INTRODUCTION
The frontal lobes, particularly the prefrontal cortices, are critical
for normal development, because of their rich connections with other
cerebral regions and their central role in efficient executive function.
These structures develop rapidly through childhood and early adolescence,
paralleled by increases in “higher-order” executive abilities
such as attentional control, planning, and mental flexibility (Gogtay et al., 2004). Emerging research suggests that
damage to prefrontal regions during childhood may interrupt normal
maturational processes, leading to irreversible changes in brain structure
and organization, and associated impairments in neurobehavioral
development (Anderson et al., 2002). With
respect to attentional processes, there is also a suggestion from adult
literature that the laterality of prefrontal pathology may contribute to
the nature of the attentional impairment, with different patterns of
impairment described for left- versus right-frontal pathology
(Mecklinger et al., 1999; Stuss et al., 1999). However, whether attentional
skills can be similarly lateralized in childhood is not known, due to
incomplete understanding of normal developmental trajectories, limitations
in reliable and valid assessment options, and the relatively rare
occurrence of discrete prefrontal lesions during childhood.
The application of neuropsychological models of attention and
information-processing skills has provided some insight regarding links
between frontal pathology, lesion laterality, and neurobehavioral
function. One of the earliest neuropsychological models of attention was
that postulated by Alexander Luria (1973). He
proposed a posterior- and an anterior-attentional system working in
parallel in the mature brain, allowing the individual to monitor the
environment for events that might require a response (posterior system),
while pursuing various goals guided by intentional behaviors (anterior
system). Consistent with Luria's model, Posner (1988, 1995) hypothesized a
dual component model, with one aspect directed towards selective
attention, and shifts in spatial attention, and predominantly located in
the posterior cerebral cortex, in particular the parietal lobes. A second,
anterior element, incorporating the anterior cingulate gyrus and areas of
the prefrontal cortex, he argued represented higher-order function,
although with substantial neural links to the posterior system.
More recent attempts to describe attentional function have argued for
greater complexity, describing multicomponent systems with separate but
interacting attention modules (e.g., Dove et al.,
2000; Mirsky et al., 1991; Rubia et al., 1999; Shimamura,
1995; Stuss et al., 1999). Specific
details of these models remain controversial. In keeping with early
concepts, it is generally agreed that attention is subsumed by an
integrated neuroanatomical network. This network includes the brain stem,
aspects of the subcortex and posterior cortical regions, and prefrontal
cortex, with a critical role for the right hemisphere (Mirsky et al., 1991; Posner &
Petersen, 1990; Stuss et al., 1995; Stuss et al., 1999; Woods &
Knight, 1986). Definitions and operationalization of this
attentional system remain problematic, but a number of separate,
interdependent components are consistently identified: (i) sustained
attention, referring to the ability to maintain attention to a task
for prolonged periods; (ii) selective attention, the capacity to
attend to, and focus on, relevant stimuli, while filtering out extraneous
information; (iii) the ability to shift or switch attention,
moving flexibly from one concept or set to another; and (iv) divided
attention, the capacity to attend to competing stimuli
simultaneously. In keeping with its role as “manager” or
“central executive” of the brain, the prefrontal cortex is
argued to be critical to all aspects of the system, but primarily the
higher-order components of the attentional system, including shifting and
divided attention (Dove et al., 2000; Mesulam, 1981; Mirsky et al.,
1991; Shimamura, 1995; Stuss et al., 1995).
Clinical and functional imaging studies in adult samples have
identified impaired attention in association with frontal lobe pathology,
arguing for a particular role for the right prefrontal cortex. These
findings provide evidence that right hemisphere dysfunction is associated
with deficits in arousal, sustained attention, selective attention,
response speed, response inhibition, self-monitoring, and motor activation
(Heilman et al., 1991; Mesulam, 1981; Posner, 1988;
Robertson, 1999; Stuss et al., 1995, 1999; Woods & Knight, 1986). In contrast, for patients
with left frontal pathology, deficits are reported to be primarily
language-based, with some suggestion that skills in initiation and divided
attention may also be impaired (Godefroy et al.,
1996; Godefroy & Rousseaux, 1996;
Mecklinger et al., 1999; Smith & Jonides, 1999). Lesions to either
hemisphere may result in deficits in attentional shift and cognitive
flexibility (Grattan et al., 1994; Owen et al., 1991). Within the developmental
literature, recent imaging studies have identified a specific role for the
right frontal cortex for children with attention deficit/hyperactivity
disorder (Semrud-Clikeman et al., 2000; Sowell et al., 2003), suggesting that lateralization
of attentional skills may be present early in childhood.
To fully understand the potential impact of prefrontal pathology on
attentional skills in children, it is critical to have some understanding
of normal maturational processes in this domain. Without such
developmental markers, it is difficult to discriminate normal from deviant
levels of ability. Recent developmental research has provided such data,
identifying different developmental trajectories for specific attentional
components. For example, a number of studies document relatively early
development of basic selective attention skills, indicating rapid
maturation in infancy and early childhood, with adult level performances
demonstrated by children as young as 6 (McKay et al.,
1994; Manly et al., 2001; Rebok et al., 1997; Ruff &
Rothbart, 1996). Shifting and divided attention skills have been
found to progress slowly in early childhood, with more dramatic
development into adolescence (Anderson et al.,
2001a; Chelune & Baer, 1986; Manly et
al., 1999, 2001).
