Neurocognitive Outcomes and Recovery After Pediatric TBI: Meta-Analytic Review of the Literature

Neuropsychology
© 2009 American Psychological Association
2009, Vol. 23, No. 3, 283–296
0894-4105/09/$12.00
DOI:10.1037/a0015268
Neurocognitive Outcomes and Recovery After Pediatric TBI:
Meta-Analytic Review of the Literature
Talin Babikian and Robert Asarnow
David Geffen School of Medicine at UCLA
Traumatic Brain Injury (TBI) continues to be one of the leading causes of death and disability in the
pediatric population. Although the literature on neurocognitive outcomes is relatively rich, studies vary
signi?cantly in the methods used to group subjects on several moderating variables, including age at
injury, injury severity, and time since injury, making it dif?cult to combine and summarize the data for
comparison. Further complicating this effort is the wide range of measures used to document functional
outcomes in key neurocognitive domains. In this meta-analytic review, 28 publications (1988 to 2007)
that met inclusion criteria were summarized based on three distinct injury severity and time post injury
groups for 14 key neurocognitive domains. Effect sizes were calculated to re?ect the extent to which the
above groups differed in case-control and case-case studies, as well as address recovery based on
longitudinal studies. To the best of our knowledge, this is the ?rst published quantitative summary of the
literature on neurocognitive outcomes after pediatric TBI. Limitations of the current state of the literature
as well as recommendations for future studies are discussed.
Keywords: traumatic brain injury, pediatric, meta analysis, neurocognitive outcomes
Traumatic Brain Injury (TBI) is the single most common cause
studies and the heterogeneity of injuries need to be systematically
of death and disability in children and adolescents (CDC, 2000). A
taken into account when summarizing the available data on neu-
major cause of the signi?cant disability frequently associated with
rocognitive outcomes after pediatric TBI.
pediatric TBI is acquired neurocognitive impairments that ad-
We conducted a meta-analytic review of studies that examined
versely affect academic, behavioral, and interpersonal functioning.
the effects of injury severity and time since injury on neurocog-
Numerous studies have examined the neurocognitive sequelae
nitive outcomes and recovery after pediatric TBI using cross-
after a brain injury in childhood. There are often striking incon-
sectional and longitudinal reports. By de?nition, a meta-analysis
sistencies among studies about the nature and extent of neurocog-
uses effect size as the metric to delineate the magnitude of the
nitive impairment and recovery following TBI. These inconsisten-
difference between groups. In this report, the analyses were de-
cies in part result from several intrinsic characteristics of TBI,
signed to answer the following three separate, but related, ques-
including the heterogeneity of injuries (e.g., location of lesion,
tions regarding outcomes and recovery of key neurocognitive
mechanism of injury) and the in?uence of social and developmen-
domains after a pediatric TBI:
tal processes (e.g., age at injury, premorbid status, and family
1.
What is the magnitude of the effect of injury severity
resources) on the injured brain’s capacity to recover neurocogni-
over the course of three time bands following injury?
tive functioning. The variability in reported neurocognitive out-
(Case-control studies)
comes among studies also re?ects methodological differences
2.
What is the magnitude of the difference among severity
among studies (e.g., sample characteristics such as injury severity,
groups over the course of three time bands following
age at injury and time since injury, as well within subject over time
injury? (Case-case studies)
vs. cross-sectional study designs). For example, some studies of
3.
What is the magnitude of change over time (i.e., recov-
mild TBI have reported signi?cant functional impairments (Boll,
ery) in key neurocognitive domains across severity
1983) whereas others have reported minimal functional morbidity
groups? (Longitudinal studies)
once premorbid status has been accounted for (Asarnow et al.,
1995). Further complicating attempts at integrating the extant data
There have been a number of excellent reviews on neurocogni-
are the wide range of measures used to assess key domains of
tive outcomes after pediatric TBI. They have, however, focused on
neurocognitive outcome. Such methodological differences among
a subsample of this population (e.g., mild injuries [Satz et al.,
1997]) or on a speci?c cognitive domain (e.g., executive function
[Levin & Hanten, 2005]). Further, to the best of our knowledge,
there has not been a systematic meta-analytic review of neurocog-
Talin Babikian and Robert Asarnow, Psychiatry and Biobehavioral
Sciences, David Geffen School of Medicine at UCLA.
nitive outcomes across domains, accounting for time post injury,
Support for this work was funded by an NIH/NINDS fellowship
injury severity, and age at injury. In this review, we attempted to
(F32NS053169) and a fellowship through the Della Martin Foundation
provide a systematic, quantitative summary of the literature.
awarded to the primary author.
Correspondence concerning this article should be addressed to Talin
Method
Babikian, Psychiatry and Biobehavioral Sciences, David Geffen School of
Medicine at UCLA, 760 Westwood Plaza, Room 58-217, Los Angeles, CA
We attempted to identify all studies that reported neurocognitive
90095. E-mail: tbabikian@mednet.ucla.edu
outcomes subsequent to pediatric TBI. Studies were included that
283

