Professional Papers
Comprehensive Assessment of Memory Functioning Following Traumatic Brain Injury in Children
Last Updated:
This study examined specific memory functions in 52 children with mild-moderate or severe traumatic brain injury (TBI) and 29 noninjured controls using the Wide Range Assessment of Memory and Learning (WRAML). Children's recall varied as function of injury severity and task demands. The participants with severe brain injuries performed worse than controls on global measures of visual memory, learning, and general memory functioning, as well as on specific subtests measuring recall of contextual verbal information. Children with mild-moderate brain injuries performed similarly to controls except for poorer performance on 2 subtests measuring sound-symbol learning and recall of geometric designs. Results suggest that the WRAML provides clinically useful information and that specific aspects of memory processing need to be evaluated following childhood TBI.
Children, like adults, are vulnerable to impairments in memory functioning following traumatic brain injury (TBI; Begali, 1992; Savage & Wolcott, 1994). Post-traumatic amnesia is common early in recovery from TBI ( Ewing-Cobbs, Levin, Fletcher, Miner, & Eisenberg, 1990). In the case of more severe injury, children and their families have also reported memory problems that persist over time ( Knights et al., 1991). Even after apparently mild injury during childhood, some individuals have described difficulties with memory and learning as long as 23 years post-injury ( Klonoff, Clark, & Klonoff, 1993).
Empirical studies have helped define the nature of memory impairments following childhood TBI (see Table 1). One of the most consistent findings is that children with more severe injuries display impairments on measures of verbal learning relative to noninjured controls ( Bassett & Slater, 1990; Fay et al., 1994; Levin et al., 1994; Levin, Eisenberg, Wigg, & Kobayashi, 1982; Yeates, Blumenstein, Patterson, & Delis, 1995). For example, using the California Verbal Learning Test for Children (CVLT-C; Delis, Kramer, Kaplan, & Ober, 1994), Yeates and colleagues ( 1995 ) found that children with severe TBI demonstrated impairments in word list learning, delayed recall, and recognition compared to controls. Children with mild-moderate TBI were less impaired, showing intact word list learning and recognition relative to controls, but decreased spontaneous retrieval after a delay. Other aspects of verbal memory functioning following TBI in children have been less well studied. Increasing injury severity has been inconsistently associated with decreased memory for sentences ( Knights et al., 1991; Winogron, Knights, & Bawden, 1984) and recall of story paragraphs ( Bassett & Slater, 1990; Donders, 1993).
In the domain of visual-spatial memory, severely injured children have shown greater deficits in visual recognition memory ( Levin et al., 1982, 1988) and visual learning ( Ewing-Cobbs et al., 1990) compared to mildly injured groups. Visual reproduction memory was impaired in children with severe injuries compared to noninjured controls ( Bassett & Slater, 1990), but not compared to children with mild-moderate injuries ( Donders, 1993).
Although this body of research has provided important information about the type and extent of memory impairments in children following TBI, these studies have several methodological limitations.
| Study | Memory Tests | ||||
|---|---|---|---|---|---|
| |
|||||
| N | Age Range | Verbal | Visual-Spatial | ||
| Levin & Eisenberg ( 1979 ) | 45 | 13-18 | SRT* | -- | |
| Levin, Eisenberg, Wigg, & Kobayashi ( 1982 ) | 60 | 5-19 | SRT* | CRT* | |
| Winogron, Knights, & Bawden ( 1984 ) | 51 | 4.7-17.6 | Sentence recall | TPT* | |
| Levin et al. ( 1988 ) | 58 | 6-15 | SRT* | CRT* | |
| Hannay & Levin ( 1989 ) | 91 | 13-19 | -- | CRT* | |
| Ewing-Cobbs et al. ( 1990 ) | 37 | 4-15 | SRT* | VSRT* | |
| Bassett & Slater ( 1990 ) | 29 | M age = 15 | WMS-LM,* SRT,* WMS-LM, Delay* | WMS-VR*, WMS-VR, Delay* | |
| Knights et al. ( 1991 ) | 76 | 5-17 | Sentence recall* | TPT | |
| Donders ( 1993 ) | 30 | 10-16 | Denman story | Rey Figure | |
| Fay et al. ( 1994 ) | 72 | 6-15 | CVLT-C* | -- | |
| Levin et al. ( 1994 ) | 62 | 5-16 | CVLT-C* | -- | |
| Yeates et al. ( 1995 ) | 47 | 5-16 | CVLT-C* | -- | |
| Massagli et al. ( 1996 ) | 30 | 6-15 | CVLT-C* | -- | |
| Note. SRT = Buschke Selective Reminding Test (for children 12 and under, animal names were substituted for the adult versions); VSRT = Visual Selective Reminding Test; CRT = Visual Continuous Recognition Test; TPT = Tactual Performance Test-Block Memory; WMS-LM = Wechsler Memory Scale-Logical Memory; WMS-VR = Wechsler Memory Scale-Visual Reproduction; CVLT-C = California Verbal Learning Test-Children's Version. |
|||||
| *Significant differences between TBI severity groups were found on these measures, with worse performance among children with more severe injuries. |
|||||
As shown in Table 1, many of these evaluations relied on adult measures of memory functioning because few appropriate children's measures were available. Others employed non-standardized measures developed primarily for research purposes. In addition, these studies often examined isolated aspects of memory functioning such as word list learning or visual recognition. More research is needed to examine the range of memory problems experienced by children following TBI, using age-appropriate, standardized measures.
