doi:10.

1093/brain/awn293

Brain 2009: 132; 4556

| 45

BRAIN
A JOURNAL OF NEUROLOGY

Childhood brain insult: can age at insult help

us predict outcome?
Vicki Anderson,1,2,5,6 Megan Spencer-Smith,1,5 Rick Leventer,1,3,6 Lee Coleman,4
Peter Anderson,1,5 Jackie Williams,1,5 Mardee Greenham1 and Rani Jacobs1,5
1
2
3
4
5
6

Murdoch Childrens Research Institute,

Department of Psychology,
Neuroscience,
Department of Radiology, RCH, Melbourne,
School of Behavioural Sciences and
Department of Paediatrics, University of Melbourne

Correspondence to: Vicki Anderson,

Department of Psychology,
Royal Childrens Hospital, Flemington Road,
Parkville, Victoria, 3052, Australia
E-mail: vicki.anderson@rch.org.au

Until recently, the impact of early brain insult (EBI) has been considered to be less signicant than for later brain injuries,
consistent with the notion that the young brain is more exible and able to reorganize in the context of brain insult. This study
aimed to evaluate this notion by comparing cognitive and behavioural outcomes for children sustaining EBI at different times
from gestation to late childhood. Children with focal brain insults were categorized according to timing of brain insult,
represented by six developmental periods: (i) Congenital (n = 38): EBI: rstsecond trimester; (ii) Perinatal (n = 33); EBI: third
trimester to 1 month post-natal; (iii) Infancy (n = 23): EBI: 2 months2 years post-birth; (iv) Preschool (n = 19): EBI: 36 years;
(v) Middle Childhood (n = 31): EBI: 79 years; and (vi) Late Childhood (n = 19): EBI: after age 10. Groups were similar with
respect to injury and demographic factors. Children were assessed for intelligence, academic ability, everyday executive function
and behaviour. Results showed that children with EBI were at increased risk for impairment in all domains assessed.
Furthermore, children sustaining EBI before age 2 years recorded global and signicant cognitive decits, while children with
later EBI performed closer to normal expectations, suggesting a linear association between age at insult and outcome. In
contrast, for behaviour, children with EBI from 7 to 9 years performed worse than those with EBI from 3 to 6 years, and
more like those with younger insults, suggesting that not all functions share the same pattern of vulnerability with respect to
age at insult.

Keywords: brain injury; plasticity; outcome; IQ; executive function; behaviour

Abbreviations: AL = age at lesion; BRIEF = Behavioral Rating Inventory of Executive Function; EBI = early brain insult;
ES = emotional symptoms; FSIQ = Full Scale Intelligence Quotient; HYP = Hyperactivity-Inattention Scale; PIQ = Performance
Intelligence Quotient; SDQ = Strengths and Difculties Questionnaire; SES = Socio-economic status

Introduction
Despite a lively and continuing interest in this area, recovery
from early brain insult (EBI) remains imperfectly understood.

Pathological conditions that would almost certainly lead to

severe cognitive dysfunction in an adult have quite different consequences for children (Aram, 1988; Bates et al., 2001; Jacobs
and Anderson, 2002; Anderson et al., 2005). Children with focal

Received May 20, 2008. Revised September 1, 2008. Accepted October 10, 2008
The Author (2009). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: journals.permissions@oxfordjournals.org

46

| Brain 2009: 132; 4556

left-hemisphere insult, for example, may go on to acquire age

appropriate language abilities, free from the symptoms of aphasia
observed following similar lesions in adulthood (Heywood
and Canavan, 1987; Taylor and Alden, 1997). Similarly, early
vascular accidents need not preclude normal or higher
intellectual and academic achievements (Smith and Sugar, 1975;
Ballantyne et al., 2008). Even when an entire cerebral hemisphere is removed, children may develop relatively normal cognitive function (Dennis and Whittaker, 1976). In contrast, children
sustaining generalized cerebral insult (e.g. traumatic brain injury)
display slower recovery and poorer outcome than adults with
similar insults (Anderson and Moore, 1995; Gronwall et al.,
1997; Taylor et al., 2002; Anderson et al., 2004, 2005). These
somewhat unpredictable recovery patterns after EBI are puzzling
and restrict health professionals capacity to identify children at
high risk for sequelae who may need more intensive follow-up
and intervention.
In search of a more accurate prognostic formula, researchers
have considered a range of factors that might reasonably be
assumed to inuence recovery. However, apart from the established relationship between insult severity and outcome, these
studies have failed to identify consistent links between outcome
and specic insult characteristics (e.g. diffuse versus focal pathology, laterality) (Bates et al., 2001; Herz-Pannier et al., 2002;
Chilosi et al., 2005; Stiles et al., 2008), presence of residual disability (e.g. hemiparesis, epilepsy) (Hartel et al., 2004; Chilosi
et al., 2005; Ballantyne et al., 2007), pre-insult child and family
factors (Ponsford et al., 1999; Anderson et al., 2006), and environmental parameters (e.g. socio-demographics, access to interventions, parent/family function) (Breslau, 1990; Taylor et al., 2002;
Anderson et al., 2006; Catroppa et al., 2008). While each of these
factors appears to contribute incrementally to outcome, we fall
short of providing a complete picture of relevant predictors and
their interactions.
A further potential piece in the puzzle is the developmental
stage of the child at time of insult, with major controversy existing
regarding the potential impact of this dimension for neurobehavioural and psychological outcome. This debate is best illustrated by the contrasting plasticity versus early vulnerability
approaches, which dispute whether the immature brain has a
greater capacity for recovery than the mature or adult brain.
This debate is argued at both biological and cognitive levels,
although this article will limit its focus to the functional dimension. Plasticity theorists postulate that the young brain is immature, less committed and thus less susceptible to the impact of
cerebral damage. Plasticity is thought to be maximal early in
development when the central nervous system (CNS) is less rigidly
specialized (Kennard, 1936, 1940; Huttenlocher and Dabholkar,
1997), and synapses and dendritic connections remain unspecied.
Such exibility provides the capacity for transferring or reorganizing functions from damaged brain to healthy tissue. In contrast,
early vulnerability proponents postulate that the young brain is
uniquely sensitive to insult, and thus EBI is detrimental to development. Donald Hebb (1947, 1949) argued that plasticity theories
ignored the possibility that brain insult will have different consequences at different times throughout development. He concluded
that EBI may be more detrimental than later injury, because

V. Anderson et al.
cognitive development is critically dependent on the integrity of
particular cerebral structures at certain stages of development.
Thus, if a cerebral region is damaged at a critical stage of cognitive
development it may be that cognitive skills dependent on that
region are irreversibly impaired (Kolb, 1995; Luciana, 2003).
A review of the literature relevant to these theories provides
little clarication. While it is now evident that the young
brain has some capacity for neural restitution, via either neural
regrowth or anatomical reorganization (Kolb, 2005; Giza, 2006),
there is ongoing controversy as to the implications of these
processes. Even if neural restitution does occur, full recovery
may be limited by either: (i) inappropriate connections being
established (Stein and Hoffman, 2003; Kolb et al., 2004) resulting
in dysfunctional behavioural recovery; or (ii) a crowding effect
(Vargha Khadem et al., 1992; Aram and Eisele, 1994), where
functions normally subsumed by damaged tissue are crowded
into remaining healthy brain areas, with a general depression
of all abilities. In support of such concerns, studies of children
with pre-natal lesions, or those sustaining insults during the
rst year of life, consistently report poorest functional outcomes
(Riva and Cassaniga, 1986; Duchowny et al., 1996; Anderson
et al., 1997; Ewing-Cobbs et al., 1997; Leventer et al., 1999;
Jacobs et al., 2007).
While the debate continues, there is little disagreement that
developmental factors play a central role in outcome from EBI.
The challenge remains to describe the nature of this relationship.
To date, most research has employed single-condition approaches
(e.g. dysplasia, traumatic brain insult), examining age effects
within such conditions. Such designs are unable to investigate
consequences of insults sustained from gestation to adolescence,
as these conditions are necessarily age-specic (e.g. traumatic
brain injury is post-natal). To investigate developmental inuences
comprehensively, studies need to incorporate conditions occurring
throughout gestation and childhood. Further, previous research
has often assumed that age effects will be linear, that is, the
younger the insult the poorer the outcome. Such an assumption
is inconsistent with knowledge of brain maturation, where development is step-wise (Casey et al., 2000; Gogtay et al., 2004),
with critical maturational periods for processes such as myelination
and synaptogenesis (Klinberg et al., 1999; Gogtay et al., 2004),
separated by more stable periods. Cognitive theorists describe
similar stage-like processes (Piaget, 1963; Flavell, 1992). It is
likely that disruption during one of these predetermined, neural
or cognitive growth periods will cause ow on effects, as the
establishment of other later emerging skills is thrown off course
(Mosch et al., 2005). To date, insufcient evidence is available
to pinpoint the timing or scope of these critical periods for
humans, however, animal literature provides some insights (Kolb,
1995; Kolb et al., 2005), suggesting that age and recovery are not
linearly related, but are associated, via underlying neural processes
such as synaptogenesis, dendritic aborization and myelination.
This study, we believe, is the rst to attempt to systematically
address these age-related hypotheses, by examining outcomes
from EBI sustained across gestation and childhood, when brain
development is most rapid. To address the potential non-linearity
between age and outcome, we have constructed age at lesion
(AL) groups, dened according to developmental timetables