---

284
BABIKIAN AND ASARNOW
either directly reported neurocognitive outcomes as a primary
relatively little change is expected thereafter. We used the mean
objective or reported data on neurocognitive outcomes for a related
time post injury for a study when assigning it to one of the three
objective (e.g., imaging study). Studies published in English report-
time post injury categories. We are fully aware that there were
ing neurocognitive outcomes after a TBI were identi?ed using the
several studies that included patients with time post injury span-
PubMed database. Search terms included combinations of neuropsy-
ning across these three time bands. We attempted to minimize the
chology/cognition/neurocognition, head injury/TBI, child/pediatric.
overlap as much as possible by carefully selecting our cutoffs.
Studies of in?icted injuries on infants and children were not
Finally, the following two age at injury groupings were created
included, because nonaccidental head trauma patients are unique in
based on the group means reported in studies: Injury age 1: 0 –5
their presentation and their prognoses, tend to be younger (Gilles,
years and Injury age 2: 6 –16 years. This cutpoint was chosen for
1999), and the neuropathology and neurophysiology of nonacci-
several related reasons: (1) in general, different types of neuro-
dental trauma is different from accidental brain injury (Geddes,
cognitive measures are used for these two age groupings, (2) by 6
Hackshaw, Vowles, Nickols, & Whitwell, 2001). The search
years, most children have begun formal education, signi?cantly
yielded 115 publications. To calculate effect sizes, only studies
affecting their development, and (3) by about 6 years, children
that reported descriptive group statistics or group difference sta-
begin to show a greater functional utilization of language and
tistics were included, thereby excluding 25 studies, many of them
metacognitive skills, which separates them from their younger
pivotal. These studies described growth curve analyses or statisti-
counterparts. There were several studies that included patients with
cal associations between neurocognitive variables and other fac-
age at injury ranges spanning across our two age at injury group-
tors (e.g., imaging correlates). Of the remaining studies, an addi-
ings. Again, we chose our cutoff to minimize this overlap as much
tional 25 were excluded because injury severity was unclear (N
as possible.
7) or severity groups were combined (N
18), leaving 65 studies.
The statistics extracted from publications included the minimal
There were only a few studies using other injury controls instead
descriptive data necessary to calculate effect sizes. These included:
of healthy controls and because these comparisons were different
(1) mean and standard deviation (when standard deviation was not
in nature and few in number, they were not included in the
reported, it was calculated using con?dence intervals or standard
analyses. An effort was also made to identify and exclude studies
errors of the means), (2) sample size and t-value, (3) difference in
using duplicate subject pools (e.g., same lab and same recruitment
means and common standard deviation, or (4) sample size and p
period), except when the reported outcomes were different, thereby
value. Rosenthal and DiMatteo (2001) review the mathematical
excluding an additional eight studies. Despite our best efforts, there
calculations used to derive effect sizes from the above descriptive
may be some subject overlap, which we were unable to identify from
data (Rosenthal & DiMatteo, 2001). The extracted descriptive and
the study descriptions provided. An additional 29 were excluded
group difference statistics were entered into a database in the
because they did not include data on the discrete neurocognitive
software program Comprehensive Meta Analysis (Borenstein,
domains we were interested in for this meta analysis. The resulting
Hedges, Higgins, & Rothstein, 2005). Hedge’s g was used as the
?nal database included a total of 28 publications, dating from 1988
estimate of effect because it is adjusted for sample size (McCart-
to 2007.
ney & Rosenthal, 2000). Random instead of ?xed effects models
To answer the questions posed by this review, one would ideally
were chosen because effect sizes were expected to vary across
include only longitudinal studies with clearly de?ned injury sever-
studies depending on the sample characteristics and the neurocog-
ity, age at injury, and narrow time since injury bands. Unfortu-
nitive measures used, and the weights assigned to each study were
nately, many studies were not organized in this manner. Conse-
balanced based on study sample size (Borenstein, Hedges, &
quently, we attempted to verify the results from the relatively few
Rothstein, 2007). A general interpretive guide to effect sizes
longitudinal studies of children with TBI by examining the same
includes: large effects (.8), moderate effects (.5), and small effects
neurocognitive domains at similar times post injury in cross-
(.2) (Cohen, 1988). We calculated effect sizes for subgroups based
sectional reports. To group the studies, we set cut points for each
on injury severity and time since injury. Although separate statis-
of the following key moderating variables: (1) injury severity, (2)
tics were not calculated for the age at injury variable (because
time since injury, and (3) age at injury.
relatively few studies were identi?ed that ?t the lower age at injury
In a majority of the studies, acute injury severity was based on
band), studies were tracked on this variable to make it possible to
Glasgow Coma Scale (GCS) score and frequently con?rmed by
identify effect sizes that were driven by study samples including
clinical ?ndings (e.g., presence/length of loss of consciousness and
younger groups for post hoc discussion.
posttraumatic amnesia, and/or positive neuroimaging ?ndings).
There were very few longitudinal studies, which would ideally
Consistent with convention, most studies used the following GCS
be used to describe recovery of functions over time. For this
score ranges to group samples based on injury severity: mild (GCS
reason, when possible, scores from the same measure but using a
13–15), moderate (GCS 9 –12), and severe (GCS 3– 8). We used
different cohort of patients were matched to form comparable
the severity groups published in the studies to designate the injury
“within-subject” data. These analyses were used to compliment the
severity groups used in our analyses.
true within subject analyses and not replace them. Statistics using
Three mean time post injury intervals were de?ned using the
this combination approach are reported separately and clearly
following cutoffs: Time 1: 0 –5 months post injury; Time 2: 6 –23
identi?ed in the tables. (These statistics are always presented on
months post injury, and Time 3: 24
months post injury. In
the right side of “/” in the tables). Further, for the “within-subject”
general, Time 1 represents the postacute period, where the greatest
longitudinal studies, we only included studies reporting standard
functional impairments are observed. Time 2 represents the period
scores in the analyses and excluded all reported data using raw
when the greatest recovery is observed. Time 3 is the chronic
scores. We decided to do so because we could not account for
stage, when most of the recovery has already taken place and
normal developmental changes that may, in part or in full, explain

