n;europswholoqro. Vol. 28. No 2. pp. 217-222. Pnnted m Great EMam
002X-3932 90 53.00+0.00 ,. 1990 Pcrgamon Prera plc
1990.
NOTE DIGIT SPAN AND VISUAL NEGLECT:
Department
of Psychology,
A PUZZLING
IAN H. ROBERTSON Astley Ainslie Hospital, Grange Scotland. U.K.
Loan,
RELATIONSHIP
Edinburgh
EH9 2HL.
(Received 2 May 1989; accepted 18 September 1989) Abstract-Weinberg et al. found a relationship between the presence of extinction and the discrepancy between backward and forward digit span among left hemiplegic patients. The present study replicates this finding in a group of 37 patients suffering from unilateral left visual neglect, demonstrating a strong relationship between degree ofneglect and digit span discrepancy. A principal components analysis of a neurospychological examination revealed three main independent components-Neglect, Verbal Memory and Non-verbal Memory respectively. Digit Span Discrepancy had one of the highest loadings on the neglect factor, and loaded significantly on no other factor. Multiple regression analysis showed that the relationship of Digit Span Discrepancy with Degree of Neglect exists independently of visuo-spatial ability, and this is also true for paced auditory serial addition (PASAT) scores for one of the two main indices of neglect. The implications of this for understanding neglect are discussed.
INTRODUCTION WEINBERGet al. [22] showed
that among 45 left hemiplegic subjects, those showing two or more omissions on confrontation testing-and therefore showing more extinction-showed a significantly greater discrepancy between forward and backward digit span than did left hemiplegic patients who made one or no errors on confrontation testing. They interpreted this result as being due to impaired visual scanning and eye movements resulting in defective visuo-spatial representation of the string of digits which could not therefore be “read ofI”’ backwards so readily. The Forward Digit Span test is less demanding ofvisuo-spatial representation, and hence it is less affected by visuo-spatial problems, resulting in a bigger discrepancy between backward and forward scores in a sub-group of subjects showing more visual neglect. so the authors of this study argued. If the above interpretation of the results is correct, then controlling for impaired visuo-spatial representation ability should abolish the correlation between extinction and digit span discrepancy. If there remains a relationship between neglect and the digit span variable after controlling for visuo-spatial representation abilitv, however. then some alternative explanation should be found which would have implications for an understanding of unilateral neglect
METHOD Subjects
Subjects were all patients in Edinburgh hospitals who showed significant unlilateral left visual field neglect according to the Behavioural Inattention test (BIT) [25]. The presence of neglect was operationalized as failure on at least three out of nine behavioural tests, as such a cutoffdistinguished neglect patients from controls in a study by WILSON et al. [24]. Patients were also aged 80 or under, oriented for time and place, and not suffering from progressive or degenerative neurological conditions. They also all had to give informed consent for participating m the research. The subjects included 19 men and 1Xwomen. with a mean age of62.5 (SD 12.5). Thirty-four had suffered cerebrovascular accidents, two head injuries. and one a meningioma. The subjects were a mean of 15.5 weeks post-infarct (or injury/operation) (SD 16.5). Of these subjects 23 had left homonymous hemianopias. and of these, 10 showed no macular sparing. Their mean Digit Forward score was 5.9 (SD 0.97). Backward 3.7 (SD 1.21). and Line Orientation 217
‘18
Non
I I .O (SD 5. I ) (the latter out ofa maximum of 30). The mean Logical Memory score was 7.6 (SD 2.7). Of the subjects 25 showed clear neglect of the left side of a complex figure on copying. Further informatton is given in Table 1.
All subjects were examined by a consultant neurologist as part of a wider study of the efficacy of computertzed rehabilitatton of neglect. On the basts of this examination, a score based on Bond’s neurophystcal scale [6] was computed. which quantified the degree of neurophysical impairment in terms of cranial nerve involvement. sensory and motor loss and a number of other neurological variables included in Bond’s scale. Visual field loss was assessed usmg the method described by DILLERet al. [7]. where the examiner‘s two hands are held at three posittons of increasing distance from the body mid-lines ofpatient and tester who sit facing each other. Fingers are moved in each hand, in both hands. and not at all in a random order. and the omissions recorded. In addition, a Bjerrum screen was used to assess the inner 30 degrees of the visual fields, and in particular to assess whether macular vtsion was impaired. A 5 mm white disc attached to a black wand was used for this examination. Wtthin the limitations of considerable difficulties in achieving eye fixation in this population, all the hemianopias detected were complete and homonymous. and no quandrantanopias were found. The mam dependent variables were two measures of neglect (the behavioural section of the Behavioural Inattention test-BIT and Letter Cancellation (c’s and e’s) [7], The Behavioural Inattention test (behavioural items) measures neglect on a number of standardised situations ranging from reading to picking up coins. These behavioural tests are highly correlated with more conventional measures of neglect such as letter cancellation [24, 251 and thus can be used with confidence in the present study. (The correlation in the present study between BIT scores and the Letter Cancellation Neglect Index used was -0.86). An index of neglect [ 173 derived from the Letter Cancellation test was used