Processing speed, which underpins performance on many attentional tasks,
shows a gradual progression, with regular increments documented from age 6
to midadolescence (Anderson et al., 2001a; Kail,
1986, 1988). These
late developing skills may be particularly susceptible to the effects of
disruption from cerebral insult due to their protracted developmental
course (Dennis, 1989).
A number of studies have examined the neurobehavioral impact of very
early focal brain lesions occurring in either the pre- or perinatal period
(Ballantyne et al., 1992; Bates et al., 1999; Riva &
Cassaniga, 1986; Vargha-Khadem et al.,
1994), with the focus generally on language outcomes and
intellectual abilities. However, the long-term developmental implications
of focal prefrontal injury in childhood are as yet unknown, with the
incidence of such lesions relatively rare in childhood and adolescence.
Animal work investigating this developmental period provides some insights
into the possible long-term effects of these early frontal lesions.
Margaret Kennard, in her seminal work in the 1930s and 40s, described
relatively better functional outcome from early versus late focal
lesions in monkeys, but noted that some subtle deficits emerged over time
even following early lesions (Kennard, 1936,
1940, 1942). More
recently, Kolb and colleagues, also using animal models, have argued for a
complex, nonlinear relationship between age at injury and outcome. Their
work with rat models has identified a number of specific developmental
periods corresponding to better outcome, while damage during other
developmental periods may result in very poor outcome (Kolb et al., 1994, 1998, 2000). They argue that these “windows of
opportunity” for better outcome are linked to peak periods of neural
generation and synaptogenesis, which allow for changes in dendritic and
synaptic structure and connectivity (Kolb & Gibb,
1993; 2002; Kolb et al., 1994, 1996, 1997). The applicability of these results to humans is
questionable for a number of reasons, including the greater complexity of
the human brain, particularly the prefrontal cortex, differences in rates
of brain development across species, variations in lesion characteristics
and associated complications with human insults (e.g., seizures), as well
as the differential impact of environmental factors (Vargha-Khadem et al., 1985; Yeates
et al., 1997).
Evidence supporting the presence of attentional difficulties in the
context of early prefrontal pathology derives primarily from a small
number of human case studies. These cases are characterized by
disinhibition, reduced attention and executive deficits in association
with prefrontal pathology (Ackerly & Benton,
1948; Anderson, 1988; Anderson et al., 2000b; Eslinger
& Biddle, 2000; Eslinger et al., 1992, 1997, 1999; Marlowe, 1992). In
keeping with the developmental model of emerging deficits (Dennis, 1989), Eslinger and colleagues (1992) report a pattern of delayed onset of
impairments, with difficulties only identified over time as new skills
fail to “come on-line” and mature at critical stages
throughout development. In particular, these authors identify poor
development of attention and self-regulation as core deficits following
prefrontal lesions sustained in childhood.
To date, group-based studies of this population are rare. Even in the
adult literature, seminal studies have included participants with
extrafrontal injuries in addition to the more focal prefrontal pathology
of interest here (Stuss et al., 1999, 2000, 2001a, 2001b). Most research examining the impact of frontal
pathology in childhood has described samples of children with traumatic
brain injury where frontal lobe pathology often occurs in the presence of
more global cerebral pathology (Anderson &
Pentland, 1998; Dall'Oglio et al.,
1994; Ewing-Cobbs et al., 1998; Kelly, 2000; Levin et al.,
1997). Such samples, while valuable, do not enable the
fractionation of deficits specifically related to frontal pathology. In
one such study, Mateer (1990) examined a small
group of children who had sustained early cerebral insult. Her findings
documented intact or mildly depressed intellectual ability, despite
presence of frontal pathology, consistent with adult findings (Walsh,
1978, 1985). In
contrast, these children demonstrated symptoms of perseveration, reduced
attention, rigidity, and social difficulties. Several other studies (Anderson & Moore, 1995; Garth
et al., 1997; Levin et al., 1997; Pentland et al., 1998; Todd et al.,
1996), also employing samples of children with traumatic brain
injury, have reported similar findings.
Differential consequences of lateralized frontal pathology following
focal brain lesions have not been investigated within the pediatric age
range. While individual case studies have reported particular clinical
features associated with unilateral prefrontal pathology, such
descriptions have limited generalizability (Anderson et
al., 2000b; Eslinger & Biddle, 2000).
Some preliminary results, including those from our own research, suggest
that laterality effects in children may be somewhat different than those
observed in adults (Anderson et al., 2002; Bates et al., 1999; Goldberg &
Costa, 1981; Vargha-Khadem & Polkey,
1992). Specifically, with respect to prefrontal pathology, children
with right-sided insults have been found to perform more poorly than those
with left prefrontal lesions of similar severity on executive and
attention measures, regardless of the primary domain of the task. Such
results imply that the right prefrontal cortex plays a critical role in
the early development of basic attentional skills. On the basis of these
findings, and in keeping with the well-established theory of nonverbal
learning disability (Rourke, 1989), we have
postulated that, in early childhood, the right prefrontal cortex may play
a critical role in the mediation of executive functions, including
attention, with the left hemisphere being recruited for language tasks and
for complex tasks which require more global activation of frontal brain
regions in the immature brain. This hypothesis is consistent with data
from functional imaging studies that show more generalized frontal lobe
activation in adolescents, compared with adults, on inhibition tasks
(Casey et al., 1997; Tamm et
al., 2002).