---

NEUROCOGNITIVE OUTCOMES AFTER PEDIATRIC TBI
285
changes in raw scores. Further, we are aware that in cases where
injury severity and time post injury. The effect sizes derived from
more than one measure of a function (e.g., attention) were re-
the three sets of analyses provide complimentary ?ndings and a
ported, we did not observe the independent observations prereq-
unique perspective on neurocognitive outcomes and recovery after
uisite for inclusion in the effect size calculations. We decided to
pediatric TBI that none of the analyses can do independently.
include multiple measures of the same neurocognitive domain
In subsequent paragraphs, the ?ndings for each severity group is
within the same study because we believe that each measure
reviewed with references made to the corresponding summary
contributes uniquely to the study of a given neurocognitive domain
statistic tables. The tables include three sets of statistics. Cross-
(e.g., sustained vs. divided attention) and that eliminating mea-
sectional studies summarizing case-control and case-case (more
sures may bias the results. We also had no basis of eliminating one
severe compared to less severe) studies are presented in the “a”
subskill wihin a given domain in favor of another.
and “b” rows of Table 1, respectively. In these analyses, a negative
The following neurocognitive domains were included in the
case-control effect size indicates that the cases performed more
analyses: General Intellectual Functioning (FSIQ or its equivalent;
poorly than the controls. Similarly, a negative case-case effect size
Verbal IQ; and Performance IQ); Attention/Executive Functions
indicates that the more severe groups performed more poorly than
(working memory, processing speed/reaction time, attention, ?u-
the less severe group. The magnitude of these differences is
ency, inhibition, and problem solving); Memory (verbal/visual
re?ected in the respective weighted mean effect sizes listed in
immediate/delayed); and Visual Perceptual/Motor skills. Although
corresponding sections in Table 1. Further, summaries from lon-
there were several pivotal studies of language functions, academic
gitudinal studies addressing recovery over time are presented in
skills, and motor skills, we focused on more traditional neurocog-
the corresponding “c” rows of Table 1, and compare the more
nitive domains to simplify our analyses. Global measures of func-
proximal evaluation to a later evaluation. For example, a negative
tioning were also excluded (e.g., combined visual and verbal
effect size comparing the Time 1 and Time 2 data points from
memory composites) because they provided redundant information
longitudinal analyses indicate that the Time 2 scores are compar-
and potentially erroneously combined skill areas that are uniquely
atively higher. The magnitude of the effect sizes provides a stan-
affected after injury.
dardized measure of the degree of the differences, which we have
interpreted as recovery. Table 2 lists all of the neurocognitive
measures contributing to the effect sizes tabulated by domain.
Results
Twenty-three of the 28 studies included in the meta-analysis
Outcomes After Mild TBI
used exclusion criteria that were similar across studies and in-
cluded, in large part, prior neurological (including previous head
Case-control studies (Table 1, rows 1-14a) showed negligible or
injury), developmental, learning/attention, and psychiatric diag-
small differences between the Mild and Control groups at all time
noses, penetrating injuries, injuries secondary to abuse, and in
points for FSIQ, PIQ, working memory, problem solving, visual
some cases, non?uent English speaking patients and/or parents.
immediate memory, and visual perceptual functioning. Negligible
Three of the studies did not indicate whether exclusion criteria
differences were also noted for VIQ and attention through Time 2,
were used. The remaining two explicitly stated that all patients
but small effects were noted at Time 3. The apparent effect size
who met inclusion criteria were recruited. Interestingly, one of the
increase in VIQ was present even if only the single study contrib-
latter two studies of severe TBI outcomes showed the largest
uting to all three time points was considered. The small effect
case-control effects across many neurocognitive domains (Parry et
noted for attention at Time 3, on the other hand, was due largely
al., 2004).
to a study of a younger age at injury cohort (not in earlier
In Table 1, effect sizes, an estimate of the associated standard
analyses). Small group differences were apparent up to 2 years
error (SE), the statistical signi?cance of the associated statistic (p),
post injury in verbal immediate and delayed memory, which re-
and the number of individual studies or individual reports from a
solved by Time 3. Small to moderate effects were noted for
given study (N) are reported. The speci?c studies from which
processing speed at all time points. Finally, generally negligible
descriptive statistics are extracted from are also identi?ed as ref-
differences in ?uency were noted at Time 1. This difference,
erences. Of note, in a few domains, there are more observations
however, appeared to get larger over time and was in the moderate
than studies because more than one measure of a given neurocog-
range by Time 3. (The single study contributing to this statistic was
nitive domain was included. The three sets of analyses for each
also based on data from a younger age at injury cohort).
neurocognitive domain correspond to the three questions posed in
Case-case studies (Table 1, rows 1–14b) showed no differences
this meta-analysis, addressing magnitude of (a) the effect of injury
between the Mild and Moderate groups at any time point in verbal
severity over the course of three time bands post injury, (b) the
immediate memory and ?uency, and at Time 3 for visual delayed
difference among severity groups over the course of three time
memory. Although no meaningful group differences in problem
bands post injury, and (c) change over time (recovery) in key
solving were apparent at Time 1, the groups are better differenti-
neurocognitive domains across severity groups. The ?rst two sets
ated over time, with a large effect size noted by Time 3. Across all
of questions are addressed by the cross-sectional case-control and
three time points, small (and frequently negligible) effects were
cross-sectional case-case reports, respectively. Corresponding to
apparent for attention, working memory, and verbal delayed mem-
these analyses, the ?rst two sets of effect sizes presented in Table 1
ory whereas small to moderate effects were noted for VIQ (re-
(sections 1a-14a and 1b-14b) re?ect the magnitude of effects listed
solved by Time 3), PIQ, processing speed, and visual perceptual
separately by injury severity and by time post injury. The third
functioning. For visual immediate memory, a moderate effect size
question is addressed using longitudinal within group data re?ect-
(although inconsistent across studies) was apparent at Time 1, with
ing the magnitude of change over time, also listed separately by
groups less differentiated over time.