as the other main dependent variable. This index was the percent of responses expected on the left of the Cancellation test (i.e. 50%) minus the percent actually observed. Other central variables were the difference between Forward and Backward Digit Span (Digit Span subtest of the Wechsler Adult Intelligence Scale (Revised Version) [21] and visuo-spatial orientation (Benton Line Orientation test, [4]). The latter was chosen as a relatively “pure” measure of simple spatial representation ability (see [3] for justification of this assumption), though other tests of visuo-spatial functioning were given as mentioned below. The other tests given were as follows: the Logical Memory sub-test of the Wechsler Memory Scale [ZO]: the Rey-Osterreith Complex Figure test, copy and recall 116, 181; the National Adult Reading test [IS]; the Word Fluency test [2]; the Wisconsin Card Sorting test (Nelson version-14); the Paced Auditory Serial Addition test4 set interval (PASAT) [S]: the Neale Reading test [13]; the Picture Completion sub-test of the Wechsler Adult Intelligence Scale (Revised Version) [21]: the Block Design sub-test of the Wechsler Adult Intelligence Scale (Revised Version); Motor Impersistence test [5]. Procedure
Tests were given in a fixed order over two 1 hr sessions. The procedure for administering the Line Orientation test was amended slightly for the purposes of the present study. so as to minimize the influence of visual neglect on the test performance. If the patient only pointed to one line on the target, he or she was reminded that there were two. and that he or she should look to the left of the target and matching pages. In this way omissions interfered less with assessment of ability to represent a given angle of orientation. Data analysis
The data were then analysed in two ways. Firstly. a principal components analysis was carried out. with the hypothesis that tests ofneglect would load on a factor with Digit Discrepancy and PASAT, but not on other factors. No varimax or other type of rotation was carried out. and components with an eigenvalue of greater than one were selected. Secondly. multiple regression analysis was implemented in order to establish to what degree the relationship between neglect and serial processing capacity was independent of visuo-spatial capacity.
RESULTS Table 2 shows the loadings of variables on the three principal components which had at least one variable loading 7 O/bof the total variance in test scores, which constitutes 58% of the variance above 0.7. These together explain I_.8 explained by all the components in the analysis. Table I only shows the loadings for each of the three factors. The first component could be labelled a “neglect” component. with the added “serial processing” variables of Digit Discrepancy and PASAT (longest string). The second component loads at the criterion level on only three variables, namely Logical Memory, Macular Splitting and Digit Backward: it does not load on any neglect-related variables. The third component loads only on Rey-Osterreith recall, suggesting a non-verbal memory component.
219
NOTE Table 1. Subject characteristics
Age 67 37 63 61 69 64 69 57 60 72 60 39 71 27 79 52 56 68 71 70 57 65 72 77 69 62 56 61 76 65 70 72 71 79 65 64 29
Weeks postonset
Diagnosis (1)
Location (2)
8 7 8 11 92 12 8 17 22 26 49 7 7 22 12 9 5 8 29 24 13 12 5 10 5 42 8 4 27 10 5 7 7 13 9 6 14
cva cva cva cva cva cva cva men. cva cva cva chi cva cva cva cva cva cva cva cva cva chi cva cva cva cva cva cva cva cva cva cva cva cva cva cva cva
na na tps !a na f
f:apS na tP
fip fu na na na P S
na P ; na na na ftp ftD
itp na na na na na na na na na na
Hemianopia (3) No No No No No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes No Yes Yes No No Yes No No No Yes Yes Yes Yes Yes No Yes
Macula? (4)
Neurophysical scale (5)
No
10
No No No No Yes No Yes No Yes No Yes Yes Yes No Yes No
9 9 8 5 14 4 13 13 12 12 14 11 14 4 13 13
FO
No Yes No No No No No na No No No na
13 12 12 13 3 12 7 8 9 15 9 10 8 14
Ks No Yes No No
14 15 7 na 8 12
WAIS Block Design (6)
WAIS PC (7)
Bit (8)
3 6 7 5 5 3 7 6 7 3 na 2 3 na 6 2 4 3 5 6 2 6 3 3 2 8 7 6 3 8 3 3 4 6 4 2 1
2 8 6 4 9 4 6 8 5 7 na 7 5 na 12 5 11 5 8 7 6 7 2 7 3 7 3 8 5 9 6 2 7 5 4 4 5
40 72 68 68 44 13 60 57 68 25 36 39 3 36 61 28 18 12 71 69 50 72 53 69 30 74 37 67 64 52 12 24 50 71 4 71 38
Digit diff Rey score Copy (9) (10) 3 1 1 2 2 2 2
1 2 1 3 3 3 3 1 4 5 3 0 2 3 2 2 1 3 1 0 2 2 3 2 3 1 1 5 2 2
22 41 23 34 14 8 17 43 31 6 na 5 na na 33 16 11 16 26 30 10 35 10 12 7 31 18 37 32 28 8 22 11 28 7 21 19
Notes: (1) cva: cerebra-vascular accident; chi: closed head injury; men: excised meningioma. (2) ct scan based. I: frontal, t: temporal, p: parietal, s: subcortical, na: not available. (3) See text for method. (4) Yes indicates macular splitting, based on Bjerrum screeen using 5 mm white disc on wand. (5) Neurophysical scale by Bond. (6) WAIS-R Block Design age scaled score. (7) WAIS-R Picture Completion age-scaled score. (8) Behavioural subtests of Behavioural Inattention test, WILSON et al. [25]; maximum score =81. (9) WAIS-R digit span: difference between maximum number of digits forward and maximum number backward. (10) Rey-Osterreith complex figure copy: number of items correctly copied out of a maxtmum of 47. Multiple regression on BIT SLOWS An automated stepwise regression was then implemented, including as the independent variables three measures of visuospatial functioning (Line Orientation, Block Design, and Rey-Osterreith copy scores) as well as Digit Span Discrepancy (an G(level of 0.1 for both entering and removing variables was chosen). The dependent variable of interest was the BIT score. This regression yielded only two significant predictive variables-the first being Digit Discrepancy scores (explaining 46.1% of variance in BIT scores) and the second Rey-Osterreith copy scores (explaining a further 15.5% of the variance). to give a total explained variance of 61.6%. This result was checked by carrying out a further stepwise multtple regression, loading in the visuospatial variables first. Together, Line Orientation, Block Design and Rey-Osterreith scores predicted 45.6 of variance in BIT scores. Adding in Digit