Can age at insult predict outcome?

Brain 2009: 132; 4556

for key neurological processes in the prefrontal cortex as well as

developmental timetables for cognitive processes which recruit this
region. These groupings are consistent with principles emerging
from animal studies (e.g. Kolb et al., 2004), and preliminary
child studies (Pavlovic et al., 2006; Jacobs et al., 2007) which
identify critical periods for neural development, and account
for parallels in brain structure and function. These parallels are
highlighted by Goldman Rakic (1987), who has shown that skill
emergence is tightly linked to the peak period of synaptogenesis
in the brain region by which it is underpinned. These studies
suggest that both neurological and cognitive developmental
processes occurring at the time of brain insult are central to outcomes. Of note, these AL groups were necessarily heterogeneous
for cause of insult as many CNS insults occur only at specic
stages of development (e.g. penetrating head injury, developmental malformations). To minimize any confounding effects caused
by this heterogeneity: (i) only children identied as having focal
abnormalities on MRI scan were included in the sample; and
(ii) AL groups were compared with respect to lesion characteristics
(size, location, laterality). In this context we have addressed the
following questions: (i) Is EBI associated with decits in intelligence, academic achievement, and executive abilities in the daily
context and emotional and psychological function? and (ii) Does
age at brain injury have long-term implications for these outcomes
in later childhood and adolescence?

Method

| 47

Eligible children were identied via hospital records and consecutive

referrals to neuroscience outpatient clinics.
Inclusion criteria were: (i) aged 1016 at assessment; (ii) MRI evidence of focal brain pathology; (iii) brain insult at least 12 months
prior to assessment, to allow for stabilization of recovery processes;
(iv) cognitive skills sufcient to participate in study protocol. Exclusion
criteria were: (i) evidence of diffuse pathology (e.g. closed head injury)
on MRI scan; and (ii) non-English speaking. We did not exclude children based on low IQ as we were interested in achieving a representative sample. Eleven children were excluded based on study criteria.
Approaches were made to 215 families, with 51 declining to participate (77% participation rate) due to time burden (n = 18), lack of
interest (n = 29) or distance (n = 3). Table 1 provides demographic
information on the sample.
The sample was divided into six AL groups, based on timing of
cerebral growth spurts (van Praag et al., 2000; Kolb et al., 2004):
(i) Congenital (n = 38): EBI during rst and second trimester;
(ii) Peri-natal (n = 33); EBI within the third trimester to 1 month
post-natal; (iii) Infancy (n = 23): EBI 2 months2 years post-birth;
(iv) Preschool (n = 19): EBI 36 years of age; (v) Middle childhood
(n = 31): EBI 79 years of age; and (vi) Late childhood (n = 19): EBI
after age 10.
Diagnoses were necessarily diverse, in order to provide sufcient
children with EBI across the developmental span of interest, and
included stroke, contusion from falls, penetrating head injury,
tumour, malformation, dysplasias, cyst and abscess. Details of the
mechanism of insult and the extent, laterality and region of lesion
across the groups are provided in Table 2.

Materials

Sample

Demographic information

The sample comprised 164 children, including 92 (56.1%) males, aged

between 10 and 16 years at recruitment (Mean = 13.07, SD = 1.88),
with a history of EBI. Participants were ascertained between 2005 and
2007, through the Royal Childrens Hospital, Melbourne, Australia.

In a structured interview parents provided information on their childs

medical and developmental history, academic progress and parental
occupation and educational level. Socio-economic status (SES) was
determined using Daniels Scale of Occupational Prestige (Daniel,

Table 1 Demographics of sample

Congenital

Perinatal

Infancy

Preschool

Mid Childhood Late Childhood Total group

n
38
33
23
19
31
Gender, n (%) males
19 (50.0)
23 (69.7)
13 (56.5)
12 (63.2)
16 (51.6)
SES Mean (SD)
4.40 (1.4)
4.07 (0.84) 4.04 (1.06) 4.09 (1.13) 4.21 (1.29)
Age at testing Mean (SD)
12.97 (1.86) 13.24 (1.98) 12.48 (1.97) 12.57 (1.72) 12.90 (1.72)
Age at insult Mean (SD)
NA
NA
1.35 (0.93) 4.80 (1.07) 8.30 (0.80)
Time since insult Mean (SD)
NA
NA
11.10 (2.19) 7.78 (1.98) 4.59 (2.10)
Age at diagnosis Mean (SD)
3.56 (3.91) 1.96 (2.19) 1.60 (1.23) 4.95 (1.03) 8.47 (1.05)
Time from diagnosis Mean (SD) 9.40 (3.68) 10.97 (3.68) 10.79 (1.85) 7.68 (1.99) 4.30 (2.05)
Handedness (Right), n (%)
19 (50.0)
23 (69.7)
13 (56.5)
12 (63.2)
16 (51.6)
Developmental delays
Speech delay, n (%)
16 (42.1)
12 (36.4)
4 (17.4)
0
3 (9.7)
Motor delay, n (%)
19 (50.0)
15 (45.5)
5 (21.7)
2 (10.5)
2 (6.5)
No delays, n (%)
14 (36.8)
11 (33.3)
17 (73.9)
17 (89.5)
24 (77.4)
Academic assistance
Special school/aide, n (%)
15 (39.7)
16 (48.5)
9 (39.1)
4 (21.1)
9 (29.0)
Extra tuition, n (%)
12 (31.3)
5 (15.2)
7 (30.4)
3 (15.8)
7 (22.6)
Normal progress, n (%)
11 (30.0)
12 (36.3)
6 (26.1)
12 (63.1)
15 (48.4)
Seizures
24 (63.1)
16 (48.5)
14 (60.9)
5 (26.3)
11 (35.5)


P50.001, P50.01.

20
9
4.25
14.45
11.85
2.50
11.91
2.54
9

(45.0)
(1.11)
(1.46)
(1.60)
(1.30)
(1.59)
(1.28)
(45.0)

0
1 (5.0)
16 (80)
2
5
13
5

(10.0)
(25.0)
(65.0)
(25.0)

164
92
4.20
13.07
NA
NA
5.20
7.79
92

(56.1)
(1.06)
(1.86)

(4.27)
(4.14)
(56.1)

35 (21.3)
44 (26.8)
101 (61.6)
56
39
69
75

(34.1)
(23.8)
(42.1)
(46.9)

48

| Brain 2009: 132; 4556

V. Anderson et al.

1983), which rates parent occupation on a 7-point scale, where a high

score reects low SES.

(developmental, infective, ischaemic, neuroplasm or traumatic).

Details are provided in Table 3. A subset of four scans was double
coded, demonstrating 97.5% internal consistency.