Recently, Sheslow and Adams ( 1990 ) developed the Wide Range Assessment of Memory and Learning (WRAML), a comprehensive measure of memory functioning for children ages 5-17. The WRAML has Verbal, Visual, and Learning Indexes and also provides a General Memory Index as a global measure of the child's functioning. Individual subtests allow comparisons between verbal and visual-spatial memory and between immediate and delayed recall. Similar in scope to the Wechsler Memory Scale-Revised ( Wechsler, 1987) for adults, this standardized instrument has the potential to further clarify changes in memory following childhood TBI.
Williams and Haut ( 1995 ) used the WRAML to examine differences in memory functioning among broad diagnostic groups, including children with TBI, epilepsy, substance abuse, and non-neurologically impaired psychiatric controls. The TBI group, which consisted of all children referred for assessment regardless of injury severity, did not differ significantly from the other diagnostic groups on the WRAML. The authors pointed out that the combined range of severity levels in their TBI group may have obscured important information about the nature of memory functioning in children with TBI.
The purpose of this study was to conduct a comprehensive evaluation of memory functioning using the WRAML in a sample of children and adolescents with mild-moderate and severe TBI. Specific goals were to (a) determine whether children with mild-moderate and severe TBI differ from noninjured control participants across a range of memory tests, and (b) identify which aspects of child memory functioning are most susceptible to impairment in each group. Performance on both immediate and delayed memory tasks was expected to decline as a function of severity of injury. In addition, previous studies suggested that participants with TBI would be particularly likely to differ from controls on word list learning and visual recognition tasks. Too few data have been reported to make clear predictions for other subtests.
METHOD
Participants
Fifty-two children with TBI, ranging in age from 6 to 17 years old, participated in the study. These children were consecutive referrals to child neuropsychological assessment services at three university-affiliated medical centers: Rusk Rehabilitation Center (University of Missouri-Columbia School of Medicine), Chestnut Ridge Hospital (West Virginia University School of Medicine), and Arkansas Children's Hospital (University of Arkansas for Medical Sciences). The study excluded 33 children with documented pre-injury academic difficulties, behavioral problems, previous or subsequent TBI, and medical conditions affecting the central nervous system; 2 children were dropped from the study because of insufficient data about TBI severity.
Injury severity (mild-moderate or severe) was determined by reviewing medical records to identify length of impaired consciousness (LIC; time to first following one-step commands) and Glasgow Coma Scale (GCS; Teasdale & Jennett, 1974). GCS was available for many but not all children (72% of mild-moderate and 65% of severe). Therefore, LIC was used as the primary determinant of severity. This method of classification was supported by previous studies that showed that LIC was a better predictor of cognitive outcome than GCS ( Ewing-Cobbs, Fletcher, Levin, Hastings, & Francis, 1996; Knights et al., 1991; Levin et al., 1982). Labeling of group severity based on LIC incorporated standards recommended and used in the TBI literature ( Dikmen, Donovan, Loberg, Machamer, & Temkin, 1993; Fletcher, Ewing-Cobbs, Francis, & Levin, 1995; Lezak, 1995). To be considered mildly to moderately injured, the child had to show an LIC less than or equal to 24 hr. The injury was classified as severe if LIC was more than 24 hr. All three referral centers contributed participants to the mild-moderate and severe groups (Missouri--14 and 11; Arkansas--7 and 9; West Virginia--8 and 3).
Table 2 illustrates the defining characteristics of each TBI group. Compared to participants in the Mild-Moderate group, children in the Severe group showed increasing duration of impaired consciousness and greater evidence of brain injury on CT/MRI. The mean GCS for each group, based on available data, approximated definitions used in other studies of TBI (e.g., Fay et al., 1994; Levin et al., 1988; Massagli et al., 1996). There was one case of a child with a moderate injury based on a GCS of 9, but with no available LIC. The data were analyzed both with and without this participant with no difference in results, so the child's data were retained in the Mild-Moderate group. The TBI groups did not differ in median time from injury until assessment, Kruskal-WallisX2(1, N = 52) = .12, p = .73, mean age at injury, F(1, 52) = .09, p = .77, or on a gross estimate of socioeconomic status based on payment source for the evaluation (e.g., Medicaid, Commercial Insurance; Fisher's Exact Test p = .50).