220
NOTE Table 2. Loadings
on three principal
components
of neuropsychological
Test
Component
loadings
I
0 RX -0.81 0.77 -0.73 -0.72 0.71 0.69 0.52 0.50 -0.16 -0.22 0.45 -0.30 0.03 -0.35 0.41
Behavioural Inattention test Forward Minus Backward Digit Span ReyyOsterreith copy Neale Reading test-accuracy Letter Cancellation accuracy PASAT 4 set Longest String Block Design PASAT 4 set total score Line Orientation Logical Memory Macular Splitting Digit Backward Digit Forvvard Word Fluency Hemianopia Rey-Osterreith Recall Variance
Percent
explamed by components. 1 2 5.6640;0 4.1oo”io of total variance explained
I
2
25.145
18.638
variables
0.05 -0.1'
0.14 0.39 0.19 0.34 0.07 0.44 0.48 0.81 0.75 0.71 0.69 0.58 0.56 -0.32
3 -0.31 0.4 1 0.07 0.17 0.40 0.18 0.28 -0.12 0.05 0.34 0.14 -0.14 0.25 -0.30 -0.30 0.81
3 3.068% 3 13.946
discrepancy to the analysis increased that by a further 18.5 to 64.1%. Gtven that Block Design is a heavily problemsolving. and hence attentional. task. the analysis was repeated excluding Block Design. Digit Discrepancy increased the explained variance of Lure Orientation and Rey copy by 20.1%. from 41.9 to 62.0%. An identical set of analyses was carried out substituting PASAT total score for Digit Discrepancy. Again the automated stepwise analysis revealed only two significant variables-Rey -0sterreith copy and PASAT tomI score. The variance in neglect explained by Line Orientation and Rey copy scores (41.9%) was increased by 10.9”,0 to 52.8% by the addition of the PASAT variable. All the above percent increases in vartance are statistically significant at least at the 0.05 level of significance. and no predictor variables in any regression were significantly correlated. Digit Discrepancy and PASAT were not mcluded in the same regresston analysis because their high intercorrelation violated one of the assumpttons of mulitiple regression.
In order to check that the results are not a feature ofone particular test of neglect. an index of neglect derived from a Letter Cancellation test 1171 was substituted for BIT. Digit Discrepancy scores increased the variance explained by Lure Orientation, Block Destgn and Rey Ostereith copy by 13.5% to 32.8 to 46.3%. This effect was statistically significant at the 0.1 level. There was a trend. but no statistically significant effect, for the inclusion of PASAT total scores.
DISCUSSION The prmciple components analysts conlirms the hypotheses of a loading of Digit Discrepancy and PASAT on the neglect prmctpal component ofthe analysts. Thts was independent ofthe second factor which loaded significantly on Logical Memory and Word Fluency. The reason why Digtt Span Discrepancy should be the variable ofchoice rather than either Forward or Backward Digit Span alone 1s quote sample. Forward Digit Span is not commonly reduced following brain damage m the absence oflanguage tmpatrment and Backward Dtgit Span is much more commonly affected 141. A Backward Digit Span of 3 tn one person who had a forward span of8, and the same in another with a forward span of 5. might mean two very diflerent thmga. The first would seem to represent a greater degree ol specific impairment on this rev*ersal task than the latter. The former IS more likely to suggest a greater degree of general impairment or of low general intelligence (with the exceptton of rare cases of specificshort-term verbal memory deficits).
Thus it is only at first sight that the loading of Digit Backward on the second component appears to contradict the hypothesis. It is unlikely to be the case given that Forward Digit Span also loads on the same component to almost exactly the same degree: these variables seem to be measuring a general intellectual deterioration rather than a speciiic deficit with reverse digits. This is supported by the fact that Loelcal Memory is the highest loading on this component, as memory is a key measure of general intellectual decline [8]. With respect to the multiple regression analysis. it is clear from the above that the tirst hypothesis is confirmed. insofar as Digit Forward minus Backward explains a significant portion of the variance m neglect. independent11 of the variance explained by tests of visuo-spatial capacity. and this was true for two different indices of neglect. BIT and Letter Cancellation. The finding was also replicated for PASAT total scores. though only for the BIT neglect scores. If then the Digit Discrepancy score still correlates with degree of neglect irrespective of lisuo-spatial representation ability. what might account for this relationship? One possible factor is that a deticlt m general attentional capacity is associated with severity of unilateral visual neglect as suggested by ROBEKTSOS [IS]. and supported by several other recent studies [ll. 12. 17. 23.1. The fact that this study found a separate test which purports to measure the same type ofcentral executive function (PASAT) to show similar results further strengthen this hypothesis. though other explanations are possible. Of course this study does not rule out the spatial processmg hypothesis of Weinberg er al. in explainmg the relationship, but it does strengthen the view that such a hypothesis does not account for all of the data. Acknowledgements-I would like to thank the following colleagues at Astley Ainslie Hospital support in carrying out this study: Brian Pentland, Mary Ramsay. John Gray and Lorna Waite. thank an anonymous reviewer for valuable comments on an earlier draft of this paper.
