Attentional and executive impairments in children with spastic cerebral palsy


  • Acknowledgements
    The study was funded by grants from the Faculty of Social Sciences, University of Copenhagen, and the Elsass Foundation.

Louise Bottcher at the Department of Psychology, University of Copenhagen, Oester Farimagsgade 5A, 1353 Copenhagen K, Denmark. E-mail:


Aim  Children with cerebral palsy (CP) are reported to have learning and social problems. The aim of the present study was to examine whether children with CP have impairments in attention or executive function.

Method  We examined attention and executive function with standardized neuropsychological measures in a group of children with unilateral (n=15) or bilateral (n=18) spastic CP (14 females, 19 males, mean age 11y 4mo, SD 1y 1mo, range 9y 1mo–13y 7mo; Gross Motor Function Classification System level I n=22, II n=3, III n=6, and IV n=2). Performance was compared with test norms.

Results  Verbal cognitive functioning fell within the normal range, whereas sustained (p=0.001) and divided attention (p<0.001) were found to be impaired. Greater impairment was observed in executive function in general (p<0.001) and in inhibition (p=0.038) and shifting (p<0.001) in particular. No significant difference was found between types of CP (unilateral and bilateral). Performance of all timed tasks was slower than the test norm (p<0.00).

Interpretation  The finding of slower performances across tasks may indicate a general impairment in efficiency of information processing in relation to white-matter lesions. Impairments in attention and executive functions are present in children with CP and may help to explain why these children have increased social and learning problems.

List of Abbreviations

Behaviour Rating Inventory of Executive Function


Contingency Naming Test


Test of Everyday Attention for Children


Verbal Comprehension Index

Cerebral palsy (CP) designates a group of neurodevelopmental conditions beginning in early childhood and persisting into adulthood. CP is due to a pre- or perinatal non-progressive brain disorder that often affects the periventricular white matter, either unilaterally or bilaterally.1 Although CP is primarily a disorder of movement and posture, it often involves other developing functions, such as perception and cognition. CP varies substantially in type and severity of symptoms. Seizure disorders are a common complication and are associated with a decreased level of cognitive functioning.2

CP is associated with restrictions in participation in day-to-day activities in childhood3 and a higher prevalence of specific learning disabilities,4 which affect academic achievement, school acceptance, and social participation. Children with CP are two to three times more likely than unaffected children to experience problems in their relations with peers.5 Later in life, CP is associated with limited educational and vocational possibilities.6 To obtain a better understanding of both learning difficulties and restrictions on participation in general, studies of the specific cognitive impairments associated with CP are highly relevant.

A large proportion of spastic bilateral CP is caused by periventricular leukomalacia, particularly in children born preterm. Anterior lesions to white-matter tracts are associated with attentional and executive dysfunctions.7 Lesions of the basal ganglia and thalamic functional systems may affect both focused attention8 and executive functions.

Another type of lesion associated especially with unilateral CP is infarction of the middle cerebral artery,1 which supplies several cortical and subcortical areas thought to support focused attention and motor executive function.8

Executive functions cover a range of different functions, all thought to be served by neural circuits involving the prefrontal cortex. Executive functions are defined as the ability to control impulses, anticipate consequences, establish goals, plan, monitor results, and use feedback and they are thought to regulate both immediate behaviour and planning towards long-term goals. Executive functioning is the last cognitive area to mature, and adult level is not reached until well into adolescence.9 The integrity of the white matter is important in this development because of the reliance of executive functions on an extensive interconnectivity with other parts of the brain. Impairment of executive function has been found in other neurodevelopmental disorders, along with impairments in social participation,10 and speculations can be raised whether executive functional impairments may partly explain why CP is so constraining on social participation and educational possibilities. A general picture of executive functions in children with spastic CP is missing, as the few previous studies conducted in this field have involved small subdomains only. Our aim in the present study was to examine whether children with spastic CP have specific executive or attentional impairments. Furthermore, we examined whether attentional or executive impairments are associated particularly with unilateral or bilateral spastic CP.



Thirty-three children with CP (19 males, 14 females) participated in the study (Table I). The children were recruited through the Danish Registry of Cerebral Palsy.11 In the first round of recruitment (n=39 children recruited), children were considered eligible if they had participated in the Study of Participation of Children with Cerebral Palsy Living in Europe,12 had congenital spastic CP, were born between 1993 and 1996, and had an estimated cognitive function within or above the normal range. A second round of recruitment became necessary to increase the sample size, and children from the registry who had not participated in the European participation study were also considered.