Twenty-nine normal control participants, ranging in age from 9 to 15 years old, were also recruited from the community near the West Virginia University School of Medicine. According to information provided by their parents, the children had no history of TBI, other significant medical disorders, academic failure, special education services, behavioral concerns, or psychiatric problems. As shown in Table 2, the noninjured control participants and the groups of children with TBI did
TABLE 2 Demographic Features of the Participants
| Mild-Moderate Severe | Control | ||
|---|---|---|---|
| Number of participants | 29 | 23 | 29 |
| Gender (percentage male) | 48.3 | 73.9 | 48.3 |
| Ethnic status (percentage White) | 82.8 | 91.3 | 89.7 |
| Family structure | |||
| Two parents (biological) | 51.7% | 52.2% | 51.7% |
| Two parents (one step-parent) | 20.7% | 21.7% | 10.3% |
| Single parent | 24.0% | 26.1% | 34.5% |
| Other | 3.5% | 0.0% | 3.4% |
| Age | |||
| M | 12.37 | 11.50 | 11.95 |
| SD | 3.57 | 2.74 | 1.74 |
| Age at injury | |||
| M | 10.74 | 10.43 | -- |
| SD | 4.00 | 3.62 | -- |
| Time since injury in months | |||
| Median | 8.28 | 3.45 | -- |
| Range | 0.16-87.78 | 0.56-99.35 | -- |
| Length of impaired consciousness in days | |||
| Median | 0.00 | 6.00 | -- |
| Range | 0.00-0.46 | 1.05-60.00 | -- |
| Glasgow Coma Scale | |||
| M | 12.29 | 6.33 | -- |
| SD | 2.80 | 2.38 | -- |
| CT/MRI findingsa | |||
| Bilateral | 10.3 | 47.8 | -- |
| Diffuse edema | 0.0 | 8.7 | -- |
| Frontal only | .5 | 0.0 | -- |
| Right | 17.2 | 21.7 | -- |
| Left | 13.8 | 3.0 | -- |
| Linear skull fracture | 20.7 | 4.4 | -- |
| None | 34.5 | 4.4 | -- |
| Cause of injurya | |||
| MVA | 44.8 | 60.9 | -- |
| Bike | 10.3 | 8.7 | -- |
| Pedestrian | x6421">3.5 | 17.4 | -- |
| Fall | 10.3 | 8.7 | -- |
| Assault | 6.9 | 0.0 | -- |
| Other | 24.1 | 4.4 | -- |
| Method of reimbursementa | |||
| Medicaid | 10.3 | 13.0 | -- |
| Self-pay | 17.2 | 8.7 | -- |
| Insurance | 69.0 | 60.9 | -- |
| Insurance + Medicaid | 3.5 | 17.4 | -- |
| Mild-Moderate | Severe | Control | |
|---|---|---|---|
| Intelligence | |||
| VIQ | |||
| M | 100.2 | 95.7 | -- |
| SD | 15.1 | 16.2 | -- |
| PIQ* | |||
| M | 96.8 | 85.0 | -- |
| SD | 18.2 | 17.1 | -- |
| FSIQ* | |||
| M | 98.24 | 89.52 | 102.60 |
| SD | 16.5 | 17.4 | 6.94 |
| Note. MVA = motor vehicle accident; VIQ Verbal Intelligence Quotient; PIQ Performance Intelligence Quotient; FSIQ = Full Scale Intelligence Quotient. |
|
| aPercentage of cases per group | |
| *p < .05. |
not differ in age, F(2, 81) 0. 64, p = .53; race (Fisher's Exact Test p = .88); gender, X2(2, N = 81) = 4.38, p 11; or family structure (Fisher's Exact Test p = .87).
Memory Measure
The VIRAML ( Sheslow & Adams, 1990) consists of three verbal, three visual, and three learning subtests, which yield a Verbal Memory Index, a Visual Memory Index, and a Learning Index, respectively. These nine subtests are also combined to provide a General Memory Index. Each subtest measures immediate recall of the material presented. The Verbal subtests include (a) the Story Memory subtest, which requires the child to retell information from two stories that are read aloud; (b) the Sentence Memory subtest, which requires exact repetition of sentences that increase in length and complexity; and (c) the Number-Letter subtest, which asks for repetition of random series of numbers and letters that increase in length. The Visual subtests are (a) the Picture Memory subtest, a visual recognition task that presents two similar pictures serially and asks the child to identify differences in the second picture; (b) the Design Memory subtest, a visual recall and reproduction task with four separate geometric designs that must be drawn from memory after a brief exposure; and (c) the Finger Windows subtest, a visual recall task involving imitation of a visual-spatial pattern after a demonstration.
Evaluation of learning following repeated exposure involves administration of the following Learning subtests: (a) the Verbal Learning subtest, which provides four complete repetitions of an unrelated word list; (b) the Sound Symbol subtest, which gives the child four trials to learn a series of paired associations between novel sounds and abstract visual symbols; and (c) the Visual Learning subtest, which requires the child to learn the spatial location of visual designs on a stimulus board over four trials. The score derived for each of the Learning subtests reflects the total amount of information learned over four practice trials.
Delayed free recall is measured following four of the subtests (Story Memory, Verbal Learning, Sound Symbol Learning, and Visual Learning). There is also a multiple-choice Story Recognition task presented after the delay, to allow assessment of cued recall. The original WRAML manual ( Sheslow & Adams, 1990) recommended that these Delay subtests be evaluated using difference scores. The test authors later compiled a table of normative means and standard deviations by age groups for the Delay subtests ( Adams, Sheslow, & Wilkinson, 1991). These normative data were used to derive age-adjusted scaled scores from each participant's raw scores on the Delay subtests. Both the immediate and delayed subtests from the WRAML have a mean of 10 and a standard deviation of 3. The index scores have a mean of 100 and a standard deviation of 15.
Procedure
Clinical group. Children were tested individually in a child assessment laboratory. All participants with TBI were given the WRAML as part of a battery of neuropsychological tests, including an age-appropriate Wechsler measure of intelligence ( Wechsler, 1981, 1989, 1991). These estimates of intellectual functioning showed that TBI groups did not differ in Verbal IQ, F(1, 52) = 1.07, p = .31, but the children with severe injuries performed significantly worse on Performance IQ measures than children with mild-moderate TBI, F(1, 52) = 5.66, p =.02 (see Table 2). The entire evaluation typically required 4 to 5 hr for completion. To avoid problems with participant fatigue, the WRAML was always administered either early in the day or after a break. At one referral site, the Sound Symbol and Visual Learning subtests were not administered at times to reduce the length of the clinical neuropsychological assessment. Therefore, for these two subtests and the composite Learning and General Memory Indexes, data analysis was conducted on 24 children in the Mild-Moderate group and 15 children in the Severe group.