From a functional perspective, attentional skills are of particular
significance during childhood, being critical for the development of
cognitive and neuropsychological systems, which in turn influence
adaptive, social and academic functioning (Cooley &
Morris, 1990; Dennis et al., 1995; Douglas, 1983). If attentional skills are impaired,
then children may be less able to learn and acquire skills from their
environment, to function independently in day-to-day life, and to make use
of teaching and instruction. Accurate mapping of a child's
attentional profile may enable the implementation of appropriate and
accurately targeted intervention.
The present study aimed to examine components of attention and
information processing in childhood and their cerebral lateralization
using a pediatric sample with documented prefrontal pathology. The study
is unique, in that no previous child-focused research has addressed this
domain in the context of focal prefrontal pathology, that is, without
additional significant and potentially contaminating extrafrontal cerebral
pathology. The study evaluated several aspects of
attention/processing, consistent with current theoretical models,
including selective attention, the ability to flexibly shift attention,
and divided attention. Speed of processing was also examined due to its
close links with efficient attentional function, and its potential
confounding effect on attention measures incorporating a timed component.
We predicted that, in keeping with adult models, children with prefrontal
lesions would exhibit deficits in shifting and divided attention, skills
thought to be subsumed by prefrontal regions. By contrast, we expected
selective attention skills, argued to be mediated by posterior systems, to
be relatively intact. We also considered the possibility of lateralization
of attentional processes. Based on a developmental framework in which the
right prefrontal region plays a key role in the growth of attentional
skills, our expectation was that children with right-frontal pathology
would demonstrate greater deficits in attentional function than those with
left-sided pathology.
METHOD
Participants
Between 1996 and 2002, children meeting study criteria were
ascertained via neuroscience clinics at the Royal Children's
Hospital, Melbourne, and the Sydney Children's Hospital, Australia.
Recruitment was consecutive for children meeting the following criteria:
(i) magnetic resonance imaging (MRI) evidence of focal frontal pathology,
as judged by a pediatric neuroradiologist and neurologist, with exclusion
of children with lesions confined to frontal motor cortex, extrafrontal
pathology, diffuse or progressive lesions; (ii) aged 7.0–16.11 years
at time of assessment; (iii) at least 6 months since diagnosis of frontal
pathology, to reduce confounding of outcomes with acute deficits and
treatment effects; (iv) English as the first language; (v) Full Scale IQ
≥ 70, based on the Wechsler Intelligence Scale for Children–Third
Edition (Wechsler, 1991); and (vi) absence of a
preexisting psychiatric condition, including attention
deficit/hyperactivity disorder (ADHD). A subset of the sample was
using anticonvulsant medication to control seizures.
All families of children in the clinical sample that were invited to
participate agreed to do so. The resultant clinical sample comprised 36
children (20 males), mean age 11.5 (SD = 2.8) years. All
participants had documented evidence of circumscribed frontal pathology,
involving prefrontal cortex, based on MRI scans. Timing of lesion varied,
with pathology resulting from either pre-natal abnormalities (n =
12), or peri/postnatal disorders (n = 24). Etiology of
frontal pathology was necessarily diverse, due to the low incidence of
such lesions in childhood. Prenatal abnormalities included focal cortical
dysplasia (n = 5), developmental tumours (n = 4), and
migrational disorders (n = 3). Peri-/postnatal pathologies
included stroke (n = 6), focal brain injury (n = 12),
space-occupying pathologies (n = 5), and acute disseminated
encephalomyelitis (n = 1). Extent of lesions varied across the
sample, with 18 participants having pathology circumscribed within
dorsolateral prefrontal cortex. Seven children had lesions extending to
both dorsolateral and orbitofrontal regions, two had lesions involving
dorsolateral and medial areas, and seven had lesions involving all three
areas. One child had a lesion specifically involving medial prefrontal
cortex, and another had pathology confined to the orbitofrontal region.
Examples of these lesions are provided in Figure
1.
MR scans of four participants, illustrating nature of prefrontal
pathology. Note radiological conventions apply to all scans). (a) 12-year
old female (ID = 33) who sustained a stroke causing focal insult to the
left prefrontal cortex six months prior to imaging. (b) 6 year old female
(ID = 23) who presented with a cerebral haemorrhage causing bilateral
prefrontal injury. (c) 12 year old male (ID = 13) with a developmental
tumor involving left dorsolateral prefrontal cortex, diagnosed following
onset of seizures at age 12 years. (d) 8 year old female (ID = 7) with
left hemisphere ganglioglioma involving the dorsolateral prefrontal
region, identified following emergence of seizures at age 8
years.
Laterality of lesion also varied with 15 children having left-sided
lesions, 10 right-sided lesions, and 11 bilateral lesions. Individual case
details of cerebral pathology, lesion timing, and age at symptom onset
(age at injury for acquired lesions, age at seizure onset for prenatal
lesions) are provided in Table 1.
Clinical and lesion data for children with prefrontal
lesions.
Forty healthy children (22 males), mean age 11.0 (SD: 2.8)
years, were recruited to match the demographic characteristics of the
prefrontal group (gender, age, and socioeconomic status or SES as measured
by the Daniel Scale of Occupational Prestige (1983). These children were ascertained via local
schools, where information outlining the study and letters of invitation
were sent to parents of children meeting the gender and age
characteristics required. Families agreeing to participate were considered
for involvement on a consecutive basis. Children in the control sample
were excluded from participation if they failed to meet criteria (ii),
(iv), (v) and (vi) as described above for the clinical sample. The
resulting controls group was similar to the prefrontal group on all
matching variables, with no significant differences. Mean SES was 4.2
(SD = 1.8) for the controls and 4.8 (SD = 1.3) for the
prefrontal group.