for their help and
I would also like to
REFERENCES 1. BADDELEY, A. D. U’orking Memory. Oxford University Press, Oxford. 1986. 2. BENTON, A. and HAIISHER. K. Mulrilingual Aphasia Euaminafion. University of Iowa Press. Iowa. 1976. 3. BESTON. A.. HANNAY. J. and VARNAY. N. Visual perception of line direction in patients with unilateral brain damage. NeurnloyJ 25, 907-910. 1975. 4. BEXTON, A.. HAMSHER.K.. VARNEY. N. and SPREES. 0. Conrrihutiotls IO Neurops~c/tological.4s.se.ww~1r. Oxford University Press. Oxford. 1983. 5. BEN-YISHAY. Y., DILLER. L.. GERSTMAN. L. and HASS. A. The relationship between impersistence. mtellectual function and outcome of rehabilitation in patients with left hemiplegia. Neuroloy!, 18, 852-861. 1968. 6. BOND. M. Assessment of psychosocial outcome after severe head injury. In Ourcome c$Serere Donlcu/t’ 10 r/u Central Nervous Sysrem. Ciba Foundation Symposmm 34 (New Series). Elsevier-Exerpta Medica. North Holland. 1975. 7. DILLER. L.. BEIU‘-YISHAY.Y.. GERSTMAN. L.. GOODKIN. R.. GORIXX. W. and WEINBERG. J. Studies irt C~rg~irr~r~ and Rehahilitarim iu Hemipleyrc Rehabilitation Monograph No. 50. New York University Medical C‘cntcr, New York. 1974 8. GIBSON. A. and KE\;uRIT~;. D. T/I@ Hendrick Battery for fhc Drrecrim of Dememia in thrr Ndwly. NFER. Windsor, 1979. 9. GRONWALL. D. and SAntpsov. H. T/I@Ps~clto/o~~~col Eflecfs of‘Concu.ssion. Auckland University Press Oxford University Press. Auckland. 1974. 10. HEILMAS. K. M. and VALFL~STEIX.E. Mechamsms underlying hemispatial neglect. Am. Nrurol. 5. I66 170. 1979. 11. LECCXRS. A. R.. MEHLER. F.. PARENTE. M A. rf ul. Ilhteracy and brain damage: II. Manifestations ofunilateral neglect in testing, “auditory comprehension” with iconographic materials. Brain C~cgt~ir. 6, 243 265. 19X7. 12. MARK. V. W.. KCK)ISTRA.C. A. and HEILMAN. K. M. Hemispatlal neglect affected by non-neglected stimuli. h’euroloyy 38, 1207 121 I. 19X8. 13. NEALE. M. ,Yua/e .Jucr/~s~s c!/ Rradimg Ahilir!. MacMillan. London. 1958. 14. NELSON, H. A moditied card sortmg test sensitive to frontal lobe deficits. Corteu 12. 313 324. 1976. 15. NELSON. H. Tllr .h’e/sor~ .4dulr Rcttdrmg Trst (NART) Tcs/ Manual. NFER, Windsor. 1982. 16. OSTERREITH. P. Le lest de coplc d‘une figure complexe. Archs de Psychologie .X0, 106 156. 1944. 17. RAPCSACK. S. Z.. VEHFAELLIE.M.. FLEET. S. and HCILMAN. K. M. Selective attention in hemispatial neglect. Arch.5 Nrurol. 46, 172 17X. 19X9. 18. REY. A. L‘examen psychologlque dans le cas d’encephaiopathie traumatique. Archs de f’.s~c/t~~/~, in the laterallty of omissions in left unilateral visual field neglect: ImpllcatlonT fob an attentional theory of’ neglect. ~europ.\~~llolt,ylrl 27, 157 165. 1989. 20. WECHSLER. D. A standardlsed memory scale for clinical use. J. P.sycho/. 19, 87 95. 1945. 21. WECHSLEK.D. Tltts Wech&r Adulf I~t~rll~qotc~~ Scnlr-Rcrwrd. The Psychological Corporation. London. 19x1.
222
NOTE
22. WE~XBERG.J.. DILLER. L.. GERSTMAN. L. and STHULMAN. P. Digit span in right and left herniplegics. J. c/in. Psychol. 28, 361, 1972. 23. WEINTRAUB, S. and MESULAM. M. Right cerebral dominance in spatial attention. Further evidence based on ipsilateral neglect. Archs Neural. 44, 621-625. 1987. 24. WILSON, B., COCKBURN. J. and HALLIGAN, P. Development ofa behavioural test of visuospatial neglect. Archs Phys. Med. Rehabil. 68, 98-102, 1987. 25. WILSON, B., COCKBURN. J. and HALLIGAN, P. Behakxu-a/ Inamnrion Test. Thames Valley Test Company. Fareham. Hampshire. 1988.