MRI scans
(i) Acquisition: MRI scans were conducted as part of routine clinical
practice prior to recruitment. For those who had not undergone scanning, or whose scans were unavailable, scans were conducted simultaneously with neurobehavioural evaluation. All scans were conducted
on a 1.5 Tesla scanner, and axial and coronal slices were obtained. (ii)
Coding protocol: A coding protocol developed by Leventer et al.
(1999) was employed to describe brain insult characteristics including:
brain regions affected (lobes, subcortical structures), laterality (left,
right, bilateral), extent of insult (focal, multifocal), volume of brain
affected (number of regions) and mechanism of brain insult

Table 2 Summary of terms for brain insult characteristics

and related factors
Variable
Region
Frontal
Extra-frontal
Subcortical
Posterior fossa
Laterality
Left hemisphere
Right hemisphere
Bilateral
Extent
Focal
Multifocal
Diffuse

Denition
Brain pathology involves the
Brain pathology involves the
or temporal lobe.
Brain pathology involves the
thalamus or basal ganglia.
Brain pathology involves the
cerebellum.

frontal lobe.
parietal, occipital
corpus callosum,

Brain insult
Timing of brain insult was based on a combination of MRI, brain
biopsy, and medical record (clinical history, medical investigations).
For acquired brain insults, where timing is generally precise, ratings
were based on information provided by clinical history. In contrast,
for pre- and peri-natal events, rating of injury timing was not precise,
but was divided into trimesters, based on current understanding of
the likely timing of specic structural abnormalities. Where timing
of insult was not evident, consensus was reached through discussion
with a paediatric neurologist and neuropsychologist. Ten cases were
double-rated, with 100% consistency. Mechanism of insult was coded
as: developmental, infective, ischaemic, neuroplastic, or traumatic.
Presence of seizure history and neurological abnormalities were
noted. Age at diagnosis indicated the time at which the brain condition was identied and diagnosed. For acquired injuries age at diagnosis and time of insult were identical. For pre- and peri-natal insults,
diagnosis was frequently delayed, although for the majority of children
developmental delay had been detected and early interventions implemented from an early age.

Neurobehavioural measures

brain stem or

Brain pathology conned to left hemisphere.

Brain pathology conned to right hemisphere.
Brain pathology located in both left and right
hemispheres.
Brain pathology is conned to one area.
Brain pathology involves two or more areas.
Brain pathology involves the whole brain.

(i) Intelligence: The 4-subtest version of the Wechsler Abbreviated

Intelligence Scale (WASI: Wechsler, 1999) was administered.
Scores derived were Verbal (VIQ), Performance (PIQ) and Full
Scale Intelligence Quotients (FSIQ) (Mean = 100, SD = 15).
(ii) Academic Ability: The Wide Range Achievement Test-3
(WRAT-3: Wilkinson, 1993) assessed reading, spelling and arithmetic. Scaled scores were used in analyses (Mean = 10, SD = 3).
(iii) Executive function in everyday life: The Behavioral Rating
Inventory of Executive Function (BRIEF: Gioia et al., 2000)
(parent and teacher) provided an index of attention and

Table 3 Insult mechanism across age at lesion groups

n
Mechanism of pathology
Developmental, n (%)
Infective, n (%)
Ischaemic, n (%)
Neuroplastic, n (%)
Traumatic, n (%)
Regiona
Frontal, n (%)
Extrafrontal, n (%)
Subcortical, n (%)
Laterality
Left, n (%)
Right, n (%)
Bilateral, n (%)
Extent
Focal, n (%)
Multifocal/diffuse, n (%)

Congenital

Perinatal

Infancy

Preschool

Middle Childhood

Late Childhood

Total Group

38

33

23

19

31

20

164

30
0
2
5
0

(78.9)
(0)
(5.3)
(13.2)
(0)

5
0
25
3
0

(15.2)
(0)
(75.8)
(9.1)
(0)

0
1
6
14
2

(0)
(4.3)
(26.1)
(60.9)
(8.7)

0
1
6
8
4

(0)
(5.3)
(31.6)
(42.1)
(21.1)

0
1
12
10
8

(0)
(3.2)
(38.7)
(32.3)
(25.8)

0
1
6
6
7

(0)
(5.0)
(30)
(30)
(35.0)

35
4
57
46
21

(21.3)
(2.4)
(34.8)
(28.0)
(12.8)

21 (55.3)
21 (55.3)
29 (76.3)

23 (69.7)
23 (69.7)
23 (69.7)

8 (24.8)
8 (24.8)
17 (73.9)

9 (47.2)
9 (47.2)
11 (57.9)

14 (45.2)
14 (45.2)
14 (45.2)

11 (55.0)
11 (55.0)
11 (55.0)

86 (52.4)
105 (64.0)
91 (55.5)

8 (21.1)
9 (23.7)
21 (55.3)

11 (33.3)
6 (18.2)
16 (48.5)

9 (39.1)
8 (34.8)
6 (26.1)

8 (42.1)
5 (26.3)
6 (31.6)

14 (45.2)
7 (22.6)
10 (32.3)

4 (20.0)
9 (45.0)
7 (35.0)

54 (32.9)
44 (26.8)
66 (40.2)

18 (47.4)
20 (52.6)

18 (54.5)
15 (45.5)

16 (69.6)
7 (30.4)

14 (73.7)
5 (26.3)

18 (58.1)
13 (41.9)

13 (65.0)
7 (35.0)

97 (59.1)
67 (40.9)

a There is some overlap across categories for this variable.


P50.001.

Can age at insult predict outcome?

executive abilities in everyday life. The BRIEF comprises 86 items
over two subscales: (i) Behavioral Regulation (BRI); and (ii)
Metacognition (MCI). A Global Executive Composite (GEC)
was also calculated (Mean = 50, SD = 10).
(iv) Psychological function: The Strengths and Difculties Questionnaire (SDQ: Goodman, 1997) (parent and teacher) rated childrens
behaviour over the previous 6 months. The SDQ is scored on a
likert scale and includes 25 items, providing ve subscales: Emotional Symptoms (ES), Conduct Symptoms (CS), HyperactivityInattention (HYP), Peer Problems (PP), Prosocial Behaviour
(PSB). A Total Difculties (TOT) score was derived and ranked
as normal, borderline and abnormal, as per scoring instructions.

Procedure
This study was approved by the Human Research Ethics Committee,
Royal Childrens Hospital, Melbourne, Australia. Eligible children were
identied via medical records, neuroradiology meetings or outpatient
clinics. Families were contacted to seek their participation and mailed
details of the study and requests for written consent. Consenting
families were seen at an outpatient clinic, with a small number of
children assessed at home, for family convenience. Children were evaluated on an individual basis, by a trained child psychologist.
Assessments lasted approximately one hour and during this time parents completed questionnaires.

Statistical analysis
For the WASI and WRAT-3, some children were unable to complete
measures due to low functioning. In these cases, missing data were
recoded conservatively to two SDs below the mean. For the WASI,
eight children in the Prenatal and one in the Perinatal group were
recoded. For the WRAT-3, eight children in the Prenatal, two in the
Perinatal and one in the Infancy group were recoded. Data missing
for other reasons (e.g. failure to return a questionnaire) were not
recoded.
Quantitative analyses were conducted using SPSS (version 14.0). AL
groups were compared (ANOVA) to identify demographic differences.
To address hypothesis 1, the total sample was compared to published
test norms, using single sample t-tests. For hypothesis 2, preliminary
analyses were conducted to determine group differences for demographic and medical variables, using ANOVA. Tukeys HSD analyses
were used to identify individual group differences. MANOVA was
used to compare AL groups across each domain. Effect size was determined by 2. When group differences were observed, Tukeys HSD
was calculated.
Power analysis was conducted prior to study commencement,
based on results from our previous studies (Anderson et al., 2004,
2005), and indicated that the study required a sample of 20 participants per group (one-tailed alpha 0.05, power set at 0.95) to detect
a difference of 2/3 to 1 SD, that is a clinically signicant difference,
between the groups. AL group size was determined accordingly.
Individual impairment scores were also derived, based on test manuals. For the WASI and WRAT-3: (i) normal function: within 1 SD
of test mean; (ii) mild impairment: one to two SD below test mean;
(iii) severe impairment: 42 SD below test mean. For the BRIEF:
(i) normal function: 565; and (ii) clinical range: 65 or above. For
the SDQ, total scores were classied as normal, borderline or abnormal, using the following cut-offs: parent version: normal: 013;
borderline 1416; abnormal: 1740; teacher version: normal: 011;
borderline 1215; abnormal: 1640. Chi-square analyses were

Brain 2009: 132; 4556

| 49

conducted on these data. Due to small numbers in some cells, mild

and severe impairment groups were collapsed for these analyses.