Screening measures
Medical and demographic variables: Data were obtained via
medical records (clinical participants) and from a demographic/medical
questionnaire completed by parents (all participants). The latter
questionnaire requested information on the child's medical and
developmental history, handedness, parental education and occupation, and
family constellation. Age at symptom onset, neurological symptoms and
seizures were noted. Seizures were coded into none, controlled by
medication, and uncontrolled. Socioeconomic status (SES) was
coded using Daniel's Scale of Occupational Prestige (1983) which rates parent occupation on a 7 point
scale, where a high score represents low SES.
Radiological data. All clinical participants underwent MRI
scan. Lesions were rated and coded by a pediatric radiologist and
neurologist, according to the following parameters: laterality
(left, right, bilateral), location
(lateral, medial, orbitofrontal); extent
(global: diffuse frontal pathology; multifocal: multiple frontal
lesions; and focal: single lesion site), and timing (pre-,
peri-, postnatal).
Intellectual evaluation. The Wechsler Intelligence Scale for
Children-III (WISC-III: Wechsler, 1991) was
administered to all children. Verbal, Performance and Full Scale
Intellectual Quotients were calculated.
Cognitive measures of attention
The following attention tests were chosen to measure selective,
shifting and divided attention, and to cover both verbal and spatial
domains, as summarized in Table 2.
Digit span. (Wechsler, 1991):
Children were asked to repeat strings of digits of increasing length, both
forwards and backwards. Scaled scores were employed in analyses. An
additional score was recorded for maximum number of digits forwards
repeated reliably (that is, on both trials).
Contingency Naming Test (CNT: Anderson
et al., 2000a; Taylor &
Alden, 1997). This test includes four subtests, of
increasing difficulty. The child is presented with a laminated card on
which are printed rows of shapes, of different colors. Within each
“outside” shape a second, “inside” shape, of the
same color, is drawn. Above some of the stimuli a reverse arrow is drawn.
For Subtest 1 the child is required to name the stimulus color, while for
the second subtest the aim is to name outside shapes. Subtests 1 and 2 act
as baseline measures, and tap selective attention skills. The third and
fourth subtests (CNT3, CNT4) involve a “shift” in attention.
For CNT3 the child is given two rules: if inside and outside shape are the
same, the child must name the stimulus color. If the shapes are different,
the correct response is the name of the outside shape. On the fourth
subtest, the child is instructed to follow the rules for CNT3, except when
a reverse arrow is above a stimulus, in which case the child is directed
to reverse the rules from CNT3 (i.e., where the shapes are the same, the
correct response is the shape of the stimuli). Time taken to completion
and efficiency scores (incorporating speed and accuracy measures) were
recorded.
Trail Making Test (TMT: Reitan &
Davison, 1974). In this test children are required to
draw lines connecting consecutive numbers (Part A) or numbers alternating
with letters (Part B). Time to completion and number of errors were
recorded. A ratio score was also employed (Trails B: time to
completion–Trails A: time to completion)/Trails A: time to
completion) as an index of shifting attention.
Sky Search (SS: TEA-Ch: Manly et al.,
1999). Children were given a laminated sheet
depicting rows of spacecraft and instructed to find all targets, indicated
by two of the same ships within a pair, as quickly as possible. To control
for differences attributable to motor speed rather than visual selection,
children completed a motor control task, identical to that of the Sky
Search test with the exception that all of the distractor items were
removed. For each part of the test, an “attention score” was
calculated (targets found/time). Subtraction of the
“motor” time-per-item from the more attentionally demanding
condition time-per-item produced an “attention” score that was
relatively free from the influence of motor speed. Time to completion for
Sky Search Motor and the attention score were also employed in
analyses.
Creature Counting (CC: TEA-Ch: Manly
et al., 1999). For each item, a variable number of
“creatures” were depicted in their burrow. Interspersed
between the creatures were arrows either pointing up or down. Children
were asked to begin counting creatures from the top left hand corner of
the page, and through the “burrows,” but to use the arrows as
a cue to switch the direction of their count. The accuracy of the response
and the time taken to complete the page were recorded. A timing score was
calculated (seconds-per-switch) by dividing the time taken to complete
correct items by the number of switches within those items.
Score Dual Task (Score DT: TEA-Ch: Manly et al., 1999). This task was
divided into a series of 10 items or “games” in which the
child was asked to listen to a tape, and to count a series of tones, while
listening to a news broadcast. The tone counting aspect of the task
required children to silently count tones separated by silent
interstimulus intervals of variable duration (between 500 and 5000 ms),
and to give the total at the end. In addition, meaningful, auditory
speech—in the form of news bulletins—was simultaneously
presented. Children were asked to keep a count of tones whilst listening
for the mention of an animal during the news broadcast. Scores used in
analyses included number of games correct (i.e., the number of items where
the child correctly counted number of tones presented), number of animals
correctly identified and a score combining accuracy for both measures.