Control group. Control group participants were recruited with the use of a flyer, or by verbal announcements about the study to parents and children in the community. The control participants first completed the WRAML, followed by a battery of neuropsychological tests. Eighty-six percent of the control group completed an estimate of intelligence using selected subtests of the Wechsler Intelligence Scale for Children, 3rd edition (Information, Vocabulary, Digit Span, Picture Completion, Coding, & Block Design). All were making average progress in school according to parent report. At the end of testing, which required between 2 and 3 hr, participants received their choice of two movie passes or $10 for participation. Statistical analyses. All analyses of group differences on the WRAML were conducted using the Jonckheere-Terpstra Test for ordered alternatives ( Daniel , 1990). This nonparametric procedure tests the null hypothesis against an alternative hypothesis in which the order of the means is specified and at least one population mean shows a strict inequality (i.e., is less than the others). Based on previously cited studies of TBI on memory functioning in children, the expected order of the means was M Severe ¡Ü M Mild-Moderate ¡Ü M Control . This statistical test was also considered appropriate because the WRAML data were non-normally distributed on several of the subtests. The Jonckheere-TerpstraJ test statistic is reported for all univariate analyses of variance (ANOVAs). The p values are based on the normal distribution with a-1 degrees of freedom, where a is the number of groups being compared. In this study, familywise Bonferroni adjusted p values were used for the initial tests of significance to reduce the likelihood of error due to multiple comparisons. Significant group differences were followed by post hoc tests using one-tailed, Bonferroni adjusted p values.
The first step in the analysis was to examine group differences (Mild-Moderate, Severe, and Control) on the four Index measures (Verbal, Visual, Learning, & General). Next, group differences on the nine immediate recall subtests and the five delayed recall and recognition subtests were analyzed individually using one-way ANOVAs. These subtest analyses were conducted because previous factor analytic studies questioned the composition of the WRAML Indexes ( Aylward, Gioia, Verhulst, & Bell, 1995; Callahan, Haut, Haut, & Franzen, 1993; Gioia, 1991; Haut, Haut, Callahan, & Franzen, 1992). These studies provided evidence that the WRAML immediate recall subtests assess separate factors or skills rather than three cohesive Verbal, Visual, and Learning domains. Furthermore, delayed subtests had to be examined separately because there is no global Delay index score available.
RESULTS
Index Scores
As shown in Table 3, the groups differed significantly on the Visual Memory Index, Learning Index, and General Memory Index, but not on the Verbal Memory Index (Bonferroni adjusted p = .013). Children with severe injuries performed significantly worse than noninjured control participants on these three indexes. The Severe group also displayed deficits compared to the Mild-Moderate group on the Visual Memory Index. The participants with mild to moderate injuries did not differ from the non-injured control children on any of the Indexes.
Relative to normative data, Index scores were in the average range in the Mild-Moderate and Control groups (see Table 3). The severely injured group displayed low average abilities on all indexes except the Learning Index, which was at the low end of the average range.
| Test | Controla | Mild-Moderateb | Severec | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |
|
|
|||||||||||
| M | SD | M | SD | M | SD | J | p | ||||||
| Verbal Memory Index | 98.07 | 13.13 | 94.41 | 15.79 | 88.87 | 15.92 | 1291.5 | .04 | |||||
| Visual Memory Index | 100.79d | 11.11 | 95.21d | 14.73 | 83.43e | 20.21 | 1477.0 | .0004* | |||||
| Learning Index | 107.62d | 13.51 | 97.96de | 18.60 | 91.60e | 19.51 | 987.5 | .003* | |||||
| General Index | 102.34d | 12.40 | 96.17de | 15.85 | 85.47e | 19.10 | 986.0 | .003* | |||||
| Note. Means in the same row that do not share subscripts differ based on one-tailed Bonferroni adjusted p values. There were 24 participants in the Mild-Moderate group and 15 in the Severe group for |
|||||||||||||
Subtest Scores
Verbal Memory subtests. The Severe group displayed deficits relative to the Control group on the immediate recall of the Story Memory subtest (Bonferroni adjusted p = .01; see Table 4). At the delay condition of the Story Memory subtest, severely injured children showed significantly worse recall than both the Control group and the Mild-Moderate group. However, on the Story Recognition subtest, the Severe group improved to the average range and no longer differed from the other two groups. There were no group differences on the Sentence Memory or Number-Letter Memory subtests.
Visual Memory subtests. The children with severe injuries demonstrated significantly weaker skills on the Picture Memory subtest compared to Control group participants (Bonferroni adjusted p = .017; see Table 4). Participants in both the Mild-Moderate and Severe groups performed worse than control participants on the Design Memory subtest. Groups did not differ significantly on the Finger Windows subtest.
