. From the entire sample the within subject variability was estimated as SD=11.37. 3.2.2. Accuracy When bisecting horizontal lines, boys are signi®-
892
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
3.4. Reading Reading skills had no signi®cant impact on HAIs (R1: F(2,647)=0.59, ns; R2: F(2,647)=0.10, ns) nor on VAIs (R1: F(2,647)=0.52, ns; R2: F(2,647)=1.63, ns). Subjects exposed to bidirectional reading instruction (n = 11) did not present with signi®cantly dierent HAIs (F(1,648)=0.69, ns) nor VAIs (F(1,648)=0.16, ns) than their from left to right reading classmates. 4. Discussion
Fig. 4. Example of the stimulus design.
cantly more accurate than girls, but when they bisect vertical lines, both groups perform equally well (Fig. 3). In terms of accuracy, there is no handedness eect in horizontal nor in vertical bisection (Fig. 3). A signi®cant hand eect is only seen in vertical bisection: when the right hand has been used, the VAIs are smaller, and thus the performance more accurate (Fig. 3). Accuracy clearly increases with age: vHAIsv vVAIsv (F(4,645)=9.63, P < 0.001) and (F(4,645)=12.83, P < 0.001). However, the pace by which signi®cantly dierent levels of accuracy are reached is dierent in horizontal and in vertical bisection (Fig. 3). 3.2.3. Mastery Both boys' and girls' vHAIsv deviated signi®cantly from the objective midpoint (as can be deduced from the bold typed t-values in Fig. 3). And also when bisecting vertical lines both boys and girls signi®cantly underestimated the bottom part. In terms of mastery, no handedness or hand eect was demonstrated. All age groups displaced the marks signi®cantly, although, with increasing age a tendency towards increasing mastery was observed. 3.3. Attention Dierences in attention performance-level were not re¯ected in dierent HAIs (A1: F(2,647)=1.04, ns; A2: F(2,647)=0.49, ns) nor in VAIs (A1: F(2,647)=0.23, ns; A2: F(2,647)=0.04, ns). Furthermore, no signi®cant dierences were found when IQ-levels were compared with HAIs (F(3,646)=1.17, ns) and VAIs (F(3,646)=0.95, ns).
Results obtained with our protocol show that normal children's LB performances are subject to biases. In horizontal lines left- and rightward shifts of the subjective midpoint occur, along the vertical axis an upward shift seems to be the rule. HAIs and VAIs are independent: performances in the horizontal dimension do not allow predictions as to the vertical bisection and vice versa. Moreover, our childrens' VAIs are signi®cantly larger than their HAIs, a ®nding which corroborates previously reported data in normal adults [47]. Thus, horizontal and vertical LBs seem to require separate skills in both young and adult normal subjects. The eects discussed below, seem to consolidate this distinction, especially with respect to children. Indeed, performances become more accurate with age, and LB may perhaps be mastered much later in life, but certainly not before the age of 11. But strikingly, AIs and accuracy measures do not evolve at the same pace in horizontal and in vertical LB. Moreover, in horizontal bisection only girls and non-right-handers present with an important directional bias, whereas a sex or handedness eect is absent in vertical bisection. The importance of individual variations in LB has repeatedly been stressed [2,33]. Therefore, it is important to note that the sex linked pattern found here is not due to an averaging eect: there are indeed more girls with a rightward bias. With respect to the impact of handness and hand performance, one can not but observe that in our study both eects coincide. This may be due to the aselect nature of our population (and thus to the not very impressive number of non-right-handed performers). Moreover, subjects completed the forms only once, using one hand, which makes our data far from ideal in view of the assessment of the hand eect. Nevertheless, some observations can be made. Along the vertical axis, our right-handed, left-handed and ambidexterous subjects obtain comparable AIs and in terms of accuracy the three groups perfom equally well. This is in contrast with Scarisbrick et al.'s [45] normal left-handed adults, who bisected vertical lines far more accurately than their right-handed participat-
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
ing colleagues. Moreover, Scarisbrick et al. [45] failed to demonstrate a hand eect in vertical LB, whereas our subjects who used their right hand are signi®cantly more accurate than those who held the pencil in their left hand. Both these dierences may be linked to the age of our subjects and to the fact that vertical bisection assumes a skill acquired apparently after the age of 11. Moreover, a lot of variability in that skill has been described in adults: Rapcsak et al.'s [41] normal subjects, when bisecting vertical lines, tended to underestimate slightly the top part, whereas Shelton et al.'s [47] normal subjects tended to underestimate the bottom part. In accordance with the authors who predicted (on the basis of Weber's law) and reported signi®cant length eects in normal adults [9,18], the signi®cance