Sky Search Dual Task (SSDT: TEA-Ch: Manly et al., 1999). Children were
asked to complete a parallel version of the visual search stimuli used in
Sky Search. As they did this they were asked to count the number of tones
presented within each item or game of a counting task, as described in
Score Dual task, above. Tones were presented at 1-second intervals. The
test was ended when the child completed the visual search component. Time
taken to find each visual target was calculated, together with the
percentage of the counting items correct.
Behavioral evaluation
In addition to tests of attention, parents completed the Behavioral
Rating Inventory of Executive Function (BRIEF: Gioia et
al., 2000) to assess day-to-day executive and attentional function.
The BRIEF yields T scores (mean 50, SD = 10), with scores above 65
representing significant impairment. Scores employed in analyses included:
(i) Global Executive Composite (GEC): total score; (ii) Inhibit Scale:
assesses inhibitory control and the ability to resist the temptation to
act on impulse; (iii) Shift scale: evaluates the ability to move freely
from one situation, activity or aspect of a problem to another according
to task demands or environmental circumstances; and (iv) Monitor scale:
taps the child's ability to check his or her own work for errors and
to monitor the effect the child's own behaviour may have on
others.
Procedure
Potential clinical participants were identified via neuroscience
clinics, based on the criteria described above. For children with
postnatal conditions, families were contacted at least six months
post-event, to allow for acute recovery to take place. For all others,
families were contacted after an initial chart review. According to
requirements of the Royal Children's Hospital, Melbourne, and Sydney
Children's Hospital Human Ethics Committees, letters describing the
study, together with a brief questionnaire seeking information on the
child's developmental, neurologic, and psychiatric history, parent
occupation and ethnicity, were distributed to families. Participants were
children for whom signed consent was obtained and who met inclusion
criteria. Children were then assessed at the Australian Centre for Child
Neuropsychology Studies.
Healthy control children were identified from local schools, in
accordance with education department ethics procedures. Children were
chosen from class lists, to match clinical participants as closely as
possible with respect to age, gender, and SES. Information on the study
and the background questionnaire were sent to families via the school.
Children were assessed within the school context once signed consent was
obtained, and inclusion criteria checked.
Assessments were conducted on an individual basis, and by a qualified
child psychologist. All tests were administered in fixed order. Assessment
duration was approximately 2 hours, divided into two sessions by a short
break. Intellectual testing was completed in the first session and
attentional measures in the second.
Statistical analyses
To examine between-group differences, the attention measures were
subjected to analysis of covariance (ANCOVA), covarying for age, gender,
and SES. IQ was not employed as a covariate, as the development of IQ is
likely to be intimately linked to deficient executive functions in
children (Hebb, 1947; Williams & Mateer, 1992).
Although preliminary analysis failed to reveal outliers, children who
were unable to complete a task were assigned the lowest (age-normed) score
possible for a given measure. These scores were applied to individual
scores for a small number of children (11% of prefrontal sample and 2.5%
of control sample). Two sets of ANCOVA were carried out, one set comparing
the total prefrontal group with controls, and the second comparing the
left prefrontal, right prefrontal, bilateral prefrontal, and control
groups. As comparisons were considered exploratory in nature, a p
level of .05 was adopted for these analyses. Post-hoc analyses of simple
effects were conducted using Bonferroni adjusted p levels. Effect
size was computed using partial eta-squared.
A series of correlation analyses was conducted, to examine the
potential relationship between a number of specific factors and
attentional function. These included timing of lesion
(pre-/postnatal), age at symptom onset and presence of seizure
activity.
RESULTS
IQ and Demographic Characteristics
Mean FSIQ (SDs) for the prefrontal and control groups were
89.7 (10.8) and 104.6 (13.4), respectively, F(1,75) = 27. 7,
p < .001. Children with prefrontal lesions thus performed more
poorly than controls, with the mean score for the prefrontal group at the
lower end of the average range. This finding is contrary to adult-based
evidence, which has argued that individuals sustaining frontal lobe damage
present with intact IQ scores.
The lesion laterality subgroups did not differ from one another with
respect to age, gender, VIQ, PIQ, or FSIQ (see Table
3). No group differences were identified for presence of seizures,
onset of seizures, or age at symptom onset. There were no differences
between left and right lesion subgroups for extent and location of lesion
or for timing of lesion. There was a trend for a greater proportion of the
bilateral subgroup to be peri- or post-natal in origin, and for pathology
in this subgroup to extend across multiple prefrontal regions (e.g.,
lateral, medial, orbitofrontal).
Characteristics of frontal lesion group according to lesion
laterality
Cognitive measures of attention
Comparisons of prefrontal and control groups. As summarized
in Table 4, children with prefrontal lesions
performed significantly worse on measures of shifting and divided
attention. Reduced scores were evident on tasks involving both verbal
(CNT, Score) and nonverbal (TMT, CC, SS) domains, and for accuracy (TMT: B
errors, CC: number correct, Sky Search DT: percent games correct, Score
DT: animals correct, games correct, total) and speed measures (CNT 3 &
4: efficiency). Processing speed was also slower within the prefrontal
sample, with significant group effects identified for SS Motor: time and
CNT trial 1: time.
Mean adjusted scores, F, p values and partial
eta-squared values for attention measures across pre-frontal and control
groups
For selective attention, group differences were identified for Digit
Span (total subtest score and digits forwards) only, with the prefrontal
groups performances not significantly different from healthy controls for
CNT: trial 1: efficiency, CNT: trial 2: efficiency, TMT: A: time and
errors and SS attention.