Results
Sample characteristics
No group differences were identied for gender, SES or handedness. A signicant age at test difference was identied,
F(5,158) = 3.21, P = 0.009, 2 = 0.09, revealing that Late
Childhood group was older than the Congenital (P = 0.04),
Infancy (P = 0.007), Preschool (P = 0.02), and Middle Childhood
(P = 0.037) groups. Age was not used as a covariate in analyses,
however, as all measures reported are age-standardized. As
expected given the nature of the groups, group differences were
also present for age at diagnosis, F(5,149) = 64.25, P50.001,
2 = 0.68, and time since diagnosis, F(5,149) = 39.95, P50.001,
2 = 0.57. AL groups also showed distinct differences with respect
to outcomes. Risk of developmental delay, as reported by primary
caregiver, was associated with earlier AL, 2(20, n = 163) = 39.12,
P50.001, V = 0.50, with high frequency of such delays in
Congenital and Peri-natal groups. For academic assistance, again
the earlier AL groups (Congenital, Peri-natal and Infancy) had
high rates, but group differences failed to reach signicance,
2(10, n = 162) = 17.93, P = 0.06, V = 0.24. Finally, signicant
group differences were identied for presence of seizures, 2(5,
n = 158) = 17.44, P = 0.004, V = 0.332, with a large proportion
of children in the Congenital group with epilepsy/seizures
(SR = 1.7) and a small proportion in the Preschool group (Table 1).
Mechanisms of insult are provided in Table 2, illustrating signicant group differences, 2(20, n = 163) = 150.10,
P50.001, V = 0.48, consistent with the heterogeneity of
the sample. There were no group differences for region of insult
[frontal, 2(5, n = 164) = 7.84, P = 0.17, V = 0.22; extrafrontal,
2(5, n = 164) = 9.74, P = 0.08, V = 0.24; subcortical, 2(5,
n = 164) = 5.08, P = 0.41, V = 0.18], or extent of insult (unifocal/
multifocal), 2(5, n = 164) = 5.46, P = 0.36, V = 0.18, suggesting
that groups did not differ signicantly with respect to lesion
characteristics.

Comparing EBI to normative

expectations
As illustrated in Table 4, using total group data, children with EBI
achieved poorer scores than the normal population (P50.001) on
all measures. For the WASI, the EBI group means ranged from
approximately 3/5 to 1 SD below normative means, with greatest
discrepancies recorded for VIQ (14.12 points). For academic ability, all mean differences were signicant (P50.001) and greater
than 2/3 SD. Of note, Arithmetic abilities were particularly poor
in the EBI group (scaled score = 6.13), with group mean 41SD
below expectations.
Parent ratings of executive function (BRIEF) were elevated relative to test expectations (all P50.001). BRI, MCI and GEC were all
41 SD above the test mean. Teacher ratings indicated even
greater deviation from normal (all P50.001) for BRI, MCI and

50

| Brain 2009: 132; 4556

V. Anderson et al.

GEC. For both parent and teacher SDQ ratings, group differences
were signicant (P50.001).

Comparisons across AL groups

A. Comparison of group means
Intellectual ability
MANOVA detected a signicant multivariate effect, Wilks
K = 0.785, F(15,420) = 2.57, P50.001, 2 = 0.08. Univariate
analyses showed group differences for three measures: FSIQ:
F(5,154) = 5.11, P50.001, 2 = 0.14; VIQ: F(5,154) = 5.08,
P50.001, 2 = 0.14; and PIQ: F(5,154) = 5.77, P50.001,
2 = 0.16 (Table 5). For FSIQ, the Congenital group performed
more poorly than the two older AL groups (Middle

Childhood: P = 0.01; Late Childhood: P = 0.04). The Infancy and

Perinatal groups recorded lower scores than the Middle Childhood
group (P = 0.02 and 0.03, respectively). The Congenital group
achieved signicantly lower VIQ and PIQ scores than the three
older AL groups (Preschool: P = 0.04; Middle Childhood: P = 0.01;
Late Childhood: P = 0.02). For PIQ, the pattern was identical, with
the Congenital groups results poorer than the three older AL
groups (Preschool: P = 0.01; Middle Childhood: P50.001; Late
Childhood: P = 0.02). A signicant group difference was also
found between the Perinatal and Middle Childhood groups
(P = 0.03), in favour of the latter. These results indicate that
children with insults prior to preschool are at greater risk for
long-term intellectual problems than those sustaining later insults,
especially those with congenital lesions.

Table 4 Differences between clinical sample and test means for outcome domains
Measure

Variable

Test mean

Sample mean

SD

WASI

FSIQ
VIQ
PIQ
Reading
Spelling
Arithmetic
BRI
MCI
GEC
BRI
MCI
GEC
TOT
TOT

100
100
100
10
10
10
50
50
50
50
50
50
8.2
6.5

87.93
85.88
91.19
7.74
7.81
6.13
62.23
63.97
64.27
66.12
71.46
71.47
13.65
10.85

20.10
18.53
20.49
4.24
4.01
3.65
14.67
11.66
12.72
17.90
16.31
16.85
7.25
6.36

WRAT-3

BRIEF parent

BRIEF teacher

SDQ parent
SDQ teacher

t
7.60
9.64
5.44
6.74
6.90
13.44
10.14
14.57
13.65
10.77
15.74
15.23
9.48
8.32

df

159
159
159
159
159
159
147
147
147
142
142
142
158
147

50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001
50.001

Mean difference
12.08
14.12
8.81
2.26
2.19
3.86
12.23
13.97
14.27
16.12
21.46
21.47
5.45
4.35

Means for SDQ are Australian norms based on a sample of 717 years old. From website www.sdqinfo.com which sites this reference: Mellor, D. Normative data for
the Strengths and Difculties Questionnaire in Australia. Aust Psychol 2005; 40: 21522.

Table 5 Performances across groups

Intellectual ability
n
FSIQ Mean (SD)
VIQ Mean (SD)
PIQ Mean (SD)
Academic achievement
Reading Mean (SD)+
Spelling Mean (SD)+
Arithmetic Mean (SD)
Executive function
Parent ratings
BRI Mean (SD)
MCI Mean (SD)+
GEC Mean (SD)
Teacher ratings
BRI Mean (SD)
MCI Mean (SD)
GEC Mean (SD)


Congenital

Perinatal

Infancy

Preschool

Mid Childhood

Late Childhood

Total group

38
79.05 (16.10)
81.58 (16.70)
80.63 (18.68)

32
81.00 (18.40)
86.44 (19.34)
82.41 (18.27)

22
79.91 (17.53)
86.41 (20.66)
81.59 (19.73)

19
93.79 (13.67)
100.32 (20.62)
95.16 (16.54)

30
94.41 (19.99)
102.00 (18.11)
98.17 (20.22)

19
94.53 (17.08)
98.68 (21.74)
96.32 (19.99)

160
87.93 (20.10)
85.88 (18.53)
91.19 (20.49)

6.71 (4.61)
6.66 (4.09)
4.94 (3.16)

6.70 (4.07)
7.06 (4.09)
5.55 (3.96)

6.50 (4.75)
7.05 (4.12)
4.50 (3.66)

8.58 (3.36)
8.21 (3.87)
7.53 (2.88)

9.48 (3.85)
9.55 (3.68)
7.71 (3.55)

9.15 (3.48)
8.85 (3.42)
7.15 (3.31)

7.74 (4.24)
7.81 (4.01)
6.13 (3.65)

63.63 (12.75)
67.13 (11.89)
66.88 (12.09)

67.65 (18.44)
66.87 (12.46)
68.58 (14.90)

63.05 (11.52)
64.68 (7.56)
64.95 (8.44)

52.94 (12.14)
58.11 (12.31)
56.67 (12.14)

63.33 (13.74)
63.40 (10.83)
64.30 (11.71)

55.80 (13.17)
58.33 (12.16)
57.87 (12.42)

62.23 (14.67)
63.97 (11.66)
64.27 (12.72)

70.88 (21.81)
75.18 (16.08)
75.91 (18.07)

67.14 (18.33)
73.21 (18.82)
72.68 (19.15)

67.60 (14.80)
73.60 (13.62)
73.10 (13.09)

57.83 (10.47)
65.11 (14.11)
63.44 (12.80)

66.00 (16.99)
69.63 (15.01)
71.07 (15.22)

61.64 (18.50)
68.21 (19.40)
67.43 (19.76)

66.12 (17.90)
71.46 (16.31)
71.47 (16.85)

P50.001, P50.01, +P50.05.