of the misbisection increases together with line length in our population. In addition to length, stimulus position is another source of variation. Nichelli et al. [38] found that, in horizontal bisection, normal adults displace the subjective midpoint toward the end opposite to the side of hemispace presentation (i.e. body midline attraction). Our right-handed subjects (as well as all right-hand bisectors) display the opposite phenomenon: stimuli presented in the left hemi-space yielded a signi®cant leftward bias, lines presented to the right resulted in a signi®cant rightward displacement of the marks, whereas only the centred stimuli yielded no signi®cant bias. Our non-right-handed subjects (as well as those who performed the task with their left hand) misbisected signi®cantly to the left in all three conditions. The ``excentric'' shift of the right-handers' marks is compatible with the long standing view that each hemisphere orients preferentially toward the contralateral hemispace [19,45]. Stimuli presented in the left hemi-space cause a more important selective activation of the right hemisphere resulting in an enhancement of the left visual ®eld, and thus a leftward deviation of the midpoint, and vice versa. Our nonright-handed subjects showed a stronger leftward bias than our right-handed children, as was the case in Turnbull and McGeorge's [52] normal adults. This might well be the result of a ``dominant'' right hemisphere extending its leftward bias over its usual scan®eld; i.e. both hemispaces [19]. Subjects with lower scores on substitution tests performed equally well on LB as subjects with higher scores. This may indeed be seen as a remarkable ®nding since in patients with HN, neglect seems to be associated with a general de®cit in the intensity aspects of attention, i.e. with de®cient generation or maintenance of alertness [44]. Moreover, HN phenomena seem to improve with sustained attention training [42]. One might thus have expected decreasing AI with increasing substitution test results. The reason why this trend was not observed might well be that substitution tests
893
are not ``pure'' measures of attention [30] and that they are not suitable measures for the vigilance and spatial attention implicated in LB [19,42]. On the other hand, this ®nding suggests not to overestimate the possibly ``distracting nature'' of our complex protocol in which several lines are presented on one sheet. Another way of exploring the relationship between LB and attentional processes is based upon Chokron and Imbert's [9] ®ndings that LB performance and reading behaviour seem to be related, because reading instruction directs (single or multiple component) spatial attention toward one hemispace [9]. Therefore, we expected to ®nd signi®cantly dierent HAIs in poor and strong readers. However, in our population, the in¯uence of reading on LB was not as prominent as in previous reports [9]. This may be explained by the fact that we compared reading pro®ciency with LB performance in predominantly left to right readers, whereas Chokron and Imbert [9] studied the in¯uence of reading direction upon LB results in left to right readers as well as in right to left readers. Moreover, their 60 French subjects were older (range: 13.9±44.3 years) [9]. However, the relevance of this last argument depends upon whether or not future research is able to discover a signi®cant change in horizontal LB performance in subjects older than 12 years. In summary, horizontal and vertical LB appear to be distinct tasks appealing on dierent skills. This may incite to look for double dissociations in HN patients. It was found that LB is a skill which seems not to be mastered before the age of 11, and as even normal adults display signi®cant biases, certainly along the vertical axis [27], it remains doubtful whether LB ever may attain some degree of ``mathematical precision''. Furthermore, length and position seem to act as intensi®ers of the asymmetry phenomenon. If it is accepted that tasks which involve the appreciation of visual space selectively activate the right hemisphere, the overall right underestimation we found may be explained as resulting from an enhancement of the left visual ®eld and consequently a displacement of the subjective midpoint to the left. As in adult non-righthanders [45,52], this right hemisphere activation eect is also greater in non-dextral children. The fact that boys attain signi®cantly more accuracy in horizontal bisection than girls might possibly support the hypothesis of greater hemispheric specialization for visuospatial tasks in males than in females [22]. Acknowledgements This research was supported by the Provincie Antwerpen, the Belgische Stichting Roeping and the Antwerp University (UIA). We gratefully acknowledge the help of P. Wackenier and P. Bal. We wish to
894
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
thank Y. De Smet and J. Turnbull for reviewing the manuscript and M. Isacson for his kind bibliographical support. We also gratefully acknowledge the helpful suggestions of two anonymous reviewers.