Comparisons of left prefrontal, right prefrontal, bilateral
frontal and control groups. To evaluate the effect of lesion
lateralization, the clinical group was divided into children with left
(n = 15), right (n = 10) and bilateral (n = 11)
prefrontal lesions. Data pertaining to these three groups are provided in
Table 5. Group comparisons revealed surprisingly
few differences. The bilateral prefrontal group performed significantly
more poorly than the control group on the Digit Span test (Digit span,
p = .003; Digits forwards, p = .01). For shifting
attention, the bilateral prefrontal group performed significantly more
poorly than controls on trials 3 and 4 of the CNT (CNT3 efficiency,
p = .002, CNT4 efficiency, p = .004), Creature Counting
number of games correct (p = .003). On measures of divided
attention, the bilateral prefrontal group identified fewer animals on
Score DT than both the right (p = .003) and control (p
< .001) groups and achieved fewer “score” games correct
than controls (p < .001), and consistent with this
performance, obtained a lower total score compared with the control group
(p < .001) on this task. The bilateral prefrontal group also
performed more poorly than controls on the auditory/verbal component
of Sky Search DT, obtaining fewer “score” games correct on
this task than controls (p = .030).
Adjusted means, F and p values across prefrontal
group according to lesion laterality (significant clinical group and
control comparisons are also indicated)
For the left lesion group, deficits were only found on
auditory/verbal aspects of divided attention tasks. Compared with
controls, the left lesion group achieved fewer ‘score’ games
correct on Score DT (p = .05), and had a significantly lower
total score for this task (p = .020).
Impairments detected for the right lesion group were not consistent
across attentional domains, perhaps reflecting the role of the right
frontal lobe in sustained attention, online monitoring and self-regulation
(Dobler et al., 2003). This group achieved a
significantly lower Sky Search Attention score compared with both the
control and left lesion groups (p = .010, p = .030,
respectively), and were slower than the control group to complete the Sky
Search Motor task (p = .060). Within the domain of divided
attention, there was a trend for the right lesion group to obtain a lower
score than controls with respect to the number of “score”
games correctly identified on Score DT (p = .060). Of note, the
right prefrontal lesion group was the only group to have significantly
elevated scores on all subscales of the BRIEF pertaining to attention,
achieving a lower Global Executive Composite compared with controls (GEC:
p < .001) as well as raised scores on three subscales: Inhibit
(p = .010), Shift (p = .020) and Monitor (p =
.003).
Behavioral Indices of Attention (BRIEF)
Parent ratings of children's attentional skills as derived from
the BRIEF indicate consistently greater difficulties for children with
prefrontal pathology. Children with prefrontal lesions recorded
significantly higher Global Executive Composite scores than healthy
controls, F(1,75) = 27.7, p < .001. Means (SDs) for
the prefrontal and control groups, respectively, were 64.0 (12.7) and 51.2
(10.0). Further, Inhibit, Shift, and Monitor subscales (F(1,75) =
12.6, p = .001, F(1,75) = 4.5, p = .040;
F(1,75) = 14.4, p < .001, respectively) were all
significantly elevated for the clinical group. BRIEF ratings were highest
for the right prefrontal group, with the mean GEC score significantly
different from the control mean (p < .001) and falling within
the clinically impaired ranged (Adj M = 68.5, SE = 3.5). This group
obtained significantly higher subscale scores than controls for Inhibit
(p = .001), Monitor (p = .002), and Shift subscales
(p = .006).
Correlations of Lesion Characteristics with Attention
Measures
Correlations between predictor variables and attention variables were
also calculated. Results indicated strong correlations between shifting
and divided attention measures and age at symptom onset, with younger age
at onset predicting poorer performance. Specifically, age at symptom onset
was negatively correlated with Sky Search: attention (r =
−.606, p < .01) and Sky Search DT: time/target
(r = −.390, p < .05), suggesting that children
with younger age at symptom onset took longer to complete these tasks. For
Creature Counting: number correct (r = .368, p < .05)
and Score DT: total (r = .405, p < .05) younger age
at symptom onset was associated with greater performance accuracy. No
correlations were found for seizure severity or lesion timing.
DISCUSSION
Children with lesions involving prefrontal cortex exhibit attentional
impairments when compared with healthy age and gender matched controls.
These impairments are evident on both psychometric measures and for parent
ratings of day-to-day function. On both cognitive and behavioral measures
of attention, deficits were detected in all components investigated.
Further, the effects of lesion laterality suggested few deviations from
expectations for children with left prefrontal lesions, with reduced
scores evident only on the auditory-verbal task of divided attention.
Children with right prefrontal lesions also exhibited specific deficits,
demonstrating difficulties on measures of visual selective attention.
However this group was characterized by most impaired day-to-day function.
Not surprisingly, children with bilateral lesions demonstrated a trend to
poorest scores across a range of attention domains, with greatest
impairments in shifting and divided attention, possibly indicating the
greater cognitive resources required for these tasks. Results provide
evidence of the importance of prefrontal cortex in attentional function
during childhood, and indicate that the lateralization of attention
deficits observed following early prefrontal lesions may not be identical
to those described for adults.