Learning subtests. Groups differed significantly on the Verbal Learning and Sound Symbol subtests (Bonferroni adjusted p = .008; see Table 4). On the Verbal Learning immediate and delay subtests, children with severe TBI learned and retained significantly fewer words compared to control group participants. The children with mild to moderate injuries did not differ from control participants on
| Test | Controla | Mild-Moderateb | Severec | J | P | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| |
|
|
|||||||||
| M | SD | M | SD | M | SD | ||||||
| Verbal Memory subtests | |||||||||||
| Story memory, immediate | 11.07d | 2.53 | 10.14de | 2.96 | 8.22e | 3.79 | 1363.5 | .008* | |||
| Story memory, delay | 10.84d | 2.47 | 10.13d | 2.94 | 8.00e | 3.15 | 1186.0 | .002* | |||
| Story recognition | 10.58 | 3.02 | 9.80 | 3.41 | 10.28 | 2.87 | 702.0 | .29 | |||
| Sentence memory | 9.76 | 2.52 | 9.00 | 3.31 | 8.96 | 2.79 | 1188.5 | .19 | |||
| Number-letter memory | 8.24 | 2.57 | 8.45 | 2.87 | 7.70 | 2.38 | 1169.5 | .24 | |||
| Visual Memory subtests | |||||||||||
| Picture memory | 9.97d | 2.23 | 8.97de | 2.21 | 7.13e | 3.15 | 1481.5 | .0003* | |||
| Design memory | 10.28d | 2.56 | 8.52e | 3.66 | 7.52e | 3.58 | 1416.0 | .002* | |||
| Finger windows | 10.03 | 2.76 | 10.48 | 2.84 | 7.87 | 4.33 | 1274.5 | .05 | |||
| Learning subtests | |||||||||||
| Verbal learning, immediate | 11.86d | 3.10 | 9.93de | 3.52 | 8.61e | 3.81 | 1437.5 | .001* | |||
| Verbal learning, delay | 11.18d | 2.92 | 9.19de | 4.14 | 7.47e | 3.88 | 1237.5 | .001* | |||
| Sound-symbol, immediate | 11.79d | 2.92 | 9.50e | 3.31 | 9.33e | 3.99 | 970.0 | .005* | |||
| Sound-symbol, delay | 11.60d | 2.55 | 9.31 | 3.16 | 8.78e | 3.90 | 964.5 | .002* | |||
| Visual learning, immediate | 9.90 | 1.95 | 9.50 | 3.27 | 7.87 | 3.74 | 888.5 | .05 | |||
| Visual learning, delay | 10.52 | 2.52 | 10.07 | 3.42 | 8.81 | 2.53 | 857.5 | .06 | |||
| Note. Means in the same row that do not share subscripts differ based on one-tailed Bonferroni adjusted p values. There were 24 participants in the Mild-Moderate group and 15 in the Severegroup for the Sound-Symbol and Visual Learning subtests. |
|||||||||||
| an = 29. bn = 29. cn = 23. | |||||||||||
| *Significance based on familywise Bonferroni adjusted p value. | |||||||||||
Normative comparisons. As shown in Table 4, children with severe TBI displayed low average performance on the Number-Letter, Picture Memory, Design Memory, Finger Windows, delayed Verbal Learning, and immediate Visual Learning subtests, compared to normative data. All groups performed within the average range on the remaining subtests.
Clinical Significance
The data were further examined to determine the proportion of participants in each of the three groups who could be considered impaired compared to WRAML normative data ( Adams et al., 1991; Sheslow & Adams, 1990). To be considered impaired, children had to score ¡Ü 77 on the WRAML Index measures and ¡Ü 6 on the subtests (i.e., approximately 1.5 SDs below average).
The groups differed significantly in the proportion of children showing impairment on the Visual Memory, Learning, and General Memory Indexes (all Fisher's Exact Test ps < .01). As shown in Figure 1, rates of impairment on these WRAML indexes were highest for the Severe group, somewhat lower in the Mild-Moderate group, and lowest in the Control group.
The number and percentage of participants who performed in the impaired range on the WRAML subtests are shown in Table 5. The groups differed significantly in rate of impairment on the immediate Story Memory subtest, all of the Visual Memory subtests, and all of the Learning subtests except the delayed Visual Learning subtest. As shown in Table 5, the Severe group had the highest rate of impairment on most of these subtests, followed by the Mild-Moderate group with the next highest rate. At least 25% of the participants were in the impaired range on 11 of the subtests for the Severe group, 5 of the subtests for the Mild-Moderate group, and 1 of the subtests for the Control group. Thus, the WRAML indexes and subtests do appear to be sensitive to clinically significant deficits among individual children with TBI.
DISCUSSION
The present study indicates that the WRAML can provide clinically useful information about memory and new learning abilities following TBI in children. Children's performance on the WRAML varied as a function of both injury severity and task demands. Children with severe brain injuries recalled significantly less material than control participants on global index measures of visual memory, learning, and general memory functioning. Those with mild to moderate injuries did not differ from controls on any of the indexes, and they performed significantly better than the Severe group on the Visual Memory Index. The Verbal Memory Index did not differ significantly across severity groups.
On the individual WRAML subtests, children with severe brain injuries performed worse than controls on measures of contextual verbal memory, picture recognition, visual reproduction, word list learning, and sound symbol learning. The groups did not differ on two of the visual subtests, Finger Windows and Visual Learning, based on comparisons of group means. However, the groups did differ on the Finger Windows and immediate Visual Learning subtests when comparing the proportion of children who performed in the impaired range; 40%-43% of children with severe injuries demonstrated impairment on these subtests. Although they displayed abilities similar to controls on most of the individual subtests, participants with mild to moderate injuries performed significantly worse than controls on measures of visual reproduction and immediate and delayed sound symbol learning.