---

286
BABIKIAN AND ASARNOW
)
Ref
Ref
Ref
Ref
(3–5)
(3–5)
(3–5,8,9)
(3–5,10)
(3–5,10)
(3–5,10)
(3–5)
(3–5)
(3–5,8,9)
(3–5,10)
(3–5,10)
(3–5,10)
Ref
continues
(3–5)
(3–5)
(3–5,8,9)
(2–5,10,11)
(2–5,10,11)
(3–5,10,11)
(1,3,10,11)
(1,3,10,11)
(1,3,10–12)
Ref
(1,3,10)/(5,14)
(1,3,4,10,14)
(1,3,9,10,12,14)
(
table
Sig
Sig
(1,3,10)/(4,14)
(1,3,10)/(1,3,4,10,14)
(1,3,9,10,12,14)
Sig
Sig
/
/
/
Sig
)
)
)
)
(
N
Sig
(
N
(
N
(
N
SE
.14(4)
.40(4)
.14(5)
SE
.16(3)
.20(3)
.36(5)
.17(5)
.15(5)
.19(5)
SE
.16(3)
.20(3)
.20(5)
.14(5)
.16(5)
.19(5)
SE
.16(3)
.20(3)
.23(5)
.13(6)
.14(6)
.17(6)
)
(
N
)
T3
SE
(
N
.19/.55(3/4)
.16/.16(3/4)
.15/.14(4/5)
3
vs.
.148
.204
.507
3
SE
3
.298
.599
.430
.469
.778
.358
.464
.903
.292
.428
.606
.15/.14(3/4)
.16/.17(3/4)
.20/.20(4.5)
.106
.579
.846
.450
.561
.867
1.106
T1
Time
Time
Time
T3
.385
.643
T3
.253
vs.
vs.
T1
.414/.175
.458/
.692/
Ref
Ref
Ref
Ref
(2)
(2)
(2)
(2,10)
(2,10)
(2,10)
T1
.070/.127
.066/.157
.376/
(2)
(2)
(2)
(2,10)
(2,10)
(2,10)
(2)
(2)
(2,6,7)
(2,10,11)
(2,10,11)
(2,10,11)
10–13)
(2,3,10,11)
(2,3,10,11)
(2,3,10–12)/(2–7,
Sig
Sig
Ref
Group
Sig
Sig
Ref
(2,3,10)
(2,3,10)
(2,3,10,12)
)
(2,3,10)
(2,3,10)
(2,3,10,12)
)
Injury
)
(
N
(
N
(
N
SE
.20(1)
.28(1)
.31(1)
.16(3)
.19(3)
.19(3)
SE
.20(1)
.28(1)
.32(1)
.16(3)
.20(3)
.19(3)
Sig
Post
)
SE
.20(1)
.28(1)
.20(4)
.14(4)
.16(4)
.17(4)
(
N
Sig
SE
.14(4)
.14(4)
)
Time
.14/.11(5/8)
2
)
2
(
N
2
(
N
SE
.16(3)
.16(3)
.15(4)
and
.053
.416
.393
.547
.418
.923
.237
.464
.787
.396
.586
.962
.171
.487
.564
.479
.998
1.086
Time
SE
.16(3)
.16(3)
.15(4)
Time
Time
T3
.145
T3
Severity
vs.
T3
vs.
.189
.024
.055
by
T2
.019
.004
.084/
vs.
.102
.084
.105
T2
Ref
Ref
Ref
T2
(1)
(1)
(1)
(1,10,14)
(1,10,14)
(1,10,14)
(1)
(1)
(1)
(1,10,14)
(1,10,14)
(1,10,14)
(1)
(1)
(1)
(1,10,11)
(1,10,11)
(1,10,11)
Domain
Ref
Ref
Ref
Sig
Sig
Sig
(1,2,10,11)
(1,2,10,11)
(1,2,10–12)
(1,2,10)
(1,2,10)
(1,2,10,12)
(1,2,10)
(1,2,10)
(1,2,10,12)
)
)
)
Neurocognitive
(
N
Sig
(
N
Sig
Sig
(
N
SE
.19(1)
.29(1)
.32(1)
.14(4)
.16(4)
.17(4)
SE
.19(1)
.29(1)
.32(1)
.16(4)
.17(4)
.17(4)
SE
.19(1)
.29(1)
.33(1)
.15(4)
.18(4)
.18(4)
Each
)
)
)
(
N
for
(
N
(
N
1
SE
.13(4)
.13(4)
.14(5)
1
SE
.15(3)
.16(3)
.15(4)
1
SE
.15(3)
.16(3)
.15(4)
.209
.917
.786
.656
.792
1.348
.143
.665
.523
.421
.556
.915
.216
.830
.892
.635
.791
1.394
Time
Time
Time
Statistics
T2
T2
T2
vs.
.129
.213
.438
vs.
.016
.025
.350
vs.
.249
.434
.664
T1
T1
T1
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Associated
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
Severity
vs.
Severity
and
Severity
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Sev
Severity
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Sev
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Sev
Severity
Severity
1
Sizes
1a
1b
1c
Table
Effect
FSIQ
VIQ
2a
2b
2c
PIQ
3a
3b
3c