Comparisons between controls and children with prefrontal pathology
identified the expected discrepancies in attention. Children with
prefrontal lesions performed poorly on “executive” aspects of
attention, including shifting and divided attention. For divided attention
tasks, scores recorded for the clinical sample were greater than one
standard deviation below controls for all measures, reflecting a highly
significant discrepancy. For shifting attention measures, clinically
relevant group differences were also evident. These findings suggest that
children with prefrontal lesions experience substantial difficulties with
cognitively demanding activities such as dividing attention across two
tasks simultaneously, or shifting flexibly from one dimension to
another.
In keeping with attentional models (Mirsky et al.,
1991; Posner, 1995), group means for more
“posterior” based, selective attention functions were closer
to developmental expectations. With the exception of Digit Span and CNT1:
efficiency, performances on selective attention measures were comparable
across groups, suggesting no substantive impairments in association with
prefrontal pathology. Findings for the Digit Span task are worthy of
further consideration. While this measure is commonly listed as tapping
selective attention (Mirsky, 1996; Mirsky et al., 1991), it may be argued that such a
classification is problematic. Specifically, due to the inclusion of a
“digits backwards” component, the resultant score may be
better interpreted as a working memory index (Baddeley,
1990). Poorer performances may thus reflect impairments in higher
order attention processing, rather than selective attention deficits.
Speed of processing was also slowed in children with prefrontal lesions,
in keeping with adult literature which links slowed processing to frontal
impairments. The magnitude of deficits was less than those identified for
higher level processing tasks, but was present for both speech and motor
responses.
These results are consistent with expectations from adult-based
studies, which suggest that the anterior regions of the brain are
responsible for shifting and divided attention, and aspects of processing
speed, reflecting the high level of attentional resources required to
perform effectively on such tasks. Recently, adult studies have examined
the impact of frontal lobe pathology in more detail, suggesting that there
may be some localization or lateralization of function within the
prefrontal cortex (Mecklinger et al., 1999;
Stuss et al., 1999). To date, pediatric research
has not addressed this possibility for the less differentiated, developing
brain, with published data restricted to case studies, and more global
insults (Bates et al., 1999; Eslinger et al., 1999; Vargha-Khadem et al., 1994). The present study
provided an opportunity to investigate the possibility of lateralization
of “prefrontal” attentional functions in children.
Interestingly, differences across the laterality groups were few,
consistent with earlier reports comparing verbal and performances IQ in
children with congenital, diffuse, unilateral lesions (Bates et al., 1999; Vargha-Khadem
et al., 1994). Children with left hemisphere pathology performed
closest to expectations across attention domains, with no problems
identified for visually based measures or processing speed. Deviations
from controls were identified only for auditory-verbal divided attention,
which requires on-line attention skills and verbal working memory.
Children with right prefrontal pathology also demonstrated relatively
circumscribed attentional impairments. Within the selective attention
domain, poorer performances were recorded within the visual modality only.
For shifting and divided attention measures, most results were similar to
controls, although there was a tendency for the group to demonstrate lower
accuracy on more demanding tasks, consistent with suggestions of greater
impulsivity and poorer attentional control associated with right
prefrontal pathology (Eslinger et al., 1997,
1999). This possibility is supported by this
group's elevated scores for parent ratings of behavior, with
impairments noted on Monitor, Inhibit, and Shift subscales, and reflecting
difficulties with attentional control.
Not surprisingly, children with bilateral prefrontal pathology also
demonstrated impairments in many aspects of attention. With the exception
of Digit Span, and consistent with suggestions that selective attention is
mediated by posterior brain regions, this group exhibited relatively
intact selective attention. Similarly, processing speed was intact. In
contrast, this group achieved poorer results for the more demanding
shifting and divided attention tasks, with scores falling well below
controls. In keeping with adult models (Stuss et al.,
1999; Walsh, 1978), as well as with
developmental perspectives (Anderson et al.,
2001a; Kolb & Gibb, 1993; Taylor & Alden, 1997; Vargha-Khadem et al., 1994), such findings suggest
that more global prefrontal pathology may result in difficulties coping
with resource-intense cognitive tasks, regardless of the attentional
component under investigation. With respect to day-to-day behavior,
children with bilateral prefrontal lesions were similar to controls, with
mean scores within the average range.
The possibility that the results of the bilateral lesion group might
reflect an additive or cumulative effect of both left and right prefrontal
performance characteristics was considered, but results indicate that such
an explanation may be too simplistic, given that this group did not
demonstrate impairments on behavioral indices nor on several measures
which were deficient for either the left or right prefrontal groups.
Further, there were a number of areas where only the bilateral prefrontal
group recorded deficient scores. These latter uniquely poor results are
specific to the higher-order shifting and divided attention domains,
suggesting that serious deficits in these aspects of attention in children
may only occur when both prefrontal cortices are dysfunctional. One
explanation for this pattern of results may be that, where damage is
present in both prefrontal regions, the impact is similar to that seen for
early generalized brain insults, such as head injury, where there is
limited potential for “reorganization” or
“recruitment” within healthy brain tissue in either the
ipsilateral or contralateral hemisphere (Anderson et al., 1997, 2001b). The somewhat
unexpected finding that children with bilateral prefrontal lesions did not
demonstrate behavioral problems in the attention domain needs some further
consideration. Comparison of the three clinical groups identified no
differences with respect to demographic, and no differences in lesion
characteristics, with a similar distribution of prefrontal pathology
across groups. One potential explanation may be that, while focal right
prefrontal pathology reduces attentional control, concurrent left
prefrontal pathology may act to dampen these symptoms, or alternatively,
given their more generalized cognitive impairments, parent ratings may
reflect lower expectations of these children.