[23] [24] [25]
References [1] Barton J, Black S. Line bisection in hemianopia. Journal of Neurology, Neurosurgery, and Psychiatry 1998;64:660±2. [2] 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. [3] Bowers D, Heilman K. Pseudoneglect: eects of hemispace on a tactile line bisection task. Neuropsychologia 1980;18:491±8. [4] Bradshaw J, Nettleton N, Hathan G, Wilson L. Bisecting rods and lines: eects of horizontal and vertical posture on left-side underestimation by normal subjects. Neuropsychologia 1985;23:421±5. [5] Brus B, Voeten H. EeÂn minuut test. Nijmegen: Berkhout BV, 1980. [6] Bruyer R. Neglects in hemineglect: a comment on the study of Bisiach et al. (1983). Brain and Cognition 1984;3:231±7. [7] Cantagallo A, Ferraresi G, Basaglia N. Visual attentive and premotor factors of visual neglect in the line bisection test. Journal of Clinical and Experimental Neuropsychology 1993;15:413. [8] Cattell RB. Measuring intelligence with the culture fair tests. Illinois: Institute for Personality and Ability Testing, 1973. [9] Chokron S, Imbert M. In¯uence of reading habits on line bisection. Cognitive Brain Research 1993;1:219±22. [10] Dellatolas G, Coutin T, DeAgostini M. Bisection and perception of horizontal lines in normal children. Cortex 1996;32:705± 15. [11] De Vos L, Lievens S, Vallaeys A. Schoolvorderingentest BVL, voor de leerjaren 2, 3, 4, 5, en 6 van het gewoon lager onderwijs en voor het buitengewoon onderwijs. Tessenderlo: Boekenprokuur Broeders van Liefde, 1986. [12] Ferro JM, Kertesz A, Black SE. Subcortical neglect: quantitation, anatomy, and recovery. Neurology 1987;37:1487±92. [13] Ferro JM, Martins IP, Tavora L. Neglect in children. Annals of Neurology 1984;15:281±4. [14] Freedman M, Leach L, Kaplan E, Winocur G, Shulman K, Delis D. Clock drawing: A neuropsychological analysis. MultiHealth Systems Inc, 1994. [15] Godfrey K. Statistics in practice: comparing the means of several groups. New England Journal of Medicine 1985;313:1450± 6. [16] Graveleau Ph, Viader F, Masson M, Cambier J. NeÂgligence thalamique. Revue Neurologique (Paris) 1986;142:425±30. [17] Halligan P, Manning L, Marshall J. Individual variation in line bisection: a study of four patients with right hemisphere damage and normal controls. Neuropsychologia 1990;28:1043±51. [18] Halligan P, Marshall J. Line bisection in visuo-spatial neglect: disproof of a conjecture. Cortex 1989;25:517±22. [19] Halligan P, Marshall J. Toward a principled explanation of unilateral neglect. Cognitive Neuropsychology 1994;11:167±206. [20] Heilman K, Bowers D, Watson R. Pseudo-neglect in a patient with partial callosal disconnection. Brain 1984;107:519±32. [21] Heilman K, Watson R, Valenstein E. Neglect and related disorders. In: Heilman K, Valenstein E, editors. Clinical neuropsychology. New York: Oxford University Press, 1985. p. 243±93. [22] Hiscock M, Israelian M, Inch R, Jacek C, Hiscock-Kalil C. Is there a sex dierence in human laterality? II: an exhaustive survey of visual laterality studies from six neuropsychology jour-
[26]
[27] [28] [29] [30] [31]
[32]
[33] [34] [35] [36] [37] [38] [39] [40]
[41] [42]
[43] [44]
nals. Journal of Clinical and Experimental Neuropsychology 1995;17:590±610. Johnston C, Shapiro E. Hemi-inattention resulting from left hemisphere brain damage during infancy. Cortex 1986;22:279± 87. Jeerakathil T, Kirk A. A representational vertical bias. Neurology 1994;44:703±6. Kaplan RF, Verfaellie M, Meadows M-E, Caplan LR, Pessin MS, DeWitt LD. Changing attentional demands in left hemispatial neglect. Archives of Neurology 1991;48:1263±6. Kashiwagi T, Kashiwagi A, Nishikawa T, Tanabe H, Okuda JI. Hemispatial neglect in a patient with callosal infarction. Journal of Clinical and Experimental Neuropsychology 1989;11:377. Kashmere J, Kirk A. The complex interaction of normal biases in line bisection. Neurology 1997;49:887±9. Katz N, Cermak S, Shamir Y. Unilateral neglect in children with hemiplegic cerebral palsy. Perceptual and Motor Skills 1998;86:539±50. Lecours AR, Lupien S, Parente MA. Visual attention in left sylvian strokes. Journal of Neurolinguistics 1989;4:255±71. Lezak M. Neuropsychological assessment. New York: Oxford University Press, 1995. Maeshima S, Funahashi K, Ogura M, Itakura T, Komai N. Unilateral spatial