Can age at insult predict outcome?

Brain 2009: 132; 4556

Academic ability
MANOVA detected a signicant multivariate effect, Wilks
K = 0.84, F(15,420) = 1.81, P50.03, 2 = 0.06. Univariate analyses
showed group differences for Reading: F(5,154) = 3.01, P = 0.013,
2 = 0.089, Spelling: F(5,154) = 2.56, P = 0.03, 2 = 0.08; and
Arithmetic: F(5,154) = 4.21, P = 0.001, 2 = 0.12. Despite the
signicant group difference for Reading, and the trend for earlier
AL groups (Congenital, Perinatal, Infancy) to perform more poorly,
comparisons were not signicant. The Congenital and Middle
Childhood groups differed signicantly, in favour of the older AL
group on tests of Spelling (P = 0.04) and Arithmetic (P = 0.02).
In addition, a signicant difference between the Infancy and
Middle Childhood groups was identied on Arithmetic (P = 0.01).
These ndings highlight a trend towards poorer academic abilities
in children with earlier lesions.

Executive abilities
(a) Parent ratings: For all three summary indices AL group means
were abnormally elevated (460) with the exception of the
Preschool and Late Childhood groups. MANOVA detected a signicant multivariate effect, Wilks K = 0.85, F(15,387) = 1.62,
P = 0.06, 2 = 0.06, with univariate analyses detecting consistent
differences: BRI: F(5,142) = 3.19, P = 0.01, 2 = 0.101; MCI:
F(5,142) = 2.63, P = 0.03, 2 = 0.09; and GEC: F(5,142) = 3.28,
P = 0.01, 2 = 0.10. For BRI and GEC, a single difference was
found between the Perinatal and Preschool groups (P = 0.01,
P = 0.02, respectively), with the Perinatal group more impaired.
For GEC, the difference between Congenital and Preschool
groups approached signicance (= 0.06); (b) teacher ratings:
MANOVA detected no differences, Wilks K = 0.91, F(15,373) =
0.89, P = 0.57, 2 = 0.03, and univariate analyses were also
non-signicant: BRI: F(5,137) = 1.49, P = 0.20, 2 = 0.05; MCI:
F(5,137) = 1.22, P = 0.30, 2 = 0.04; GEC: F(5,137) = 1.5, P = 0.18,
2 = 0.05.

Psychological status
MANOVA identied no signicant multivariate AL group effect,
for either parent, F(5,153) = 2.0, P = 0.08, 2 = 0.06 or teacher

| 51

ratings on the SDQ, F(5,142) = 1.22, P = 0.30, 2 = 0.04, despite

the observation that total group results were signicantly elevated
in comparison to normative expectations (Table 6). Univariate
analyses revealed a signicant group difference on the HYP subscale of the parent SDQ, F(5,153) = 2.70, P = 0.02, 2 = 0.08, with
Post hoc analyses revealing a signicant difference between
Perinatal and Late Childhood groups (P50.01), in favour of the
latter. For teacher ratings only, the ES subscale showed group
differences, F(5,142) = 2.52, P = 0.03, 2 = 0.08, with the
Congenital group more impaired than the Preschool
group (P = 0.04).

B. Analysis of frequency of impairments

Intellectual ability
Figure 1 illustrates the proportion of children in each AL group
falling within the normal, mildly impaired and severely impaired
ranges. For FSIQ, 2(5, 160) = 15.60, P50.01, V = 0.31, there
was a larger proportion of children with impairments in the
Congenital group (SR = 1.5), and a smaller proportion in the
Middle Childhood group (SR = 1.5) than expected. For VIQ,
2(5, 160) = 16.37, P50.01, V = 0.32, the proportion of children
with impairments from the Late Childhood group was smaller than
expected (SR = 1.7). For PIQ, 2(5, 160) = 22.56, P50.001,
V = 0.38, impairment rates also differed across groups, with
more children with impairments in the Congenital group
(SR = 2.1), and less in the Preschool (SR = 1.7) and Middle
Childhood groups (SR = 1.8).

Academic ability
Figure 2 illustrates the proportion of children in each group
within the normal, mildly impaired and severely impaired ranges.
Chi-square analysis for Reading, 2(5, 160) = 16.19, P50.01,
V = 0.32, revealed more children with impairments in the
Perinatal group (SR = 1.7), and less in the Middle Childhood
group (SR = 1.5) than expected. For Spelling, 2(5, 160) =
15.26, P50.01, V = 0.31, the proportion of children with impairments from both the Middle (SR = 1.8) and Late Childhood

Table 6 Psychological function across groups (parent and teacher reports)

SDQ results
Parent ratings
ES Mean (SD)
CS Mean (SD)
HYP Mean (SD)+
PP Mean (SD)
PSB Mean (SD)
TOT Mean (SD)
Teacher ratings
ES Mean (SD)+
CS Mean (SD)
HYP Mean (SD)
PP Mean (SD)
PSB Mean (SD)
TOT Mean (SD)

Congenital

Perinatal

Infancy

Preschool

Middle Childhood

Late Childhood

Total group

3.33
2.25
5.08
3.44
6.78
14.11

(2.32)
(2.13)
(2.84)
(2.22)
(2.36)
(7.41)

3.82
2.48
5.91
3.91
7.06
15.67

(2.53)
(1.91)
(2.87)
(2.80)
(2.61)
(8.26)

4.00
2.30
4.78
3.22
6.96
14.30

(2.99)
(2.16)
(2.28)
(1.83)
(1.85)
(6.57)

2.95
1.47
4.32
3.00
8.32
11.74

(2.68)
(2.17)
(2.61)
(1.94)
(1.77)
(7.29)

3.40
2.53
4.70
3.47
7.40
13.97

(2.34)
(1.87)
(2.65)
(1.98)
(2.01)
(5.71)

2.28
2.22
3.17
2.00
7.06
9.67

(2.05)
(2.29)
(2.46)
(2.17)
(2.39)
(6.98)

3.38
2.26
4.83
3.30
7.19
13.65

(2.50)
(2.06)
(2.74)
(2.26)
(2.24)
(7.25)

3.49
1.49
4.37
3.09
6.23
12.43

(2.05)
(2.11)
(2.59)
(2.38)
(2.81)
(7.62)

2.28
1.69
4.10
3.14
6.17
11.21

(1.94)
(2.14)
(2.58)
(2.49)
(2.93)
(6.67)

3.43
1.14
4.43
2.00
6.76
11.00

(2.46)
(1.74)
(1.99)
(2.19)
(1.92)
(5.60)

1.67
0.72
3.44
2.67
7.56
8.56

(1.37)
(0.90)
(2.57)
(2.00)
(1.98)
(4.53)

2.77
1.33
4.23
2.27
6.33
10.93

(2.53)
(1.81)
(2.49)
(2.10)
(2.82)
(6.18)

2.67
0.87
3.47
1.87
7.80
8.87

(1.80)
(1.69)
(2.85)
(1.55)
(1.57)
(5.36)