Finally, while attention was the focus of this study, the intellectual
characteristics of the sample were somewhat unexpected. Children
sustaining prefrontal lesions demonstrated poorer overall intellectual
ability than controls, in contrast to adult findings, which suggest that
IQ is not impacted by frontal lesions (e.g. Walsh, 1978, 1985), but consistent
with child-based case reports (Eslinger & Biddle,
2000; Hebb, 1947; Williams & Mateer, 1992). However, such a
difference between adult and child findings makes intuitive sense. As has
been documented (Lezak, 1995), standardized
intellectual tests primarily assess well-learned skills. For adults
sustaining prefrontal lesions, these well-learned skills are largely
established, and less vulnerable to executive dysfunction (Anderson, 1998; Anderson et al.,
2001b). For children, acquisition of intellectual abilities is
dependent upon a logical, efficient approach to the registration,
organization and storage of new information and skills. Thus, prefrontal
damage (and associated executive dysfunction) at an early age may impede
the development of cognitive abilities in a global manner, rather than
cause specific “executive” symptoms seen in adults (Anderson, 1998).
Comparison of children with left and right prefrontal lesions detected
the predicted modality-specific pattern on attentional tasks, although
this relationship was not so clear for IQ measures. No group differences
were identified for VIQ, with both left and right lesion groups
demonstrating mean scores at the lower end of the average range. Further,
children with right-sided lesions recorded a marginally lower score on
this measure (right frontal lesion: VIQ = 91.8; left frontal lesion: VIQ=
89.2). For nonverbal skills, group differences were in the expected
direction, but remained nonsignificant. These results are consistent with
findings from other studies, which have identified global reduction in
intellectual capacity following unilateral lesions (not confined to the
frontal lobes) in the first 12 months of life. In contrast, these studies
have not detected such a significant impact on IQ for those children
sustaining similar lesions after 12 months of age (Aram
& Enkelman, 1986; Ballantyne et al., 1992: Bates et al., 1999;
Riva & Cassaniga, 1986; Vargha-Khadem et al., 1985). It may be that prefrontal
lesions, and associated higher-order cognitive impairments, have a
dramatic impact on development processes, regardless of timing.
The findings described above need to be interpreted in the context of
some methodological considerations. First, sample size, while larger than
previous studies, is relatively small, due to the low incidence of
prefrontal lesions in children. While a number of potentially confounding
variables were statistically controlled, the contribution of factors such
as lesion size, and seizure severity and medications, age at injury,
localization of cerebral pathology within the frontal lobes, and
socioeconomic status cannot be adequately canvassed with such a small
sample. In a similar vein, despite recruiting a sample with lesions
largely confined to the prefrontal lobes, the study necessarily comprises
a variety of brain conditions, which restricts interpretation across
disorder groups. Further, the limitations in sensitivity inherent in
current MR technology restrict the capacity of such methods to identify
subtle cerebral pathology. These issues may be argued to reduce the degree
of specificity that can be achieved in addressing the contribution of the
prefrontal cortex to particular aspects of cognition and behavior. Of
note, seminal studies in the adult literature have been able to clearly
demonstrate highly specific deficits linked to particular prefrontal
regions, despite including participants with similarly diverse brain
pathology (Stuss et al., 1999, 2000, 2001a, 2001b). We would argue that the lack of specificity in
our findings may not be simply explained by the nature of the sample.
Further research is required to address these outstanding areas.
In conclusion, children sustaining prefrontal lesions demonstrate
compromised intellectual and attentional function. In keeping with current
neuropsychological models of attention, when compared with healthy
controls, impairments are greatest for high-level attentional skills,
purported to be subsumed by anterior systems, including shifting and
divided attention. Selective attention skills are closer to normative
expectations. Behavioral correlates of these attentional deficits are also
present. Within the prefrontal group, subtle laterality differences in
attention were identified, consistent with adult findings. Left prefrontal
pathology was associated with relative resilience, with impairments
restricted to auditory-verbal, divided attention skills, possibly
suggesting difficulties in “on-line” processing capacities.
Behavioral ratings indicated few discrepancies from healthy controls. In
contrast, right prefrontal pathology was characterized by deficits in
spatial aspects of attention as well as high error rates and greater
impulsivity. Finally, bilateral prefrontal lesions were linked to severe
deficits in higher-order shifting and divided attention skills, with other
aspects of attention closer to normal expectations.
While some differences in performance are evident between left- and
right-prefrontal groups, in keeping with adult patterns, the integrity of
both regions appears to be required for developmentally appropriate
attentional in childhood. Such a notion is consistent with neuroimaging
studies in healthy children demonstrating the recruitment of more
prefrontal regions to support high-order cognitive processes in childhood
and early adolescence when compared with adults (Casey
et al., 1997; Tamm et al., 2002). The
implications of such deficits for both cognitive function and everyday
life are enormous, and suggest that children with frontal lobe pathology
will have difficulty coping efficiently with a range of day-to-day
activities, perhaps particularly those requiring greater cognitive
resources, more “on-line” processing, efficient
self-monitoring or cognitive flexibility.
ACKNOWLEDGMENTS
We would like to thank Dr. Anna Mandalis for her support in
recruitment of participants, and the children and families who gave of
their time to contribute to the study.