neglect due to right frontal haematoma. Journal of Neurology, Neurosurgery, and Psychiatry 1994;57:89±93. Maeshima S, Terada T, Nakai K, Nishibayashi H, Ozaki F, Itakura T, Komai N. Unilateral spatial neglect due to haemorrhagic contusion in the right frontal lobe. Journal of Neurology 1995;242:613±7. Manning L, Halligan P, Marshall J. Individual variation in line bisection: a study of normal subjects with application to the interpretation of visual neglect. Neuropsychologia 1990;28:647±55. Massironi M, Antonucci G, Pizzamiglio L, Vitale MV, Zoccolotti P. The Wundt±Jastrow illusion in the study of spatial hemi-inattention. Neuropsychologia 1988;26:161±6. McCourt M, Olafson C. Cognitive and perceptual in¯uences on visual line bisection: psychophysical and chronometric analyses of pseudoneglect. Neuropsychologia 1997;35:369±80. Mennemeier M, Wertman E, Heiman K. Neglect of near peripersonal space: evidence for multidirectional attentional systems in humans. Brain 1992;115:37±50. Monaghan P, Shillcock R. The cross-over eect in unilateral neglect, modelling detailed data in the line-bisection task. Brain 1998;121:907±21. 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. Paquier PF, Van Dongen HR. Review of research on the clinical presentation of acquired childhood aphasia. Acta Neurologica Scandinavica 1996;93:428±36. Paquier P, Van Mourik M, Van Dongen H, CatsmanBerrevoets C, Creten W, Stronks D. Clinical utility of the judgment of line orientation test and facial recognition test in children with acquired cerebral lesions. Journal of Child Neurology 1999;14:243±8. Rapcsak SZ, Cimino CR, Heilman KM. Altitudinal neglect. Neurology 1988;38:277±81. Robertson IH, TegneÂr R, Tham K, Lo A, Nimmo-Smith I. Sustained attention training for unilateral neglect: theoretical and rehabilitation implications. Journal of Clinical and Experimental Neuropsychology 1995;17:416±30. Rourke B, Fisk J, Strang J. Neuropsychological assessment of children. New York: Guilford Press, 1986. Samuelsson H, Hjelmquist E, Jensen C, Ekholm S, Blomstrand C. Nonlateralized attentional de®cits: an important component
P. van Vugt et al. / Neuropsychologia 38 (2000) 886±895
[45] [46] [47] [48] [49] [50] [51] [52]
behind persisting visuospatial neglect? Journal of Clinical and Experimental Neuropsychology 1998;20:73±88. Scarisbrick D, Tweedy J, Kulanski G. Hand preference and performance eects on line bisection. Neuropsychologia 1987;25:695±9. Schenkenberg T, Bradford D, Ajax E. Line bisection and unilateral visual neglect in patients with neurological impairment. Neuropsychologia 1987;25:695±9. Shelton PA, Bowers D, Heilman KM. Peripersonal and vertical neglect. Brain 1990;113:191±205. Smith A. Symbol digit modalities test. Los Angeles: Western Psychological Services, 1973. Stone S, Halligan P, Marshall J, Greenwood R. Unilateral neglect: a common but heterogeneous syndrome. Neurology 1998;50:1902±5. Tanghe M. Schoolvorderingentest BVL. Tessenderlo: Boekenprokuur Broeders van Liefde, 1971. Thompson N, Ewing-Cobbs L, Fletcher J, Miner M, Levin H. Left unilateral neglect in a preschool child. Developmental Medicine and Child Neurology 1991;33:636±44. Turnbull O, McGeorge P. Lateral bumping: a normal-subject
[53] [54] [55] [56] [57] [58]
[59]
895
analog to the behaviour of patients with hemispatial neglect? Brain and Cognition 1998;37:31±3. UmiltaÁ C. Domain-speci®c forms of neglect. Journal of Clinical and Experimental Neuropsychology 1995;17:209±19. Van den Bos K, lutje Spelberg H, Scheepstra A, De Vries J. De klepel, vorm A en B: een test voor de leesvaardigheid van pseudowoorden. Nijmegen: Berkhout, 1994. Van Dongen H. Clinical aspects of acquired aphasia and dysarthria in childhood. Delft: Eburon, 1988. Van Haasen PP, et al. De WISC-R, Nederlandse uitgave. Lisse: Swets and Zeitlinger, 1986. Vuilleumier P, Valenza N, Mayer E, Reverdin A, Landis T. Near and far visual space in unilateral neglect. Annals of Neurology 1998;43:406±10. Weintraub S, Daner K, Ahern G, Price B, Mesulam M-M. Right sided hemispatial neglect and bilateral cerebral lesions. Journal of Neurology, Neurosurgery, and Psychiatry 1996;60:342±4. Werth R, PoÈppel E. Compression and lateral shift of mental coordinate systems in a line bisection task. Neuropsychologia 1988;26:741±5.