Table I.   Group characteristics of participants and non-participants
 Participants (n=33)Non-participants (n=29)
Sex, n (%)
 Females14 (42)14 (48)
 Males19 (58)15 (52)
Age, y:mo
 Mean (SD)11.3 (1.2)11.5 (1.1)
Cerebral palsy (CP) diagnosis, n (%)
 Unilateral spastic15 (46)10 (34)
 Bilateral spastic18 (54)18 (62)
Gross Motor Function Classification System level, n (%)
 I22 (67)19 (65)
 II3 (9)6 (21)
 III6 (18)3 (10)
 IV2 (6)1 (3)
Schooling, n (%)a
 Mainstream26 (76) 
 Mainstream with assistance3 (9) 
 Special class/school5 (15) 
Neurological history, n (%)aTotal CPUnilateralBilateral
  1. aData recorded for participating group only.

Apgar score at 5min >712 (36)75
Neonatal depression/irritation9 (27)54
Gestational age <32wks11 (33)65
Epilepsy3 (9)12

Data on diagnosis, gross motor function, and epilepsy were obtained from the registry. Motor function was measured by the Gross Motor Function Classification System (GMFCS). CP was diagnosed according to standards agreed by the Surveillance of Cerebral Palsy in Europe,13 distinguishing between unilateral and bilateral spastic CP. Schooling was classified as attendance at a mainstream class with no assistance, a mainstream class with assistance, or a special class or school. Children were defined as having epilepsy if they were having seizures, were on antiepileptic medication, or had had seizures in the past. No child had hearing impairments. One child had a visual impairment hindering completion of some of the measures used in the study.

From the 62 children with spastic CP considered eligible for the study, informed consent was obtained for 33 (53%). The most frequent reasons for non-participation were participation in parallel assessments, child refusal to participate, and absence of parents’ response despite repeated attempts at contact by letter or telephone. No significant difference was found between participants and non-participants in demographic variables, CP diagnosis, or GMFCS level (see Table I). The high rejection rate constrained the preferred sample size of the study.

Ethical approval was sought from the ethics committee of Copenhagen and Frederiksberg municipalities, but the committee did not consider that approval was needed.

Neuropsychological measures

Studies of intelligence in children with CP have repeatedly indicated a discrepancy between Verbal and Performance IQ, measured with the revised Wechsler Intelligence Scale for Children or the third edition of that scale (WISC-III).14–16 Calculations of Global IQ, including nonverbal measures, may be confounded because they demand timed, fine motor manipulations and visuoperceptual analyses. In the present study, therefore, we assessed general cognitive functioning using the Verbal Comprehension Index (VCI) of the WISC-III. Four subtests are included in the index: information, similarities, comprehension, and vocabulary. All subtests in the index require a verbal answer only. Block design was included as a measure of timed visuoconstruction performance. Raw scores were converted into scaled scores with a mean of 10 (SD 3). Scaled scores of the four subtests were added, and the sum was converted into the VCI with a mean reaching approximately 100. The Danish version of the WISC-III was used with recent, age-graded norms for Danish children.

Attention was measured with the Test of Everyday Attention for Children (TEA-Ch),17 which is based on the broadly accepted concept that attention consists of different parts.18 The following three subtests were used: Sky Search (focused/selective attention), Score! (sustained attention), and Sky Search Dual Task (divided attention). Raw scores were converted into scaled scores similar to those of the WISC-III. The Danish version of the TEA-Ch was used, and recent age-graded norms for Australian children were used as the comparison.

Executive functions were measured with the Contingency Naming Test (CNT),19 which consists of four subtests: two baseline measures of processes of rapid naming, and two switching tasks, one simple and one complex, to evaluate the child’s rapid memory retrieval and ability to inhibit prepotent responses and to switch mental set. Each subtest produces scores for time to complete, uncorrected errors, child-corrected errors, and efficiency given time and uncorrected errors. Analyses of performance on the CNT subtests were performed by transformation of raw scores into z-scores to reduce the effect of age on scores. Non-completion of subtest 4 was included in the analysis as categorical data. The transformation into z-scores was based on the assumption that scores were normally distributed. It is possible that the actual distribution of the scores was skewed, which would increase the risk of type II error due to reduced discrimination in the analyses. Recent age-graded norms for Australian children were used for comparison.