2.79
1.23
4.09
2.60
6.64
10.85

(1.16)
(1.85)
(2.50)
(2.23)
(2.56)
(6.36)


P50.001, P50.01, +P50.05.
CS = Conduct Symptoms Scale; HYP = Hyperactivity-Inattention Scale; PP = Peer Problems Scale; PSB = Prosocial Behaviour Scale; TOT = Total Difculties.

| Brain 2009: 132; 4556

Percentage impaired on Full Scale IQ across groups

Percentage impaired on WRAT-3 Reading across groups

Severely impaired

75%

75%

50%

50%

25%

25%

0%

0%

Normal

Mildly impaired

at
rin

ni

Pe

ge
on
C

*based on test norms

Percentage impaired on Verbal IQ across groups

ta

*
ed
ct
pe

Ex

To
ta
l

La

id
d

te

C
hi
ld
G
ro
up

C
hi
ld
le

sc

ho

ol

cy
Pr
e

In
fa
n

rin

on

Pe

ge

ni

at

ta

al

100%

Severely impaired

*based on test norms

Percentage impaired on WRAT-3 Spelling across groups

Normal

Severely impaired

100%

100%

75%

75%

50%

50%

25%

25%

Mildly impaired

Severely impaired

*
ed

up

ct

ro

hi

pe

lG

Ex

ta
To

id

La

dl

te

ch

hi

ld

ld

l
oo

y
nc

d*
ec
te

Ex
p

lG

ro

hi
ta
To

La

te

C
dl
e

up

ld

ld

Severely impaired

hi

ol

at
rin

ge
on
C

pe

Pe

ni
ta

*
ed
ct

ro
Ex

te

ta
To

La

e
dl
id
M

lG

hi
C

ch
es

up

hi
ld

ld

l
oo

y
nc
Pr

at
rin
Pe

In
fa

al

l
ta

ho

0%

id

0%

25%

25%

nc

50%

sc

50%

Mildly impaired

Pr
e

75%

75%

ni

es
Pr

Normal

Severely impaired

100%

ge

fa

at
Pe

Percentage impaired on WRAT-3-Arithmetic across groups

100%

In

rin

ni
ge
on

pe
Ex

fa

Mildly impaired

ta

*
ed
ct

ro
lG

To

La

ta

te

e
dl

Percentage impaired on Performance IQ across groups

Normal

*based on test norms

*based on test norms

al

up

ld
C

hi
C

ch
M

id

Pr

hi

ld

l
oo

y
nc

es

fa
In

rin

on

Pe

ge

ni

at

ta

al

al

0%

0%

on

Mildly impaired

nc
y
Pr
es
ch
oo
M
l
id
dl
e
C
hi
ld
La
te
C
hi
To
ld
ta
lG
ro
up
Ex
pe
ct
ed
*

Normal

100%

al

Mildly impaired

fa

Normal

In

V. Anderson et al.

In

52

*based on test norms

*based on test norms

Fig. 1 Impairment ratings for AL groups for IQ. (A) FSIQ, (B)
VIQ, (C) PIQ.

groups was smaller than expected (SR = 1.5). For Arithmetic,

2(5, 160) = 11.79, P = 0.04, V = 0.27, there was a larger
proportion of non-impaired children in the Preschool group
(SR = 1.9).

Executive abilities and psychological function

No signicant group differences were recorded on any of the
BRIEF or SDQ measures, as seen in Figs 3 and 4.

Fig. 2 Impairment ratings for AL groups for academic

achievement. (A) Reading, (B) Spelling, (C) Arithmetic.

Discussion
This study explored neurobehavioural and psychological impairment after EBI to determine the impact of such insults and if
developmental stage at insult had a differential inuence on outcome, in order to add to the plasticity-early vulnerability debate.
Children sustaining EBI during six different developmental periods,
from gestation to late childhood, were compared across a range of

Can age at insult predict outcome?

Brain 2009: 132; 4556

Percentage impaired on BRIEF-BRI across groups

Normal

Percentage impaired on SDQ (parent form) across groups

Clinical range

Normal

100%

| 53

Borderline

Abnormal

100%

75%

75%

50%

50%

25%

25%

0%

on
C

*based on test norms

ni
ta
l
Pe
rin
at
al
In
fa
nc
y
Pr
es
ch
oo
M
l
id
dl
e
C
hi
ld
La
te
C
hi
To
ld
ta
lG
ro
up
Ex
pe
ct
ed
*

ge

on
ge
ni
ta
l
Pe
rin
at
al
In
fa
nc
y
Pr
es
ch
oo
M
l
id
dl
e
C
hi
ld
La
te
C
hi
To
ld
ta
lG
ro
up
Ex
pe
ct
ed
*

0%

*based on test norms

Percentage impaired on BRIEF-MCI across groups

Normal

Percentage impaired on SDQ (teacher form) across groups

Clinical range

100%

Normal

Mildly impaired

Severely impaired

100%
75%

75%
50%

50%
25%

25%

Percentage impaired on BRIEF-GEC across groups

Normal

Clinical range

100%

75%

50%

25%

*
ct
pe

Ex

ro

ed

up

ld
C

ta
l
To

La

te

C
e
dl

id
M

hi

ld
hi

ol
ch
o
es

fa

nc
y
Pr

In

at
rin

Pe

on

ge

ni

ta

al

0%

*based on test norms

Fig. 3 Impairment ratings for AL groups for everyday attention

using the BRIEF. (A) BRI, (B) MCI, (C) GEC.

outcome domains. These AL groups were derived from animal

literature to correspond with documented periods of neural
growth, and were similar on key demographic (gender, SES) and
lesion variables (lesion size, extent, laterality). Assessments were
conducted during late childhood/adolescence, when recovery processes had subsided and maturational processes were largely
complete.
Comparisons between the total EBI sample and normative
expectations revealed signicant impairments across all domains

*
ed

up

ld

ct
pe
Ex

lG

ta
To

La

te

C
e
dl
id

ro

hi

ld
hi

oo

ch

Pr

es

fa

at
rin

Pe

nc

al

l
ta
C

on

ge

ni

pe
Ex

*based on test norms

In

0%

ct

ro

ed

up

ld
ta
To

La

lG

hi

te

C
e
dl

id

Pr

hi

ld

l
oo

ch

nc
In

es

fa

at
rin

on

Pe

ge

ni

ta

al

0%

*based on test norms

Fig. 4 Impairment ratings for AL groups for psychological

function using the SDQ. (A) Parent rating, (B) Teacher ratings.

under studyintelligence, academic ability, everyday executive

function and psychological function. Results conrm that children
with EBI are at increased risk of impairment when compared to
population expectations. The second aim of the study examined
the impact of EBI at different stages of development. As predicted,
results indicated that, depending on age at insult, outcomes differed signicantly. Patterns of impairment also varied across
domains, suggesting that different stages of brain development
may be critical for different functions.

Do children with EBI differ from

population expectations?
As previously noted, children with EBI, as a group, performed
consistently poorly for all domains studied, in keeping with
much previous research (Ewing-Cobbs et al., 1997; Jacobs et al.,
2007). Mean scores for the EBI group were not severely impaired,
but fell approximately 1 SD below expectations, representing performances hovering at the lower end of average. Further, for
cognitive and academic measures, between 50% and 60% of
children recorded scores within the normal range. The exception
to this pattern was Arithmetic, where 63.1% of the EBI sample
recorded impaired function.

54

| Brain 2009: 132; 4556

While it may be argued that the use of normative data rather

than an appropriately constructed healthy comparison group limits
the interpretability of our data, these results are consistent with
previous research with other populations (Kinsella et al., 1997;
Catroppa and Anderson, 2000), and indicate that arithmetic
skills are differentially vulnerable to brain injury, regardless of
when it occurs during childhood, perhaps because of the complex
range of skills recruited in these activities. In contrast, for everyday
executive skills and behaviour, mean scores were consistently
outside the normal range. For executive skills, parents rated their
children as experiencing difculties, with teacher ratings even
higher. For these measures, impairment rates were consistently
between 45% and 55%, signifying high risk and emphasizing
the need for formal evaluation to extend further than intellectual
and academic domains.
These results support previous research documenting
the detrimental effects of EBI (Anderson and Moore, 1995;
Anderson et al., 2005), and provide little evidence to corroborate
plasticity notions, which argue for good outcome from EBI.