002X-3932 90 53.00+0.00 ,. 1990 Pcrgamon Prera plc
1990.
NOTE DIGIT SPAN AND VISUAL NEGLECT:
Department
of Psychology,
A PUZZLING
IAN H. ROBERTSON Astley Ainslie Hospital, Grange Scotland. U.K.
Loan,
RELATIONSHIP
Edinburgh
EH9 2HL.
(Received 2 May 1989; accepted 18 September 1989) Abstract-Weinberg et al. found a relationship between the presence of extinction and the discrepancy between backward and forward digit span among left hemiplegic patients. The present study replicates this finding in a group of 37 patients suffering from unilateral left visual neglect, demonstrating a strong relationship between degree ofneglect and digit span discrepancy. A principal components analysis of a neurospychological examination revealed three main independent components-Neglect, Verbal Memory and Non-verbal Memory respectively. Digit Span Discrepancy had one of the highest loadings on the neglect factor, and loaded significantly on no other factor. Multiple regression analysis showed that the relationship of Digit Span Discrepancy with Degree of Neglect exists independently of visuo-spatial ability, and this is also true for paced auditory serial addition (PASAT) scores for one of the two main indices of neglect. The implications of this for understanding neglect are discussed.
INTRODUCTION WEINBERGet al. [22] showed
that among 45 left hemiplegic subjects, those showing two or more omissions on confrontation testing-and therefore showing more extinction-showed a significantly greater discrepancy between forward and backward digit span than did left hemiplegic patients who made one or no errors on confrontation testing. They interpreted this result as being due to impaired visual scanning and eye movements resulting in defective visuo-spatial representation of the string of digits which could not therefore be “read ofI”’ backwards so readily. The Forward Digit Span test is less demanding ofvisuo-spatial representation, and hence it is less affected by visuo-spatial problems, resulting in a bigger discrepancy between backward and forward scores in a sub-group of subjects showing more visual neglect. so the authors of this study argued. If the above interpretation of the results is correct, then controlling for impaired visuo-spatial representation ability should abolish the correlation between extinction and digit span discrepancy. If there remains a relationship between neglect and the digit span variable after controlling for visuo-spatial representation abilitv, however. then some alternative explanation should be found which would have implications for an understanding of unilateral neglect
METHOD Subjects
Subjects were all patients in Edinburgh hospitals who showed significant unlilateral left visual field neglect according to the Behavioural Inattention test (BIT) [25]. The presence of neglect was operationalized as failure on at least three out of nine behavioural tests, as such a cutoffdistinguished neglect patients from controls in a study by WILSON et al. [24]. Patients were also aged 80 or under, oriented for time and place, and not suffering from progressive or degenerative neurological conditions. They also all had to give informed consent for participating m the research. The subjects included 19 men and 1Xwomen. with a mean age of62.5 (SD 12.5). Thirty-four had suffered cerebrovascular accidents, two head injuries. and one a meningioma. The subjects were a mean of 15.5 weeks post-infarct (or injury/operation) (SD 16.5). Of these subjects 23 had left homonymous hemianopias. and of these, 10 showed no macular sparing. Their mean Digit Forward score was 5.9 (SD 0.97). Backward 3.7 (SD 1.21). and Line Orientation 217
‘18
Non
I I .O (SD 5. I ) (the latter out ofa maximum of 30). The mean Logical Memory score was 7.6 (SD 2.7). Of the subjects 25 showed clear neglect of the left side of a complex figure on copying. Further informatton is given in Table 1.
All subjects were examined by a consultant neurologist as part of a wider study of the efficacy of computertzed rehabilitatton of neglect. On the basts of this examination, a score based on Bond’s neurophystcal scale [6] was computed. which quantified the degree of neurophysical impairment in terms of cranial nerve involvement. sensory and motor loss and a number of other neurological variables included in Bond’s scale. Visual field loss was assessed usmg the method described by DILLERet al. [7]. where the examiner‘s two hands are held at three posittons of increasing distance from the body mid-lines ofpatient and tester who sit facing each other. Fingers are moved in each hand, in both hands. and not at all in a random order. and the omissions recorded. In addition, a Bjerrum screen was used to assess the inner 30 degrees of the visual fields, and in particular to assess whether macular vtsion was impaired. A 5 mm white disc attached to a black wand was used for this examination. Wtthin the limitations of considerable difficulties in achieving eye fixation in this population, all the hemianopias detected were complete and homonymous. and no quandrantanopias were found. The mam dependent variables were two measures of neglect (the behavioural section of the Behavioural Inattention test-BIT and Letter Cancellation (c’s and e’s) [7], The Behavioural Inattention test (behavioural items) measures neglect on a number of standardised situations ranging from reading to picking up coins. These behavioural tests are highly correlated with more conventional measures of neglect such as letter cancellation [24, 251 and thus can be used with confidence in the present study. (The correlation in the present study between BIT scores and the Letter Cancellation Neglect Index used was -0.86). An index of neglect [ 173 derived from the Letter Cancellation test was used as the other main dependent variable. This index was the percent of responses expected on the left of the Cancellation test (i.e. 50%) minus the percent actually observed. Other central variables were the difference between Forward and Backward Digit Span (Digit Span subtest of the Wechsler Adult Intelligence Scale (Revised Version) [21] and visuo-spatial orientation (Benton Line Orientation test, [4]). The latter was chosen as a relatively “pure” measure of simple spatial representation ability (see [3] for justification of this assumption), though other tests of visuo-spatial functioning were given as mentioned below. The other tests given were as follows: the Logical Memory sub-test of the Wechsler Memory Scale [ZO]: the Rey-Osterreith Complex Figure test, copy and recall 116, 181; the National Adult Reading test [IS]; the Word Fluency test [2]; the Wisconsin Card Sorting test (Nelson version-14); the Paced Auditory Serial Addition test4 set interval (PASAT) [S]: the Neale Reading test [13]; the Picture Completion sub-test of the Wechsler Adult Intelligence Scale (Revised Version) [21]: the Block Design sub-test of the Wechsler Adult Intelligence Scale (Revised Version); Motor Impersistence test [5]. Procedure
Tests were given in a fixed order over two 1 hr sessions. The procedure for administering the Line Orientation test was amended slightly for the purposes of the present study. so as to minimize the influence of visual neglect on the test performance. If the patient only pointed to one line on the target, he or she was reminded that there were two. and that he or she should look to the left of the target and matching pages. In this way omissions interfered less with assessment of ability to represent a given angle of orientation. Data analysis
The data were then analysed in two ways. Firstly. a principal components analysis was carried out. with the hypothesis that tests ofneglect would load on a factor with Digit Discrepancy and PASAT, but not on other factors. No varimax or other type of rotation was carried out. and components with an eigenvalue of greater than one were selected. Secondly. multiple regression analysis was implemented in order to establish to what degree the relationship between neglect and serial processing capacity was independent of visuo-spatial capacity.