In addition to the CNT, executive functions were assessed with the teacher version of the Behaviour Rating Inventory of Executive Function (BRIEF),20 an 86-item questionnaire aimed at obtaining systematic information about behaviour indicative of the child’s executive functions. The BRIEF works with a measure of general executive function divided into two index scores: the behavioural regulation index and the metacognition index. These were further divided into eight subscales: inhibit, shift, emotional control, initiate, working memory, plan/organize, organize materials, and monitor. Scores were standardized on the basis of a normal distribution with a mean of 50 (SD 10). The Danish version was used, with recent, age-graded and sex-divided American norms for comparison.

After consent and assent was received from parents and children, the children’s schools were contacted, and appointments were made for neuropsychological assessment. A small group of children objected to being assessed at their school, and a home visit was arranged instead. Most children were able to complete all tests in one session with a break in the middle. Two children required two visits to complete all tests. Five children were not able to complete all tests because of motor problems. The BRIEF questionnaire was sent by mail to the child’s main teacher along with a letter with instructions for completion.

Statistical analysis

Observed mean values were compared with the standard norms using t-tests for the BRIEF, TEA-Ch, and WISC-III VCI. CNT scores were compared using Wilcoxon’s rank sum statistic. Categorical outcomes, such as likelihood of completing the CNT and non-participation, were analysed by χ2 tests. In all exploratory analyses, continuous outcomes were analysed by analysis of variance (ANOVA), with child characteristics included as categorical factors, except age, which was modelled by a basic spline with 4 degrees of freedom. Associations between BRIEF and TEA-Ch and between BRIEF and CNT were analysed by ANOVA with BRIEF score as outcome. Association between separate CNT scores was performed using Spearman’s rho. In all analyses, results were considered significant if p was less than 0.05, two-tailed. We did not correct p values for multiple comparisons.21 In all ANOVA analyses, assumptions of normality were verified by Q–Q plots, and none violated the assumptions.


General verbal cognitive functioning

The WISC-III VCI was obtained for all participating children. The mean score for the whole group was 92, which is a little below the population mean of 100, but still within the normal range. The difference between the study group and the norm population was not significant (p=0.10, Table II). The sample mean variation in the four scores that constitute the VCI was not significantly different from the mean variation in the norm population (t-test, p=0.64). No influence of GMFCS, CP diagnosis, epilepsy, or sex on VCI was found. The range of VCI within each group was very similar in children with unilateral (range=78) and bilateral (range=81) CP. The type of schooling was weakly associated with performance in the VCI, with children in normal schools scoring higher than children placed in special classes or special schools (p=0.06).

Table II.   Measures of verbal comprehension and attention in children with cerebral palsy (CP)
 Unilateral CP (n=15)Bilateral CP (n=18)All spastic CP (n=33)Norm populationt-testp value (CP vs norm)
Mean (SD)Mean (SD)Mean (SD)Mean (SD)
  1. df=32 for all reported statistical comparisons. No statistically significant difference was found between unilateral and bilateral CP. TEA-Ch, Test of Everyday Attention for Children; VCI (WISC-III), Verbal Comprehension Index of the Wechsler Intelligence Scale for Children, 3rd edition.

VCI (WISC-III)88.3 (23.0)95.5 (22.7)92.2 (22.8)100 (15)−1.710.10
 Sky Search B10.4 (2.3)10.6 (2.8)10.5 (2.5)10 (3)1.050.30
 Sky Search C8.0 (3.4)6.8 (3.2)7.4 (3.3)10 (3)−4.18<0.001
 Sky Search G8.9 (2.8)7.2 (3.4)8.0 (3.2)10 (3)−3.27<0.001
 Score!7.1 (4.3)8.0 (3.1)7.6 (3.7)10 (3)−3.62<0.001
 Sky Search Dual Task4.5 (3.4)6.6 (3.8)5.6 (3.7)10 (3)−6.32<0.001

Performance in block design was significantly below the norm population (t-test, p<0.01, data not shown). This subtest has not been further analysed in this article.


No correlation was found between the attentional subdomains measured by subtests from the TEA-Ch and the VCI, which points to attention as a cognitive domain that can be measured independently from general verbal cognitive functioning. Five children were unable to complete the Sky Search and the Sky Search DT, because of motor impairments (n=4) or motor and visual impairment (n=1). Three of these five children were not able to complete the Score! subtest either.