| Subtests | Controla | Mild-Moderateb | Severec | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| |
|
|
|||||||||||
| N | % | N | % | N | % | Fisher's Exact Test | |||||||
| Verbal Memory subtests | |||||||||||||
| Story Memory, immediate | 0 | 0 | 3 | 10 | 8 | 35 | .0006* | ||||||
| Story Memory, delay | 1 | 3 | 3 | 12 | 6 | 32 | .02 | ||||||
| Story Recognition | 3 | 11 | 3 | 14 | 0 | 0 | .47 | ||||||
| Sentence Memory | 2 | 7 | 7 | 24 | 5 | 22 | .20 | ||||||
| Number-Letter memory | 8 | 28 | 8 | 28 | 9 | 39 | .66 | ||||||
| Visual Memory subtests | |||||||||||||
| Picture Memory | 2 | 7 | 3 | 10 | 12 | 52 | .0002* | ||||||
| Design Memory | 2 | 7 | 11 | 38 | 7 | 30 | .015* | ||||||
| Finger Windows | 2 | 7 | 2 | 7 | 10 | 43 | .0009* | ||||||
| Learning Index subtests | |||||||||||||
| Verbal Learning, immediate | 0 | 0 | 5 | 17 | 7 | 30 | .003* | ||||||
| Verbal Learning, delay | 1 | 3 | 8 | 31 | 7 | 35 | .005* | ||||||
| Sound-Symbol, immediate | 0 | 0 | 6 | 25 | 4 | 27 | .004* | ||||||
| Sound-Symbol, delay | 0 | 0 | 7 | 30 | 5 | 33 | .001* | ||||||
| Visual Learning, immediate | 1 | 3 | 5 | 21 | 6 | 40 | .008* | ||||||
| Visual Learning, delay | 2 | 7 | 5 | 22 | 3 | 20 | .30 | ||||||
| Note. There were 24 participants in the Mild-Moderate group and 15 in the Severe group. The number of participants per group on some of the other subtests varied slightly. |
|||||||||||||
| an = 29. bn = 29. cn = 23. | |||||||||||||
| *Significance based on familywise Bonferroni adjusted p value. | |||||||||||||
Results from the WRAML indicated a general liability among children with severe TBI on tasks requiring rapid visual scanning and immediate visual recall. Deficits on the Visual Memory Index relative to the Verbal Memory Index among children with severe brain injuries paralleled the weakness in visual-spatial intellectual abilities compared to verbal intelligence obtained in this study and commonly observed following TBI ( Chadwick, Rutter, Brown, Shaffer, & Traub, 1981; Massagli et al., 1996; Winogron et al., 1984). This pattern of memory problems was also consistent with other studies that have shown perceptual-motor and organizational processes are more affected than basic acquired language processing abilities in children with TBI ( Fletcher, Levin, & Butler, 1995; Levin et al., 1988)
Although the WRAML Index scores captured important global differences in visual versus verbal memory functioning, performance patterns on the individual WRAML subtests must also be examined ( Aylward et al., 1995; Callahan et al., 1993; Gioia, 1991; Haut et al., 1992). Children's memory functioning was not simply spared across verbal tasks and impaired across visual tasks. For example, the TBI and control groups did not differ on the Visual Learning subtest. This subtest provides repeated exposure to visual stimuli, a task characteristic that may help children with severe TBI compensate for weaknesses in rapid visual processing.
In addition, children in the Severe group showed differential abilities across the Verbal Memory subtests. They had more difficulty than control participants on the spontaneous recall of stories, but they performed no differently from controls on multiple-choice story recognition questions. They also performed similarly to controls on the Sentence Memory and Number-Letter Memory subtests, two subtests assessing verbal short-term memory ( Aylward et al., 1995). This finding is consistent with other studies using measures of sentence repetition and digit span ( Kaufmann, Fletcher, Levin, Miner, & Ewing-Cobbs, 1993; Winogron et al., 1984). Thus, on the Verbal Memory subtests, children with severe injuries did not have difficulty encoding or storing information compared to controls, but they clearly displayed deficits in the spontaneous retrieval of semantically and syntactically complex story paragraphs.
The improvement in story recall with multiple-choice cues differed from the results of a study by Yeates et al. ( 1995 ), which found that children with severe injuries did not benefit from recognition cueing on the CVLT-C. Taken together, these data indicate that children with severe injuries may be able to encode and store meaningful verbal material more readily than the rote word list. Unfortunately, this premise could not be directly examined in this study because the Verbal Learning subtest of the WRAML does not include a recognition trial.
Results from the Verbal Memory and Verbal Learning subtests can be related to recent studies of language functioning among children with TBI. Chapman ( 1995 ) reviewed a growing body of literature indicating that children with more severe injuries typically show good recovery of basic linguistic structures (word/morpheme, clauses, sentences), but display marked impairment in narrative discourse (the ability to convey a message via a series of ideas expressed in sentences). Using story retelling as the measure of interest, Chapman found that children with severe injuries displayed deficits in information organization (decreased units of meaning, incomplete episodes, missing the main idea) beyond that explained by vocabulary or working memory (i.e., Trial 1 of the CVLT-C). Although scored differently, this story retelling task is similar to the WRAML Story Memory subtest. The studies on discourse suggest that children with TBI may have more difficulty recalling the Story Memory subtest due to weaknesses in information organization. This idea is supported by the finding that children's delayed WRAML story recall improved markedly with structured cues. A weakness in the ability to organize verbal material may also account for deficits on the Verbal Learning subtest. More research is needed to explore the relation between discourse and verbal memory and learning in children with TBI.
Examination of performance on the four delay condition subtests indicated that children had difficulty after the delay only if they showed initial problems with immediate recall. That is, the severe TBI group differed from the Control group at both the immediate and delay conditions of the Story Memory and Verbal Learning subtests, and both TBI groups differed from controls on the immediate and delay subtests of the Sound Symbol subtest. Neither TBI group showed initial recall comparable to the Control group followed by rapid forgetting on any of the delay subtests. Thus, the major problem for children with TBI appeared to be with the initial encoding, consolidation, and retrieval of the material presented, rather than with rapid information loss over time. It should be noted that some children with TBI may experience problems due to poor retention and that both encoding/retrieval and retention may be a concern in real-life settings.