RESULTS Table 2 shows the loadings of variables on the three principal components which had at least one variable loading 7 O/bof the total variance in test scores, which constitutes 58% of the variance above 0.7. These together explain I_.8 explained by all the components in the analysis. Table I only shows the loadings for each of the three factors. The first component could be labelled a “neglect” component. with the added “serial processing” variables of Digit Discrepancy and PASAT (longest string). The second component loads at the criterion level on only three variables, namely Logical Memory, Macular Splitting and Digit Backward: it does not load on any neglect-related variables. The third component loads only on Rey-Osterreith recall, suggesting a non-verbal memory component.
219
NOTE Table 1. Subject characteristics
Age 67 37 63 61 69 64 69 57 60 72 60 39 71 27 79 52 56 68 71 70 57 65 72 77 69 62 56 61 76 65 70 72 71 79 65 64 29
Weeks postonset
Diagnosis (1)
Location (2)
8 7 8 11 92 12 8 17 22 26 49 7 7 22 12 9 5 8 29 24 13 12 5 10 5 42 8 4 27 10 5 7 7 13 9 6 14
cva cva cva cva cva cva cva men. cva cva cva chi cva cva cva cva cva cva cva cva cva chi cva cva cva cva cva cva cva cva cva cva cva cva cva cva cva
na na tps !a na f
f:apS na tP
fip fu na na na P S
na P ; na na na ftp ftD
itp na na na na na na na na na na
Hemianopia (3) No No No No No Yes No Yes Yes Yes Yes Yes Yes Yes No Yes Yes Yes Yes Yes Yes No Yes Yes No No Yes No No No Yes Yes Yes Yes Yes No Yes
Macula? (4)
Neurophysical scale (5)
No
10
No No No No Yes No Yes No Yes No Yes Yes Yes No Yes No
9 9 8 5 14 4 13 13 12 12 14 11 14 4 13 13
FO
No Yes No No No No No na No No No na
13 12 12 13 3 12 7 8 9 15 9 10 8 14
Ks No Yes No No
14 15 7 na 8 12
WAIS Block Design (6)
WAIS PC (7)
Bit (8)
3 6 7 5 5 3 7 6 7 3 na 2 3 na 6 2 4 3 5 6 2 6 3 3 2 8 7 6 3 8 3 3 4 6 4 2 1
2 8 6 4 9 4 6 8 5 7 na 7 5 na 12 5 11 5 8 7 6 7 2 7 3 7 3 8 5 9 6 2 7 5 4 4 5
40 72 68 68 44 13 60 57 68 25 36 39 3 36 61 28 18 12 71 69 50 72 53 69 30 74 37 67 64 52 12 24 50 71 4 71 38
Digit diff Rey score Copy (9) (10) 3 1 1 2 2 2 2
1 2 1 3 3 3 3 1 4 5 3 0 2 3 2 2 1 3 1 0 2 2 3 2 3 1 1 5 2 2
22 41 23 34 14 8 17 43 31 6 na 5 na na 33 16 11 16 26 30 10 35 10 12 7 31 18 37 32 28 8 22 11 28 7 21 19
Notes: (1) cva: cerebra-vascular accident; chi: closed head injury; men: excised meningioma. (2) ct scan based. I: frontal, t: temporal, p: parietal, s: subcortical, na: not available. (3) See text for method. (4) Yes indicates macular splitting, based on Bjerrum screeen using 5 mm white disc on wand. (5) Neurophysical scale by Bond. (6) WAIS-R Block Design age scaled score. (7) WAIS-R Picture Completion age-scaled score. (8) Behavioural subtests of Behavioural Inattention test, WILSON et al. [25]; maximum score =81. (9) WAIS-R digit span: difference between maximum number of digits forward and maximum number backward. (10) Rey-Osterreith complex figure copy: number of items correctly copied out of a maxtmum of 47. Multiple regression on BIT SLOWS An automated stepwise regression was then implemented, including as the independent variables three measures of visuospatial functioning (Line Orientation, Block Design, and Rey-Osterreith copy scores) as well as Digit Span Discrepancy (an G(level of 0.1 for both entering and removing variables was chosen). The dependent variable of interest was the BIT score. This regression yielded only two significant predictive variables-the first being Digit Discrepancy scores (explaining 46.1% of variance in BIT scores) and the second Rey-Osterreith copy scores (explaining a further 15.5% of the variance). to give a total explained variance of 61.6%. This result was checked by carrying out a further stepwise multtple regression, loading in the visuospatial variables first. Together, Line Orientation, Block Design and Rey-Osterreith scores predicted 45.6 of variance in BIT scores. Adding in Digit
220
NOTE Table 2. Loadings
on three principal
components
of neuropsychological
Test
Component
loadings
I
0 RX -0.81 0.77 -0.73 -0.72 0.71 0.69 0.52 0.50 -0.16 -0.22 0.45 -0.30 0.03 -0.35 0.41
Behavioural Inattention test Forward Minus Backward Digit Span ReyyOsterreith copy Neale Reading test-accuracy Letter Cancellation accuracy PASAT 4 set Longest String Block Design PASAT 4 set total score Line Orientation Logical Memory Macular Splitting Digit Backward Digit Forvvard Word Fluency Hemianopia Rey-Osterreith Recall Variance
Percent
explamed by components. 1 2 5.6640;0 4.1oo”io of total variance explained
I
2
25.145
18.638
variables
0.05 -0.1'
0.14 0.39 0.19 0.34 0.07 0.44 0.48 0.81 0.75 0.71 0.69 0.58 0.56 -0.32
3 -0.31 0.4 1 0.07 0.17 0.40 0.18 0.28 -0.12 0.05 0.34 0.14 -0.14 0.25 -0.30 -0.30 0.81