---

NEUROCOGNITIVE OUTCOMES AFTER PEDIATRIC TBI
287
)
Ref
(1,3)
(1,3)
(1,3)
Ref
Ref
Ref
17,27
(1,3,10,11)
(1,3,10,11)
(1,3,10,11)
(3,4)
(3,4)
(3,4,9)
(3,4,11)
(3,4,11)
(3,4,11,15)
(1,3,10,11)
(1,3,10,11)
(1,3,10,11)
continues
Ref
(3–5)
(3–5)
(3–5,8)
(3–5,10,11)
(3–5,10,11)
(3–5,10,11,15)
Sig
(
table
Sig
Sig
Sig
(3–5,17–19)
(3–5,17)
(3–5,17,19,20)
(3–5,10,11,17,21)
(3–5,10,11,17,21)
(3–5,10,11,17,19–21)
Sig
)
)
)
(
N
(
N
)
SE
.21(1)
.31(1)
.34(1)
Sig
)
(
N
SE
.14(4)
.14(4)
.18(4)
(
N
(
N
SE
.15(4)
.14(4)
.18(4)
SE
.20(5)
.14(5)
.14(6)
.11(8)
.11(8)
.11(9)
SE
.19(3)
.21(3)
.23(4)
.13(6)
.14(6)
.15(8)
)
(
N
T3
T3
T3
SE
.10(7)
.21(4)
.15(8)
.09(9)
.09(9)
.08(13)
vs.
.035
.087
.360
3
3
vs.
.168
.052
.484
vs.
.354
.331
.570
.336
.731
.925
.385
.503
.607
.347
.316
.710
.246
.390
.570
T1
T1
Time
T1
Time
3
.092
.194
.453
.181
.124
.388
Time
Ref
(2,3)
(2,3)
(2,3)
Ref
Ref
Ref
Ref
(2)
(2)
(2)
(2,10,11)
(2,10,11)
(2,10,11)
(2,3,10,11)
(2,3,10,11)
(2,3,10,11)
(2)
(2)
(2)
(2,11,16)
(2,11,16)
(2,11,16)
(2,3,5,10,11,16)
(2,3,5,10,11,16)
(2,3,5,10,11,16)
Ref
(2,3,10,11)/
(2,3,10,11)/
(2,3,10,11)/
Sig
Sig
(2,17)
(2,17)
(2,17)
(2,10,11,16,17)
(2,10,11,16,17)
(2,10,11,16,17)
Sig
Sig
Sig
)
)
(
N
(
N
)
Sig
SE
.21(1)
.28(1)
.32(1)
.20(4)
.16(4)
.19(4)
SE
.21(1)
.31(1)
.34(1)
(
N
)
SE(N)
.19(4)
.14(4)
.18(4)
SE
.20(1)
.28(1)
.32(1)
.16(6)
.19(6)
.15(6)
)
(
N
(
N
SE
2
.14/.13(4/5)
.14/.13(4/5)
.18/.16(4/5)
T3
SE
.17(2)
.18(2)
.20(2)
.12(6)
.13(6)
.14(6)
.508
.221
.645
.255
.659
.877
T3
2
Time
vs.
.171
.410
.213
vs.
.114
.119
.048
.165
.374
.507
.201
.652
.800
T2
2
T2
Time
T3
.125
.166
.050
.183
.254
.319
.043
.391
Time
vs.
T2
.016/.088
.075/
.146/
Ref
Ref
Ref
(1,2)
(1,2)
(1,2)
(1)
(1)
(1)
(1,11)
(1,11)
(1,11)
Ref
(1)
(1)
(1)
(1,2,10,11,14)
(1,2,10,11,14)
(1,10,11,14)
(1,2,10,11)
(1,2,10,11)
(1,2,10,11)
Ref
(1,17)
(1,17)
(1,17)
(1,10,11,17)
(1,10,11,17)
(1,10,11,17)
Ref
Sig
Sig
Sig
(1,2,10,11)
(1,2,10,11)
(1,2,10,11)
Sig
Sig
)
)
)
(
N
)
(
N
(
N
Sig
SE
.19(1)
.29(1)
.34(1)
.13(5)
.15(5)
.15(5)
(
N
SE
.19(1)
.28(1)
.32(1)
.13(4)
.15(4)
.16(4)
SE
.20(1)
.28(1)
.31(1)
)
SE
.13(4)
.15(4)
.18(4)
(
N
)
SE
.16(2)
.18(2)
.20(2)
.17(5)
.13(5)
.19(5)
(
N
1
SE
.13(4)
.13(4)
.18(4)
.331
.629
.545
.692
1
T2
1.181
1.188
T2
.011
.373
.662
.334
.509
.878
Time
1
Time
vs.
.016
.344
.254
vs.
.187
.475
.554
.019
.050
.305
.309
.230
.549
T1
T2
Time
T1
vs.
.049
.018
.228
T1
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
vs.
vs.
vs.
vs.
vs.
Ctrl
Ctrl
Mod
Sev
Sev
)
vs.
vs.
vs.
vs.
vs.
vs.
vs.
Ctrl
Severity
Mild
Mod
Sev
Severity
Mild
Mod
Sev
vs.
vs.
vs.
vs.
vs.
vs.
Mild
Mod
Sev
Mild
Mod
Mild
Severity
Mild
Mod
Sev
Mild
Mod
Mild
Severity
Severity
Mild
Mod
Sev
Mild
Mod
Sev
Mild
Mod
Mild
Severity
(
continued
1
4a
4b
4c
speed
5a
5b
5c
6a
6b
6c
Table
Processing
Attention
Working
memory

---

288
BABIKIAN AND ASARNOW
)
Ref
Ref
Ref
(1,3)
(1,3)
(1,3)
(23)
(9,23)
(23)
Ref
Ref
(17)
(17,23)
(17,23)
(17)
(17,23)
(15,17)
Ref
/(6,14)
/(14,25)
(3)
(3)
(3,8,26)
(3)
(3,23)
(3,8,9,23)
(3)
(3,23)
(3,15)
Ref
continues
(3,17)
(3,17)
(17,18,20)
(5,17)
(3,5,17)
(3,5,17,20,25)
Sig
(
table
Sig
Sig
Sig
Sig
Sig
Sig
)
)
)
)
(
N
)
(
N
(
N
(
N
)
SE
.22(1)
.31(1)
.34(1)
(
N
SE
.32(1)
.14(3)
.29(3)
.35(1)
.14(3)
.24(2)
SE
.25(1)
.20(2)
.24(1)
(
N
)
SE
/.21(/1)
/.29(1)
(
N
SE
.22(1)
.34(1)
.45(3)
SE
.22(1)
.13(4)
.14(6)
.29(1)
.13(4)
.16(3)
SE
.15(3)
.19(2)
.23(7)
.30(2)
.18(3)
.15(5)
3
3
T3
T3
3
.572
.403
.316
.395
.348
1.115
.725
.389
.096
3
vs.
.775
.134
.385
3
vs.
.384
.085
.263
.878
Time
Time
.180
.402
.701
.740
.180
.305
/.151
/
T1
.061
.224
.552
.122
.434
.545
T1
Time
Time
Time
Ref
Ref
Ref
Ref
(17)
(17,22)
(17)
(17)
(17)
(17)
(2,3)
(2,3)
(2,3)
Ref
(2)
(2)
(2,6)
(2)
(2)
(2)
(2)
(2,22)
(2,6)
Ref
Ref
(2,17)
(2,17,22)
(1,2,6,7,17)
(2,17)
(2,17)
(2,17)
Sig
Sig
Sig
/(6,7,14,20,25,26)
Sig
Sig
Sig
)
)
(
N
)
)
)
Sig
(
N
SE
(
N
(
N
(
N
)
SE
.32(1)
.23(1)
.27(1)
.30(1)
.25(1)
.33(1)
SE
.21(1)
.31(1)
.34(1)
SE
.20(1)
.34(2)
.18(3)
SE
.20(1)
.28(1)
.13(5)
.24(1)
.30(1)
.27(1)
(
N
SE
.17(2)
.41(3)
.14(6)
.19(2)
.19(2)
.21(2)
)
(
N
2
2
SE
/.28(4)
2
2
T3
.419
.099
.410
.272
.279
.014
2
Time
.180
.669
.861
Time
.165
.342
.024
.397
.347
.729
vs.
.343
.064
.259
.313
.624
.809
.118
.332
.215
Time
Time
T2
Time
T3
vs.
.289
/
Ref
T2
Ref
Ref
(1)
(1)
(1)
(1)
(1)
(1)
Ref
(17)
(17)
(17)
(14,17)
(14,17)
(14,17)
Ref
(1)
(1)
(1,24)
(1,2)
(1,2)
(1,2)
Ref
(1,17)
(1,17)
(1,17,24)
(1,14,17)
(1,14,17)
(1,14,17)
Ref
Sig
Sig
Sig
Sig
Sig
Sig
)
Sig
)
)
(
N
)
(
N
)
(
N
SE
)
(
N
SE
.19(1)
.28(1)
.31(1)
.24(1)
.30(1)
.26(1)
(
N
(
N
)
SE
.31(1)
.23(1)
.27(1)
.37(2)
.22(2)
.44(2)
SE
.19(1)
.28(1)
.23(2)
SE
.20(1)
.28(1)
.31(1)
SE
.17(2)
.18(2)
.18(3)
.33(3)
.18(3)
.45(3)
(
N
SE
1
1
1
1
1
.139
.152
.288
.340
.172
.592
Time
.150
.325
.052
.183
.284
.503
T2
.327
.098
.521
.130
.338
.613
T2
.355
.492
.671
Time
Time
vs.
.440
.211
.139
Time
Time
vs.
T1
T1
Ctrl
Ctrl
Sev
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
vs.
Ctrl
Ctrl
vs.
vs.
vs.
vs.
vs.
vs.
vs.
)
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
Severity
vs.
vs.
Severity
vs.
vs.
Severity
Severity
Mod
Sev
Mod
Mild
Mod
Sev
Mild
Mod
Sev
Severity
Mild
Mod
Sev
Mild
Mod
Mild
Severity
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Sev
Mild
Mod
Mild
Severity
Mild
Mod
Sev
(
continued
1
7a
7b
8a
8b
9a
9b
9c
10a
10b
10c
Verbal
11a
Table
Fluency
Inhibition
Problem
solving
immediate
Verbal
delayed