Although statistically significant group differences were identified in this study, children with TBI exhibited generally intact (average to low average) memory functioning compared to standardized normative data. This was expected for children with milder injuries, as previous outcome research shows few lasting neurocognitive changes following mild brain injury ( Bassett & Slater, 1990; Fay et al., 1994; Jordan, Cannon, & Murdoch, 1992; Levin et al., 1994). Among children with more severe injuries, Fay et al. ( 1994 ) also found a relatively normal level of performance on many cognitive measures. Despite these encouraging results on standardized tests, their severe group displayed significant declines in academic growth relative to matched controls 3 years post-injury. As Fay et al. ( 1994 ) noted, even subtle memory dysfunction may interact with other TBI-related skill deficits (e.g., impaired attention, speed of processing, reasoning, motor skills) to result in poor school performance. Because of the range of memory skills tapped, use of the WRAML in future studies may help to clarify how changes in memory functioning contribute to long-term academic and vocational outcomes.
Clinical Implications
The results of this study suggest several implications for clinical practice. First, memory functioning should be assessed for each individual child following TBI. The large variability on many of the subtests, especially in the Severe group, underscored how difficult it is to predict individual memory functioning from ratings of injury severity alone.
Second, performance patterns across the WRAML subtests must be examined. Use of the Index scores alone has the potential to obscure important strengths and weaknesses that have functional implications. For instance, the Verbal Memory Index may overestimate the functional skills of a child who has intact memory for sentences and number-letter sequences but impaired recall of paragraph length material. Furthermore, delayed WRAML subtests also must be evaluated to capture the full range of a child's performance. The use of normative comparisons for delayed subtests ( Adams et al., 1991) may be helpful clinically to clarify level of performance relative to same-age peers.
Finally, selection of a memory and learning measure to include in a neuropsychological assessment should be based on the goals of the evaluation. If the purpose of assessment is purely diagnostic, a test such as the CVLT-C is highly sensitive to TBI and may be least stressful for the child. However, if the goal is to determine patterns of memory strengths and weaknesses to establish treatment goals and intervention strategies, a more comprehensive evaluation using an instrument such as the WRAML is warranted.
Patterns of performance obtained with the WRAML in this study suggested a number of approaches that may facilitate memory and new learning in children with severe TBI. For instance, they may benefit from repetitions and increased length of exposure to visual material, as well as decreased length and complexity of verbal material. They are likely to respond best on verbal tasks that provide a context, especially when retrieval of this information is prompted using multiple-choice cues. Tasks that involve integration across modalities or skills (auditory-visual as in Sound Symbol Learning or visual-motor as in Design Memory) may require simplification of stimulus presentations or response demands. Although such interventions are suggested by these data, they are merely speculative. The relation among clinical memory measures, everyday recollections, and functional adaptive behaviors is unknown and beyond the scope of this study. There is a clear need for additional research to investigate the ecological validity of clinical memory assessment strategies and to clarify their utility when designing individualized interventions.
Limitations of the Study
This study has several methodological limitations. The small sample size may have limited the power of statistical comparisons and allowed detection of only large effect sizes, especially on several subtests where mean group differences were suggested (e.g., Finger Windows, Visual Learning). In addition, an ascertainment bias might have been introduced because the study was based on a clinic-referred sample rather than consecutive hospital admissions. For instance, children with mild to moderate injuries who experienced significant changes in cognitive functioning may have been more likely to be referred than those who recovered without such sequelae, and on average they were referred relatively later in the recovery process than youth with more severe injuries. There was also a wide range in time since injury in both TBI groups. These clinical referral patterns must be considered when examining these data, as the results may not be generalizable to all children with TBI or to those at different points in recovery.
The clinical referral pattern also resulted in a trend toward more boys in the Severe group (p = .11), which raises the question of whether group differences between the Control and Severe TBI group may be confounded by gender differences. This is a legitimate question, because females are known to perform better on some memory measures than males ( Lezak, 1995). However, there are several reasons that gender does not likely account for the TBI group differences obtained. First, no significant gender differences were reported on the WRAML in the normative sample ( Sheslow & Adams, 1990). Second, one of the subtests most likely to show gender differences is the Verbal Learning subtest ( Lezak, 1995), yet results on this measure closely replicate previous findings that control and severe TBI participants differ in their word list learning abilities even when they are matched by gender (e.g., Yeates et al., 1995). This replication provides assurance that gender distribution is not a significant factor affecting study results.
A related concern is that the Control group for this study was selected from the normal school-age population so that it matched to the TBI sample on important general characteristics, but each control was not matched directly to a specific TBI case. Appropriate control groups are challenging to identify in studies of children with neurological disorders ( Denckla, 1994), but Massagli et al. ( 1996 ) provided data showing that case-control matches may be most sensitive to cognitive changes following TBI. Further research is indicated to replicate and extend the findings of this clinic-based study. Such research will require prospective study of larger samples of children who are matched by case-control pairs to ensure the validity and generalizability of the data.
There are also some limitations inherent in the WRAML itself. First, WRAML index scores may be misleading because they are based on theoretical concepts about memory rather than empirical factor analysis ( Aylward et al., 1995; Callahan et al., 1993; Gioia, 1991; Haut et al., 1992; Sheslow & Adams, 1990). Compared to an instrument like the CVLT-C, the WRAML provides less sophisticated information about processes involved in memory and learning (e.g., no recognition cueing on the learning subtests). Delay intervals are not carefully standardized in that they can vary from 20 to 40 min depending on each child's individual rate of subtest completion. There may be ceiling effects for older children on some of the subtests that limit test sensitivity ( Sheslow & Adams, 1990). Finally, although the WRAML allows for a comprehensive assessment of memory skills, the sheer length of the test can cause validity problems for some children due to fatigue. If the WRAML is to be used either clinically or as a research instrument, these drawbacks must be weighed against the advantages of collecting data about a range of memory and learning skills in children.