No significant difference was found between children with unilateral CP and those with bilateral CP, and the two groups were pooled. The distribution of scores on the different TEA-Ch measures is shown in Table II. Sky Search B is a non-timed measure of the ability of the child to detect visual targets placed among distracters; on this measure, the performance mean of children in the study was similar to that of the population mean. On Sky Search C, the mean performance of the children in the study was significantly below that of the population mean; however, this measure has a motor component. Given the motor slowness associated with CP, Sky Search G appears to be a more valid measure of focused attention because it adjusts for motor speed. Eight children scored at least 1SD below the population mean on Sky Search G, and four children scored 2SD or more below the population mean. In the Score!, 13 children scored more than 1SD below the population mean, and seven scored 2SD or more below the population mean. In Sky Search DT, 16 children scored more than 1SD below the population mean, and 11 scored 2SD or more below the population mean. All differences except Sky Search B were significant (see Table II). No effect of GMFCS was found (data not shown).

Executive function

Thirty-two children were able to complete at least some of the subtests of the CNT. The children who did not complete all subtests were unable to apply the rules for the two switching tasks (subtests 3 and 4) in successive trials of the learning phase; this in itself provides useful information about executive functioning in children with CP.

The children with CP performed more poorly than the norm group in all CNT subtests (Table III). The group with bilateral CP performed below the group with unilateral CP, but no significant difference was found. This is further discussed in the section on the limitations of the study. From the pooled CP group, 31% of participants scored more than 1SD below their age norm on subtest 3, and 12% found the task too difficult and had to give up on subtest 3. The differences were significant in the first three subtests (see Table III), although the number of children able to complete subtest 4 was possibly too low to reveal a significant difference in the level of performance. However, children in the age range studied are expected to be able to complete all four subtests, which implies that the performance of the study group in subtest 4 fell significantly short of expectations (χ2, p<0.05, data not shown).

Table III.   Measures of executive function in children with cerebral palsy (CP)
 Unilateral CP (n=12)Bilateral CP (n=18)All spastic CP (n=30)Norm population p value (CP vs norm)
Mean (SD)Mean (SD)Mean (SD)Mean (SD)
  1. aWilcoxon’s rank sum statistic. df=29 for all reported statistical comparisons. No statistically significant difference was found between unilateral and bilateral CP. All CNT scores are converted to z-scores. BRIEF, Behaviour Rating Inventory of Executive Function; CNT, Contingency Naming Test.

CNT subtest    W testa 
 1: time−0.6 (1.6)−2.2 (2.9)−1.5 (2.5)0 (1)62.0<0.001
 1: efficiency−0.5 (1.1)−1.4 (1.6)−1.0 (1.5)0 (1)56.0<0.001
 2: time−0.9 (1.3)−2.5 (3.6)−1.8 (2.9)0 (1)35.5<0.001
 2: efficiency−0.9 (1.2)−1.3 (1.2)−1.1 (1.2)0 (1)42.5<0.001
 3: time−1.2 (2.9)−1.2 (1.8)−1.2 (2.4)0 (1)35.5<0.001
 3: efficiency−0.8 (1.2)−0.7 (0.7)−0.8 (1.0)0 (1)34.50.001
 4: time−0.2 (1.4)−1.4 (2.3)−0.9 (2.0)0 (1)54.00.10
 4: efficiency0.2 (1.1)−0.6 (0.9)−0.3 (1.1)0 (1)91.50.16
BRIEF subtest    t-test 
 Inhibit54.9 (7.0)52.1 (8.8)53.2 (8.1)50 (10)2.170.038
 Shift59.7 (12.7)59.4 (12.8)59.6 (12.5)50 (10)4.18<0.001
 Emotional control57.5 (11.3)54.9 (11.8)56.0 (11.5)50 (10)2.840.008
 Behavioural index58.0 (8.6)56.0 (11.2)56.8 (10.1)50 (10)3.67<0.001
 Initiate58.7 (20.3)62.5 (15.1)60.5 (17.9)50 (10)3.94<0.001
 Working memory57.0 (15.9)57.9 (14.0)57.6 (14.5)50 (10)2.840.008
 Plan/organize55.1 (11.9)58.5 (12.3)57.1 (12.0)50 (10)3.240.003
 Organize materials51.4 (14.5)56.1 (11.2)54.2 (12.6)50 (10)1.830.076
 Monitor58.8 (13.5)55.2 (10.5)56.6 (11.7)50 (10)3.090.004
 Metacognition index57.5 (13.9)59.8 (13.4)58.9 (13.4)50 (10)3.620.001
 General executive function58.3 (11.2)58.2 (11.5)58.2 (11.2)50 (10)4.03<0.001

Differences in performance means of children with CP and their norm populations were greater in the measures of time than in the measures of efficiency: subtest 1 displayed a difference of 0.50, subtest 2 a difference of 0.7, subtest 3 a difference of 0.4, and subtest 4 a difference of 0.60. This difference was most pronounced in the group with bilateral CP (see Table III). This points to impairments in both quality and speed of performance.