---

NEUROCOGNITIVE OUTCOMES AFTER PEDIATRIC TBI
289
)
Ref
Ref
Ref
Ref
Ref
(8)
(3,5)
(5)
(5)
/(6,14)
(3)
(3)
(3,8,9)
(3,5)
(3,5)
(3,5)
(5)
(3,5)
(3,5,25)
/(14,25)
Ref
continues
Sig
(3,5)
(3,5)
(3,5,9,27)
(3,5,10)
(3,5,10)
(3,5,10,15,28)
Sig
(
table
Sig
Sig
Sig
)
Sig
)
(
N
)
)
)
(
N
SE
(
N
(
N
(
N
SE
.22(1)
.34(1)
.39(3)
.17(3)
.26(2)
.23(2)
SE
.37(1)
.22(2)
.40(1)
.33(1)
)
SE
.35(1)
.26(2)
.15(3)
SE
/.21(1)
/.42(1)
(
N
SE
.18(2)
.26(2)
.16(6)
.15(4)
.20(4)
.09(6)
T3
3
3
3
T3
vs.
.000
.061
.338
.029
.455
.525
.836
.031
.596
.831
.167
.559
.653
3
vs.
.039
Time
T1
Time
Time
/.120
/
.156
.171
.382
.264
.279
.448
T1
Time
Ref
Ref
(2)
(2)
(2)
Ref
Ref
(2)
(2,22)
(2,7)
(2)
(2)
(2)
/(5,22)
Ref
Ref
/(5,22)
/(6,14)
(2)
(2)
(1,27)
(2,10)
(2,10)
(2,10)
Sig
Sig
Sig
Sig
/
Sig
Sig
)
)
)
)
(
N
(
N
(
N
)
(
N
SE
.24(1)
.30(1)
.26(1)
)
(
N
SE
.20(1)
.50(2)
.17(3)
.24(1)
.30(1)
.26(1)
SE
(
N
SE
/.36(1)
SE
.20(1)
.28(1)
.19(3)
.16(3)
.23(3)
.21(3)
SE
/.36(1)
/.39(1)
2
2
2
2
T3
.172
.316
.491
.132
.694
.907
.173
.238
.495
.045
.000
.750
.338
.294
.608
Time
T3
Time
vs.
.584
Time
Time
/
vs.
T2
/-.688
/.138
T2
Ref
(1)
(1)
(1)
(1)
(1)
(1)
Ref
Ref
(1,14)
(1,14)
(1,14)
Ref
Ref
(1)
(1)
(1)
(1,10,14)
(1,10,14)
(1,10,14)
Ref
Sig
Sig
Sig
Sig
Sig
Sig
)
)
)
(
N
)
(
N
)
(
N
)
SE
.19(1)
.29(1)
.31(1)
.24(1)
.30(1)
.27(1)
(
N
SE
(
N
SE
.21(2)
.25(2)
.21(2)
(
N
SE
SE
.19(1)
.28(1)
.31(1)
.16(4)
.18(4)
.25(4)
SE
1
1
1
1
.278
.895
.346
.316
.289
.636
T2
.342
.401
.801
Time
.173
.033
.464
.326
.575
.891
T2
Time
Time
vs.
Time
vs.
T1
T1
Ctrl
Ctrl
Ctrl
Mod
Sev
Sev
Ctrl
Ctrl
Mod
Sev
Sev
Mod
Sev
Sev
Mod
Sev
Sev
Ctrl
vs.
vs.
Ctrl
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
)
vs.
vs.
vs.
Severity
vs.
vs.
vs.
vs.
Mild
Mod
Sev
vs.
Severity
Severity
Severity
Severity
Mild
Mod
Sev
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Sev
Mild
Mod
Mild
Mild
Mod
Mild
Severity
Sev
Mild
Mod
Mild
(
continued
1
12a
12b
12c
14a
14b
11b
11c
Visual
13a
13b
Visual
Table
Verbal
delayed
immediate
Visual
delayed
perceptual