Does age at insult inuence long-term

outcome?
Age at insult does impact on long-term outcome, although this
relationship may be more complex than expected. Firstly, with
respect to broad outcomes, children with earlier lesions were at
elevated risk for a number of disabilities, including developmental
delay and epilepsy, and with a trend to high risk for academic
difculties. Further, within the cognitive domain (intelligence, academic ability) group differences were substantial, with a dichotomy between EBI sustained before and after age 2, and with
younger AL associated with poorer outcome. This pattern was
consistent across group means and impairment ratings. In particular, children with Congenital and Perinatal insults performed uniformly very poorly, achieving signicantly lower scores than those
injured after age 7 years on all intellectual measures and spelling
and arithmetic. These results demonstrate that brain insult prior to
age 2 leads to poorest cognitive outcome, with later childhood
insult resulting in lesser impact.
For everyday executive function and behaviour the pattern was
somewhat different, and the discrepancy between earlier and
later AL groups was not as well dened. Once again, the Late
Childhood group appeared relatively intact, with Congenital and
Perinatal groups demonstrating signicant problems. In contrast,
the Middle Childhood group, with insults between 7 and 9 years,
was at greater risk, tending to function closer to those early lesion
groups. Children with Preschool insults (36 years) performed
best, and closest to normative expectations, while the Infant
group (1 month to 2 years) showed better outcome in executive
function.
These results support an early vulnerability perspective, with
children sustaining insults prior to and around the time of birth
being most at risk for global decits, while children with insults in
the second decade of life escape relatively unscathed. In contrast
to animal data, which suggests a non-linear relationship between
AL and outcome, our ndings illustrate a relatively linear pattern,

V. Anderson et al.
at least for cognitive and academic skills. For behaviour, where
animal researchers might argue skills are more complex, this relationship is more complex, and children with AL between 7 and
9 years have increased vulnerability, perhaps due to growth spurts
in frontal lobes during this period (Gogtay et al., 2004).
While early recovery issues have been addressed in the animal
literature (Kolb, 2005; Giza, 2006), there are concerns about
whether such data translates directly to humans, where brain
insults are less circumscribed and where cognitive abilities are
more complex. To date, human research has only contributed
partially to the eld, due to difculties in identifying children
with brain insults across development, and challenges controlling
for potential confounders such as insult severity, pathology volume
and environment. This study chose to address these previous
obstacles by recruiting children based on AL rather than the
traditional condition-based approach. In doing so, the resultant
sample necessarily included children for whom mechanism of insult
varied, creating the risk that ndings might reect differences
in brain pathology rather than AL. In order to minimize this risk,
we conned our recruitment to children with focal brain pathologies and collected detailed information on brain pathology (extent,
laterality, lesion size), allowing us to control for these potential
confounds. We believe that this approach has provided important
data to assist in understanding the impact of EBI from an empirical
perspective. Of note, we employed a categorical approach to
quantifying developmental stage. While these categories reect
CNS growth spurts, they are necessarily inexact and may mask
specic critical developmental periods. To extend these ndings,
prospective, multi-centre research facilitating larger sample sizes
is required.
Additionally, in this study we focussed our hypotheses on the
timing and characteristics of the structural lesion, with less emphasis on their neurological consequences. Thus, while some research
has demonstrated that presence of seizures has a negative inuence on development (Hartel et al., 2004; Chilosi et al., 2005;
Ballantyne et al., 2007), we chose to conceptualize presence of
seizures as a negative outcome of EBI, similar to speech delay
or motor impairment, in that it would restrict the childs capacity
to acquire skills and knowledge and function adequately within
his/her environment. Supplementary analyses, demonstrating
that presence of seizures is most frequent in earlier insults, suggests that early brain injury, together with seizures, may confer
added risk for the child, indicating that seizures should be seen
as a potential mediator of long-term function. However, as we
did not collect detailed data on age at seizure onset, frequency
and type of seizures, or medication, we are unable to examine the
specic relationships further.
A further limitation of previous literature is a failure to account
for age at testing. In the animal literature, researchers have shown
that even when keeping AL constant, different outcomes are seen
depending on the age at which outcome is assessed. In their rat
studies, Kolb and colleagues (Dallison and Kolb, 2003) showed
that, if they assessed function on a single post-natal day, recovery
seemed complete, however, if they evaluated animals on a later
post-natal day, their results were less positive. This pattern has
recently been noted in childhood stroke literature, where children
assessed at 5 years appear intact, but when tested several years

Can age at insult predict outcome?

later, clear impairments are observed (Bates et al., 2001; Stiles
et al., 2008).
Finally, while our results are consistent with increased vulnerability of the young brain, we did identify different patterns of
vulnerability, even using a limited range of outcome measures.
This suggests that inclusion of a broader range of outcome
domains may identify still more differences, reecting either
regions specic to brain development or, alternatively, critical periods for the emergence of particular behaviours.

Conclusions
Our study supports an early vulnerability model for EBI. Results
showed that children with EBI are at increased risk for impairment
in all domains assessed. Children sustaining EBI before age 2 years
recorded global and signicant cognitive decits, with pre- and
perinatal insult being particularly detrimental. Children with EBI
after age 2 functioned closer to normal expectations, suggesting
a roughly linear association between AL and outcome. In contrast,
within the behavioural domain, children with EBI from 7 to 9 years
performed worse than those with EBI from 3 to 6 years, and more
like those with younger insults, suggesting that not all functions
share the same pattern of vulnerability with respect to age at
insult. These ndings have important implications for clinical
practice, suggesting that children who sustain very EBIs will have
long-term impairments and require additional support and management across cognitive, academic and psychological domains.

Funding
National Health and Medical Research Council of Australia;
Australian Research Council.

References
Anderson V. Assessment of executive function in children. Neuropsychol
Rehab 1998; 8: 31949.
Anderson V, Bond L, Catroppa C, Grimwood K, Keir E, Nolan T.
Childhood bacterial meningitis: impact of age at illness and medical
complications on long-term outcome. J Int Neuropsychol Soc 1997; 3:
14758.
Anderson V, Catroppa C, Dudgeon P, Morse S, Haaritou F, Rosenfeld J.
Understanding predictors of functional recovery and outcome 30
months following early childhood head injury. Neuropsychology
2006; 20: 4357.
Anderson V, Catroppa C, Morse S, Haritou F, Rosenfeld J. Functional
plasticity or vulnerability following early brain injury? Pediatrics 2005;
116: 137482.
Anderson V, Jacobs R, Harvey AH. Prefrontal lesions and attentional
skills in childhood. J Int Neuropsychol Soc 2005; 11: 81731.
Anderson V, Moore C. Age at injury as a predictor of outcome following
pediatric head injury. Child Neuropsychol 1995; 1: 187202.
Anderson V, Morse S, Catroppa C, Haritou F, Rosenfeld J. Thirty-month
outcome from early childhood head injury: A prospective analysis of
neurobehavioral recovery. Brain 2004; 127: 260820.
Aram D. Language sequelae of unilateral brain lesions in children. In:
Plum F, editor. Language, communication and the brain. New York:
Raven Press; 1988. p. 17197.