BRIEF scores were obtained for 30 children. Overall comparisons (t-tests) on the BRIEF revealed significant differences between the study mean and the population mean (see Table III). No significant difference was found between children with unilateral and bilateral CP. In the pooled group, more executive problems were reported both in the subscales and in the summary index scores, with the exception of the organization of materials subscale. Inspection of the mean scores on the different BRIEF domains and the composite scores revealed that the most problematic areas were initiation (mean T score 60.5), shift (mean T score 59.6), and the composite meta-cognition (mean T score 58.9).


The finding of a slightly lower overall level of verbal cognitive functioning in children with spastic CP is in line with findings from other studies.22,23 The high VCI scores of the children with bilateral CP indicate that this group has unusually high functioning, which might account for the fact that verbal comprehension was a little higher in this group than in the group with unilateral CP. The TEA-Ch results point to attention as a vulnerable cognitive area following early brain insult. The children with CP had particular difficulty in measures of sustained and divided attention. These results are consistent with those of previous studies.23,24 Disorders of attention are associated with distractibility and inattention, which give rise to learning difficulties and problems in social relationships. More research is needed in the different aspects of attention in children with different types of CP.

The finding of impaired functioning in measures of impulse inhibition is in agreement with the few existing studies of impulse inhibition in children with unilateral CP24 or preterm diplegia.25 The finding of problems in all BRIEF subdomains but one points to a heightened level of behavioural problems at school for children with spastic CP and may shed new light on earlier reports of social and behavioural problems in these children.5 Impairments in executive functioning have been associated with behavioural problems and, in other studies, impairments in social skills with developmental disorders.10

In addition, the finding of more problems in some domains of executive functions than in others draws attention to the importance of interpreting the results in relation to different executive subdomains rather than general executive functions. Research involving larger groups of children with spastic CP is needed to examine whether executive impairments in areas of initiation, inhibition, and shifting are a robust finding, so too is research combining measures of executive and social functioning.

The reduced performance in the two CNT baseline tasks was supplementary to and consistent with another recent study of children with spastic diplegia due to periventricular leukomalacia.24 A similar result has previously been interpreted as evidence of deficits in processing speed among children with white-matter abnormalities due to phenylketonuria.26 The low performance was found in several timed tasks: the TEA-Ch Sky Search and Sky Search Dual Task subtests and the block design on the WISC-III. A lowered information processing speed in children with lesions in the periventricular areas is biologically plausible, as the rapid conduction of information between cortical areas and subcortical structures is served by fibres in the periventricular white matter. The necrosis and gliosis associated with periventricular leukomalacia interfere with the formation of oligodendrocytes, and through this interference myelination and normal rapid information processing are disrupted in the affected areas. Further research in this area is warranted.


Differences in measures of information processing speed between children with unilateral and bilateral CP would have been an interesting finding, because periventricular leukomalacia and white-matter injuries are mostly associated with bilateral spastic CP. Such results could contribute important information that would assist our understanding of neuropsychological functioning in children with spastic CP. It was rather surprising therefore that in the present study, no significant difference was found between participants with unilateral and bilateral spastic CP. One explanation might be that the participating children with bilateral CP were unusually highly functioning, thereby making the effect size between the two groups smaller than expected. The small sample size of this study has further limited the possibility of finding significant differences between subgroups, as calculations revealed rather low levels of power when the sample was separated into unilateral and bilateral groups (Table IV).

Table IV.   Power calculations for selected measures
InstrumentSubtest[Group size, n]EffectSDPower
  1. TEA-Ch, Test of Everyday Attention for Children; BRIEF, Behaviour Rating Inventory of Executive Function; CNT, Contingency Naming Test.

TEA-ChSky Search G15181.73.20.31
CNT1: time1218−
1: efficiency1218−
3: time1218−
3: efficiency1218−


The finding of impairments in attention and executive functions underscores the vulnerability of these distributed functions. Attention and executive functions are both extremely important in relation to learning and social development. Disruptiveness or excessive perseveration obstructs the child’s ability to engage in social relationships and impairs the child’s ability to learn both cognitive operations, such as reading and mathematics, and important social skills. Impairments in attention and executive functions may, therefore, help to explain why children with spastic CP often experience social problems in addition to their motor and cognitive impairments.