---

290
BABIKIAN AND ASARNOW
Longitudinal studies (Table 1, rows 1–14c) did not show
Ref
al.,
(10)
(10)
(10)
et
changes in VIQ, FSIQ, attention, working memory, or visual
(Verger
perceptual functioning. There were no longitudinal studies ad-
9
(Kaufmann
dressing ?uency, memory, or inhibition that met study inclusion
(Levin
16
criteria. However, in both verbal immediate and delayed memory,
Sig
2004);
23
combining cross-sectional data from studies that used the same
al.,
2000);
measure at Time 1 and Time 3 suggested unremarkable change
et
2004);
)
al.,
over time. Small changes for processing speed were noted, with
(
N
et
considerable improvements in problem by Time 3.
SE
.23(2)
.19(2)
.43(2)
(Parry
8
Figure 1 summarizes the effect sizes for three key ?ndings
May?eld,
(case-control postacute and chronic studies, and longitudinal stud-
&
998);
1
ies) from the analyses discussed above. Effect sizes from case-
(Brookshire
T3
control studies show that postacutely (Time 1), the majority of the
al.,
15
vs.
.129
.265
.602
et
(Lowther
neurocognitive domains reviewed were either not affected or were
T1
22
minimally affected (see Figure 1A). By Time 3 (chronic phase),
(Ong
1995);
some of the impairments apparent at earlier time points had re-
7
al.,
solved. In several cases, however, increasing case-control differ-
et
2004);
ences were apparent by Time 3 (chronic phase) (see Figure 1B).
Ref
(10)
(10)
(10)
1994);
al.,
et
With the exception of VIQ, however, many of the effects were not
al.,
(Kinsella
statistically signi?cant because of relatively large standard errors
et
2001).
14
indicating inconsistency across study results. Of note, large effects
al.,
Sig
et
in some domains (e.g., ?uency) were apparent in studies of a
(Levin
(Roncadin
6
2005);
younger age at injury cohort (2–7 at injury) (Anderson, Morse,
21
Catroppa, Haritou, & Rosenfeld, 2004). With few impairments
al.,
(Levin
)
2004);
et
noted postacutely, little, if any, recovery was observed in longitu-
(
N
2006);
28
dinal studies by Time 3, with some noteworthy exceptions, includ-
SE
.23(2)
.19(2)
.27(2)
al.,
(Ayr
al.,
et
ing problem solving, PIQ, and processing speed (see Figure 1C).
13
et
2003);
Of all the measures administered, problem solving may be one
al.,
of the more likely domains to be affected by practice effects.
et
T3
(Hawley
1997);
Assuming so, the Mild group appears to gain substantially from
5
vs.
.037
.007
.051
al.,
(Chapman
these effects while the Moderate and Severe groups gain compa-
T2
et
20
Lehnung
rably less. Also of interest, the Mild group appears to make notable
2007)
27
gains in PIQ and processing speed (factored into the PIQ on
al.,
2002);
et
Wechsler Scales). Because all scores used in the meta-analysis
al.,
were standard scores (accounting for age related development),
Ref
(10)
(10)
(10)
et
2000);
(Ewing-Cobbs
signi?cant changes in performance from Time 1 to Time 3 suggest
al.,
12
et
that the groups made improvements above and beyond that ex-
(Catroppa
4
(Levin
pected during normal development. This ?nding is somewhat
Sig
2005);
19
surprising because PIQ and processing speed are typically not
1994);
(Hanten
prone to practice effects. Also of note, the Mild and Moderate
26
al.,
2004);
groups, particularly with regard to measures of intelligence, pro-
)
et
al.,
cessing speed, attention, and working memory, appear to show the
(
N
Anderson,
et
SE
.23(2)
.19(2)
.30(2)
&
2000);
same pattern of “recovery,” despite the Mild group’s smaller
(Fay
3
al.,
discrepancy from Controls. In summary, the Mild TBI group
et
showed generally few, if any, impairments in aspects of general
1993);
(Catroppa
(Chapman
intelligence, attention/executive skills, and memory, as well as
T2
al.,
11
18
some recovery in these areas at around two years post injury.
Stefano
vs.
.099
.257
.594
et
T1
(Di
2005);
2004);
25
Outcomes After Moderate TBI
(Jaffe
2
al.,
al.,
et
et
Case-control studies (Table 1, rows 1–14a) showed negligible
1998);
overall differences between the Moderate and Control groups in
)
1992);
al.,
Mild
Mod
Sev
visual perceptual functioning, working memory, and verbal imme-
Severity
al.,
et
(Anderson
(Anderson
diate memory (with signi?cant inconsistency across studies) at any
et
10
17
time point. Generally negligible differences were noted for ?uency
(Jaffe
(Roman
by Time 2, with a larger discrepancy apparent by Time 3. Small
(
continued
1
1
1993);
24
but consistent effects were noted for attention and problem solving
2000);
14c
at all time points. Small to moderate effects for verbal delayed
Visual
al.,
al.,
Table
perceptual
Note.
et
et
1993);
memory and moderate to large effects for visual immediate mem-