Brain 2009: 132; 4556

| 55

Aram D, Eisele J. Intellectual stability in children with unilateral brain

lesions. Neuropsychologia 1994; 32: 8595.
Ballantyne A, Spilkin A, Hesselink J, Trauner D. Plasticity in the developing brain: intellectual, language and academic functions in children
with ischaemic perinatal stroke. Brain 2008 (September 2008, date
last accessed).
Ballantyne A, Spilkin A, Trauner D. Language outcome after perinatal
stroke: Does side matter? Child Neuropsychol 2007; 13: 494509.
Bates E, Reilly J, Wilfeck B, Donkers N, Opie M, Fenson J, et al.
Differential effects of unilateral lesions on language production in children and adults. Brain Lang 2001; 79: 22365.
Breslau N. Does brain dysfunction increase childrens vulnerability to
environmental stress. Arch Gen Psychiat 1990; 47: 1520.
Casey B, Giedd J, Thomas K. Structural and functional brain development and its relation to cognitive development. Biol Psychol 2000; 54:
24157.
Catroppa C, Anderson V. Recovery of educational skills following pediatric head injury. Ped Rehab 2000; 3: 16776.
Catroppa C, Anderson V. Outcome and predictors of functional recovery
ve years following pediatric traumatic brain injury. J Ped Psychol
2008; 23: 112.
Chilosi A, Pecini C, Cipriani F, Brovedani P, Brizzolara D, Ferretti G, et al.
Atypical language lateralization and early linguistic development in
children with focal brain lesions. DMCN 2005; 47: 72530.
Daallison A, Kolb B. Receovery from infant medial frontal cortical lesions
in rats is reversed by cortical lesions in adulthood. Behav Brain Res
2003; 146: 5763.
Daniel A. Power, privilege and prestige: occupations in Australia.
Melbourne: Longman-Cheshire; 1983.
Dennis M, Whitaker H. Language acquisition following hemidecortication: Linguistic superiority over the left versus right hemisphere. Brain
Lang 1976; 3: 40433.
Duchowny M, Jayakar P, Harvey AS, Altman N, Resnick T, Levin B.
Language cortex representation: effects of developmental versus
acquired pathology. Annals Neurol 1996; 40: 918.
Ewing-Cobbs L, Fletcher J, Levin H, Francis D, Davidson K, Miner M.
Longitudinal neuropsychological outcome in infants and preschoolers with traumatic brain injury. J Int Neuropsychol Soc 1997;
3: 58191.
Flavell JH. Cognitive development: Past, present, and future. Dev Psychol
1992; 28: 9981005.
Gioia G, Isquith P, Guy S, Kenworthy L. Behavior Rating Inventory of
Executive Function. Odessa, FL: Psychological Assessment Resources;
2000.
Giza C, Prins M. Is being plastic really fantastic? Mechanisms of altered
plasticity after developmental traumatic brain injury. Dev Neurosci
2006; 28: 36479.
Gogtay N, Giedd J, Lusk L, Hayashi K, Greenstein D, Vaituzis C, et al.
Dynamic mapping of human cortical development during childhood
and early adulthood. Proc Natl Acad Sci USA 2004; 101: 81749.
Goldma-Rakic P. Development of cortical circuitry and cognitive function. Child Dev 1987; 58: 60122.
Goodman R. The Strengths and Difculties Questionnaire: a research
note. J Child Psychol, Psychiat All Discip 1997; 38: 5816.
Gronwall D, Wrightson P, McGinn V. Effect of mild head injury during
the preschool years. J Int Neuropsychol Soc 1997; 3: 5927.
Hartel C, Schilling S, Sperner J, Thyen U. The clinical outcomes of neonatl and childhood stroke: Review of the literature and implications for
future research. Eur J Neurol 2004; 11: 4318.
Hebb D. The effects of early and late injury upon test scores, and
the nature of normal adult intelligence. Proc Amer Phil Soc 1942;
85: 27592.
Hertz-Pannier L, Chiron C, Jaambaque I, Renaux-Kieffer V, Van de
Moortele P, Delalande O, et al. Late plasticity for language in a
childs non-dominant hemisphere: A pre- and post-surgery study.
Brain 2002; 125: 36172.
Heywood C, Canavan A. Developmental neuropsychological correlates of
language. In: Yule W, Rutter M, editors. Language development and

56

| Brain 2009: 132; 4556

disorders: clinics in developmental medicine (No. 101102). London:

MacKeith/Blackwell; 1987. p. 14658.
Huttenlocher P, Dabholkar A. Developmental anatomy of prefrontal
cortex. In: Krasnegor N, Lyon G, Goldman-Rakic P, editors.
Development of the prefrontal cortex: Evolution, neurology, and
behavior. Baltimore, MA: Brookes; 1997. p. 6984.
Jacobs R, Anderson V. Planning and problem solving skills following
focal frontal brain lesions in childhood: analysis using the Tower of
London. Child Neuropsychol 2002; 8: 93106.
Jacobs R, Harvey AS, Anderson V. Executive function following focal
frontal lobe lesions: Impact of timing of lesion on outcome. Cortex
2007; 43: 792805.
Kennard M. Age and other factors in motor recovery from precentral
lesions in monkeys. Amer J Physiol 1936; 115: 13846.
Kennard M. Relation of age to motor impairment in man and in
subhuman primates. Arch Neurol Psychiat 1940; 44: 37797.
Kinsella G, Prior M, Sawyer M, Ong B, Murtagh D, Eisenmajer R, et al.
Predictors and indicators of academic outcome in children 2 years
following traumatic head injury. J Int Neuropsychol Soc 1997; 3:
608616.
Klingberg T, Vaidya C, Gabrieli J, Mosele M, Hedehus M. Myelination
and organization of the frontal white matter in children: a diffusion
tensor MRI study. Neuroreport 1999; 10: 281721.
Kolb B. Brain plasticity and behavior. Mahwah, NJ, USA: LEA; 1995.
Kolb B, Pellis S, Robinson T. Plasticit and functions of the orbital
frontal cortex. Brain Cog 2004; 55: 10415.
Leventer R, Phelan E, Coleman L, Kean M, Jackson G, Harvey AS.
Clinical and imaging features of cortical malformation in childhood.
Neurol 1999; 53: 71522.
Luciana M. Cognitive development in children born preterm: Implications
for theories of brain plasticity following early injury. Dev Psychopath
2003; 15: 101747.
Mosch S, Max J, Tranel D. A matched lesion analysis of childhood versus
adult-onset brain injury due to unilateral stroke. Cog Behav Neurol
2005; 18: 517.

V. Anderson et al.
Pavlovic J, Kaufmann F, Boltshauser E, Capone Mori A, Gubser
Mercati D, Haenggeli CA, et al. Neuropsychological problems after
paediatric stroke: Two year follow-up of Swiss children.
Neuropediatrics 2006; 37: 1319.
Piaget J. The origins of intelligence in children. New York: W.W. Norton;
1963.
Ponsford J, Willmott C, Rothwell A, Cameron P, Kelly A, Ayton G, et al.
Cognitive and behavioural outcome following mild traumatic brain
injury in children. J Int Neuropsychol Soc 1997; 3: 225.
Riva D, Cassaniga L. Late effect of unilateral brain lesions before and
after the rst year of life. Neuropsychologia 1986; 24: 4238.
Smith A, Sugar C. Development of above normal language and intelligence 21 years after left hemispherectomy. Neurology 1975; 25:
81318.
Stein D, Hoffman S. Concepts of CNS plasticity in the context of brain
damage and repair. J Head Trauma Rehab 2003; 18: 31741.
Stiles J, Stern C, Applelbaum M, Nass R, Trauner D, Hesselink J. Effects
of early focal brain insult on memory for visuospatial patterns:
Selective decits or global-local processing? Neuropsychology 2008;
22: 6173.
Taylor HG, Alden J. Age-related differences in outcomes following childhood brain insults: an introduction and overview. J Int Neuropsychol
Soc 1997; 3: 55567.
Taylor HG, Yeates K, Wade S, Drotar D, Stancin T, Minich N.
A prospective study of short- and long-term outcomes after traumatic
brain injury in children: Behavior and academic achievement.
Neuropsychology 2002; 16: 1527.
Vargha-Khadem F, Isaacs E, Papaleloudi H, Polkey C, Wilson J.
Development of intelligence and memory in children with hemiplegic
cerebral palsy. Brain 1992; 115: 31529.
Wechsler D. Wechsler abbreviated scale of intelligence. New York:
Psychological Corporation; 1999.
Wilkinson G. Wide range achievement test-revision 3. Wilmington:
Jastak Associates; 1993.