3 3.068% 3 13.946
discrepancy to the analysis increased that by a further 18.5 to 64.1%. Gtven that Block Design is a heavily problemsolving. and hence attentional. task. the analysis was repeated excluding Block Design. Digit Discrepancy increased the explained variance of Lure Orientation and Rey copy by 20.1%. from 41.9 to 62.0%. An identical set of analyses was carried out substituting PASAT total score for Digit Discrepancy. Again the automated stepwise analysis revealed only two significant variables-Rey -0sterreith copy and PASAT tomI score. The variance in neglect explained by Line Orientation and Rey copy scores (41.9%) was increased by 10.9”,0 to 52.8% by the addition of the PASAT variable. All the above percent increases in vartance are statistically significant at least at the 0.05 level of significance. and no predictor variables in any regression were significantly correlated. Digit Discrepancy and PASAT were not mcluded in the same regresston analysis because their high intercorrelation violated one of the assumpttons of mulitiple regression.
In order to check that the results are not a feature ofone particular test of neglect. an index of neglect derived from a Letter Cancellation test 1171 was substituted for BIT. Digit Discrepancy scores increased the variance explained by Lure Orientation, Block Destgn and Rey Ostereith copy by 13.5% to 32.8 to 46.3%. This effect was statistically significant at the 0.1 level. There was a trend. but no statistically significant effect, for the inclusion of PASAT total scores.
DISCUSSION The prmciple components analysts conlirms the hypotheses of a loading of Digit Discrepancy and PASAT on the neglect prmctpal component ofthe analysts. Thts was independent ofthe second factor which loaded significantly on Logical Memory and Word Fluency. The reason why Digtt Span Discrepancy should be the variable ofchoice rather than either Forward or Backward Digit Span alone 1s quote sample. Forward Digit Span is not commonly reduced following brain damage m the absence oflanguage tmpatrment and Backward Dtgit Span is much more commonly affected 141. A Backward Digit Span of 3 tn one person who had a forward span of8, and the same in another with a forward span of 5. might mean two very diflerent thmga. The first would seem to represent a greater degree ol specific impairment on this rev*ersal task than the latter. The former IS more likely to suggest a greater degree of general impairment or of low general intelligence (with the exceptton of rare cases of specificshort-term verbal memory deficits).
Thus it is only at first sight that the loading of Digit Backward on the second component appears to contradict the hypothesis. It is unlikely to be the case given that Forward Digit Span also loads on the same component to almost exactly the same degree: these variables seem to be measuring a general intellectual deterioration rather than a speciiic deficit with reverse digits. This is supported by the fact that Loelcal Memory is the highest loading on this component, as memory is a key measure of general intellectual decline [8]. With respect to the multiple regression analysis. it is clear from the above that the tirst hypothesis is confirmed. insofar as Digit Forward minus Backward explains a significant portion of the variance m neglect. independent11 of the variance explained by tests of visuo-spatial capacity. and this was true for two different indices of neglect. BIT and Letter Cancellation. The finding was also replicated for PASAT total scores. though only for the BIT neglect scores. If then the Digit Discrepancy score still correlates with degree of neglect irrespective of lisuo-spatial representation ability. what might account for this relationship? One possible factor is that a deticlt m general attentional capacity is associated with severity of unilateral visual neglect as suggested by ROBEKTSOS [IS]. and supported by several other recent studies [ll. 12. 17. 23.1. The fact that this study found a separate test which purports to measure the same type ofcentral executive function (PASAT) to show similar results further strengthen this hypothesis. though other explanations are possible. Of course this study does not rule out the spatial processmg hypothesis of Weinberg er al. in explainmg the relationship, but it does strengthen the view that such a hypothesis does not account for all of the data. Acknowledgements-I would like to thank the following colleagues at Astley Ainslie Hospital support in carrying out this study: Brian Pentland, Mary Ramsay. John Gray and Lorna Waite. thank an anonymous reviewer for valuable comments on an earlier draft of this paper.