---

NEUROCOGNITIVE OUTCOMES AFTER PEDIATRIC TBI
291
Table 2
List of Measures Used in Meta-Analysis for Each Neurocognitive Domain
Domain
Measures
Intelligence
Editions of the the Wechsler Scales of Intelligence (Preschool, Child, Adult), Stanford-Binet Intelligence Scales,
and the respective composite scores derived from these measures (e.g., FSIQ, VIQ, PIQ).
Processing speed/reaction time
Tests of reaction time (e.g., CANTAB) or the Processing Speed Index (PSI) of the Wechsler Scales (or subtest
scores comprising the PSI when PSI not reported).
Attention
Paper/pencil measures: Trails A and B, Letter Cancellation Test, subtests of the TEA-Ch, and contingency
naming tests.
Computerized measures: versions of the Continuous Performance Tests.
Working memory
Versions of verbal and visual span tests and n-back tests, Wechsler Scales Working Memory (WMI) or
Freedom from Distractibility (FDI) Indices (or subtest scores comprising the latter when WMT or FDI not
reported).
Fluency
Timed verbal ?uency tests (category or letter).
Inhibition
Versions of the Go-No Go test and the Interference subtest of the Stroop.
Problem solving
Measures of planning and problem solving, including WCST, the Category Test, Stockings of Cambridge
(CANTAB), delayed alternation tasks, Tower of London Test, and the 20 Questions Test.
Memory
Subtests and composites from various batteries of verbal/nonverbal, immediate/delayed memory, including
CMS, CVLT, TOMAL, WRAML, Rivermead Behavioral Memory Test, Rey-O, RAVLT, and several
nonpublished/experimental list learning tasks.
Visual spatial
Measures of visual motor and visual perceptual processing, including the Perceptual Organization Index from
Wechsler Scales, Tactual Performance Location trial, copy trial of the Rey-Osterrieth Complex Figure,
Benton’s Line Orientation Test, and various versions of the Mazes test (Kiel Locomotor Maze, Austin Maze,
Porteus Maze).
ory were apparent through Time 2, with this difference signi?-
1B). There were also signi?cant impairments in executive skills,
cantly decreasing by Time 3. Finally, large effects for FSIQ, VIQ,
including processing speed, attention, ?uency, inhibition, and
and PIQ were present at Time 1, which decreased some by later
problem solving. In contrast, working memory, memory, and
time points. Large effects for were noted for processing speed and
visual perceptual skills appeared largely commensurate with the
inhibition at Time 3.
Controls (see Figure 1B). Although few longitudinal studies were
Case-case studies (Table 1, rows 1–14b) show negligible dif-
available for review, substantial improvement in intellectual func-
ferences between the Moderate and Severe groups in working
tioning (speci?cally PIQ) and processing speed were apparent,
memory at all three time points, as well as inhibition and ?uency
with no changes in VIQ, attention, working memory, problem
at Time 3 (no earlier studies were available). Small, and in some
solving, or visual perceptual functioning noted (see Figure 1C).
cases, negligible, effects were noted at all three time points for
verbal and visual immediate memory, problem solving, and ?u-
Outcomes After Severe TBI
ency. Small to moderate differences were noted for VIQ and
processing speed (all three time points), and verbal and visual
Case-control studies (Table 1, rows 1–14a) showed small group
delayed memory at Time 3. Moderate to large effects were appar-
differences at Time 1 for ?uency and problem solving, with this
ent for FSIQ, PIQ, visual perceptual functioning, and attention,
difference appearing substantially larger over time (in the very
initially, with this difference decreasing in magnitude over time.
large range by Time 3). Small to moderate effects were noted for
There were no longitudinal studies of memory, ?uency, or
working memory and visual immediate memory, with the largest
inhibition. However, in both verbal delayed and visual immediate
difference noted for working memory at Time 3 and visual imme-
memory, combining cross-sectional data from studies using the
diate memory at Time 2. Small but statistically signi?cant effects
same measures at Time 2 and Time 3 suggested moderate changes
were apparent for inhibition at Time 3. Moderate effects were
over time. Longitudinal studies of FSIQ, PIQ, processing speed,
apparent for VIQ at Time 1, with the effect size almost doubling
attention, problem solving, and visual perceptual functioning
by Time 3, suggesting that the gap between the groups expands
showed some improvement (small to moderate) in the ?rst 2 years
over time. Moderate to large effects were observed for attention,
post injury, with no observable changes thereafter. Further, no
verbal immediate and delayed memory, and visual perceptual
improvements in VIQ or working memory were apparent. These
functioning. The latter differences persisted even at Time 3 and, in
results are presented in Table 1 (rows 1–14c).
some cases (e.g., verbal delayed memory), increased over time.
The summary graphs show signi?cant case-control effects post-
Large effects were noted for FSIQ, PIQ, and processing speed
acutely, although in some cases inconsistent across studies, in
across all three time points, with signi?cant increases in group
several neurocognitive domains, including general intellectual
differences by Time 2 for FSIQ. Large effects were also evident
functioning, processing speed, attention, problem solving, as well
for visual delayed memory at Time 3.
as visual immediate and verbal delayed memory. No impairments
Case-case analyses are presented in Table 1 (rows 1–14b). There
in verbal immediate memory and visual perceptual functioning
were no meaningful differences between the Moderate and Severe
were observed (see Figure 1A). By the chronic phase (Time 3),
groups for working memory at any time point or inhibition at
de?cits in intellectual functioning persisted, although considered
Time 3. There were, however, moderate differences in working
smaller in magnitude compared to postacute studies (see Figure
memory and ?uency between the Severe and Mild groups at

---

292
BABIKIAN AND ASARNOW
Figure 1.
Summary of effects between TBI and control groups for each neurocognitive domain in the postacute
(Time 1) and chronic (Time 3) phases, and recovery trends. (A) Time 1 (acute/post-acute) versus Control effect
sizes by injury severity and domain in cross-sectional studies. (B) Time 3 (chronic) versus Control effect sizes
by injury severity and domain in cross-sectional studies. (C) Time 3 versus Time 1 (Recovery) effect sizes by
injury severity and domain in longitudinal studies.
Time 1, which diminished over time. Small to moderate differ-
memory. No improvements in problem solving were noted. Small
ences were noted at all time points for verbal and visual immediate
changes in VIQ, working memory, attention, and visual perceptual
memory. Moderate to large effects between the Severe and both
functioning were noted over time. In contrast, moderate to large
the Mild and Moderate groups were also apparent for FSIQ, VIQ,
improvements were apparent in FSIQ, PIQ, processing speed, and
attention, problem solving, and visual perceptual functioning (all
visual perceptual functioning by Time 2, with no changes observed
of which decreased over time), verbal delayed memory (consistent
thereafter. Longitudinal study results are presented in Table 1
over time), and visual delayed memory at Time 3. Large to very
(rows 1–14c).
large differences were noted for PIQ and processing speed, with
The summary graphs indicate signi?cant impairments in general
both effect sizes decreasing over time. In all analyses, the Severe
intellectual functioning (in PIQ more so than in VIQ), aspects of
group was better differentiated from the Mild than the Moderate
executive functioning (especially processing speed and attention),
group.
as well as verbal memory (immediate and delayed) in the Severe
There were no longitudinal studies of memory, inhibition, or
TBI group compared to Controls (see Figure 1A). De?cits in
?uency. However, combining cross-sectional data from various
working memory, ?uency, and visual perceptual skills were also
studies using the same measures from different time points sug-
apparent, but studies did not report impairments in these domains
gested a small amount of improvement in verbal immediate mem-
on a consistent basis (see Figure 1A). In contrast, by the chronic
ory, with no such improvements observed for verbal delayed
phase (Time 3), signi?cant impairments in almost all neurocogni-

---