for their help and
I would also like to
REFERENCES 1. BADDELEY, A. D. U’orking Memory. Oxford University Press, Oxford. 1986. 2. BENTON, A. and HAIISHER. K. Mulrilingual Aphasia Euaminafion. University of Iowa Press. Iowa. 1976. 3. BESTON. A.. HANNAY. J. and VARNAY. N. Visual perception of line direction in patients with unilateral brain damage. NeurnloyJ 25, 907-910. 1975. 4. BEXTON, A.. HAMSHER.K.. VARNEY. N. and SPREES. 0. Conrrihutiotls IO Neurops~c/tological.4s.se.ww~1r. Oxford University Press. Oxford. 1983. 5. BEN-YISHAY. Y., DILLER. L.. GERSTMAN. L. and HASS. A. The relationship between impersistence. mtellectual function and outcome of rehabilitation in patients with left hemiplegia. Neuroloy!, 18, 852-861. 1968. 6. BOND. M. Assessment of psychosocial outcome after severe head injury. In Ourcome c$Serere Donlcu/t’ 10 r/u Central Nervous Sysrem. Ciba Foundation Symposmm 34 (New Series). Elsevier-Exerpta Medica. North Holland. 1975. 7. DILLER. L.. BEIU‘-YISHAY.Y.. GERSTMAN. L.. GOODKIN. R.. GORIXX. W. and WEINBERG. J. Studies irt C~rg~irr~r~ and Rehahilitarim iu Hemipleyrc Rehabilitation Monograph No. 50. New York University Medical C‘cntcr, New York. 1974 8. GIBSON. A. and KE\;uRIT~;. D. T/I@ Hendrick Battery for fhc Drrecrim of Dememia in thrr Ndwly. NFER. Windsor, 1979. 9. GRONWALL. D. and SAntpsov. H. T/I@Ps~clto/o~~~col Eflecfs of‘Concu.ssion. Auckland University Press Oxford University Press. Auckland. 1974. 10. HEILMAS. K. M. and VALFL~STEIX.E. Mechamsms underlying hemispatial neglect. Am. Nrurol. 5. I66 170. 1979. 11. LECCXRS. A. R.. MEHLER. F.. PARENTE. M A. rf ul. Ilhteracy and brain damage: II. Manifestations ofunilateral neglect in testing, “auditory comprehension” with iconographic materials. Brain C~cgt~ir. 6, 243 265. 19X7. 12. MARK. V. W.. KCK)ISTRA.C. A. and HEILMAN. K. M. Hemispatlal neglect affected by non-neglected stimuli. h’euroloyy 38, 1207 121 I. 19X8. 13. NEALE. M. ,Yua/e .Jucr/~s~s c!/ Rradimg Ahilir!. MacMillan. London. 1958. 14. NELSON, H. A moditied card sortmg test sensitive to frontal lobe deficits. Corteu 12. 313 324. 1976. 15. NELSON. H. Tllr .h’e/sor~ .4dulr Rcttdrmg Trst (NART) Tcs/ Manual. NFER, Windsor. 1982. 16. OSTERREITH. P. Le lest de coplc d‘une figure complexe. Archs de Psychologie .X0, 106 156. 1944. 17. RAPCSACK. S. Z.. VEHFAELLIE.M.. FLEET. S. and HCILMAN. K. M. Selective attention in hemispatial neglect. Arch.5 Nrurol. 46, 172 17X. 19X9. 18. REY. A. L‘examen psychologlque dans le cas d’encephaiopathie traumatique. Archs de f’.s~c/t~~/~, in the laterallty of omissions in left unilateral visual field neglect: ImpllcatlonT fob an attentional theory of’ neglect. ~europ.\~~llolt,ylrl 27, 157 165. 1989. 20. WECHSLER. D. A standardlsed memory scale for clinical use. J. P.sycho/. 19, 87 95. 1945. 21. WECHSLEK.D. Tltts Wech&r Adulf I~t~rll~qotc~~ Scnlr-Rcrwrd. The Psychological Corporation. London. 19x1.
222
NOTE
22. WE~XBERG.J.. DILLER. L.. GERSTMAN. L. and STHULMAN. P. Digit span in right and left herniplegics. J. c/in. Psychol. 28, 361, 1972. 23. WEINTRAUB, S. and MESULAM. M. Right cerebral dominance in spatial attention. Further evidence based on ipsilateral neglect. Archs Neural. 44, 621-625. 1987. 24. WILSON, B., COCKBURN. J. and HALLIGAN, P. Development ofa behavioural test of visuospatial neglect. Archs Phys. Med. Rehabil. 68, 98-102, 1987. 25. WILSON, B., COCKBURN. J. and HALLIGAN, P. Behakxu-a/ Inamnrion Test. Thames Valley Test Company. Fareham. Hampshire. 1988.