Dr. Janet E. Farmer, Ph.D. is an Associate Professor in the Department of Physical Medicine and Rehabilitation (PM&R). She also has a joint appointment in the Department of Child Health and an adjunct appointment in the Department of Psychology. She is licensed as a Psychologist/Health Services Provider in the state of Missouri. Her primary clinical responsibilites involve providing behavioral health services for individuals with chronic and disabling health conditions, particularly children with neurobehavioral disorders. As the Director of Pediatric Neuropsychology for PM&R/Child Health, Dr. Farmer conducts neuropsychological assessments of children with a wide range of medical, learning, behavioral, and developmental problems (for example, attention deficit hyperactivity disorder, learning disabilities, developmental delays, acquired brain injury, autism, and other health conditions such as seizures or cancer). These services are requested by physicians, schools, families, and other community-based agencies such as the Division of Family Services, First Steps, and the Bureau of Special Health Care Needs. The goal of these assessments is to consult with families and other service providers in order to create an effective plan of care for the individual child.
REFERENCES
Adams, W., Sheslow, D. V., & Wilkinson, G. ( 1991, August ). Assessment of memory and learning in children: Supplemental norms. Paper presented at the 11th annual meeting of the National Academy of Neuropsychology, Dallas, TX.
Aylward, G. P., Gioia, G., Verhulst, S. J., & Bell, S. ( 1995 ). "Factor structure of the Wide Range Assessment of Memory and Learning in a clinical population". Journal of Psychoeducational Assessment, 13, 132-142.
Bassett, S. S., & Slater, E. J. ( 1990 ). "Neuropsychological function in adolescents sustaining mild closed head injury". Journal of Pediatric Psychology, 15, 225-236.
Begali, V. ( 1992 ). Head injury in children and adolescents. Brandon, VT: Clinical Psychology Publishing Company.
Callahan, T. S., Haut, J. S., Haut, M. W., & Franzen, M. D. ( 1993, August ). Confirmatory factor analysis of the Wide Range Assessment of Memory and Learning (WRAML). Presented at the 101st annual meeting of the American Psychological Association, Toronto, Canada.
Chadwick, O., Rutter, M., Brown, G., Shaffer, D., & Traub, M. ( 1981 ). "A prospective study of children with head injuries: II. Cognitive sequelae". Psychological Medicine, 11, 49-61.
Chapman, S. B. ( 1995 ). "Discourse as an outcome measure in pediatric head-injured populations". In S. H. Broman & M. E. Michel (Eds.), Traumatic head injury in children (pp. 95-116). New York: Oxford University Press.
Daniel, W. W. ( 1990 ). "Applied nonparametric statistics". Boston: PWS-KENT Publishing Company.
Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. ( 1994 ). Manual for The California Verbal Learning Test for Children. New York: Psychological Corporation.
Denckla, M. B. ( 1994 ). "Interpretations of a behavioral neurologist". In S. H. Broman & J. Grafman (Eds.), Atypical cognitive deficits in developmental disorders: Implications for brain function (pp. 283-295). Hillsdale, NJ: Lawrence Erlbaum Associates Inc.
Dikmen, S. S., Donovan, D. M., Loberg, T., Machamer, J. E., & Temkin, N. R. ( 1993 ). "Alcohol use and its effects on neuropsychological outcome in head injury". Neuropsychology, 7, 296-305.
Donders, J. ( 1993 ). "Memory functioning after traumatic brain injury in children". Brain Injury, 7, 431-437.
Ewing-Cobbs, L., Fletcher, J. M., Levin, H. S., Hastings, P. Z., & Francis, D. J. ( 1996, February ). Assessment of injury severity following closed head injury in children: Methodological issues. Poster presented at the 24th annual meeting of the International Neuropsychology Society, Chicago.
Ewing-Cobbs, L., Levin, H. S., Fletcher, J. M., Miner, M. E., & Eisenberg, H. M. ( 1990 ). "The children's orientation and amnesia test: Relationship to severity of acute head injury and to recovery of memory". Neurosurgery, 27, 683-691.
Fay, G. C., Jaffe, K. M., Polissar, N. L., Liao, S., Rivara, J. B., & Martin, K. M. ( 1994 ). "Outcome of pediatric traumatic brain injury at three years: A cohort study". Archives of Physical Medicine and Rehabilitation, 75, 733-741.
Fletcher, J. M., Ewing-Cobbs, L., Francis, D. J., & Levin, H. S. ( 1995 ). "Variability in outcomes after traumatic brain injury in children: A developmental perspective". In S. H. Broman & M. E. Michel (Eds.), Traumatic head injury in children (pp. 3-21). New York: Oxford University Press.
Fletcher, J. M., Levin, H. S., & Butler, I. J. ( 1995 ). "Neurobehavioral effects of brain injury on children: Hydrocephalus, traumatic brain injury, and cerebral palsy". In M. C. Roberts (Ed.), Handbook of pediatric psychology, Second edition (pp. 362-383). New York: Guilford.
Gioia, G. ( 1991, February ). Re-analysis of the factor structure of the Wide Range Assessment of Memory and Learning: Implications for clinical interpretation. Poster presented at the 19th annual meeting of the International Neuropsychological Society, San Antonio, TX.
Hannay, H. J., & Levin, H. S. ( 1989 ). "Visual continuous recognition memory in normal and closed-head injured adolescents". Journal of Clinical and Experimental Neuropsychology, 11, 444-460.
Haut, J. S., Haut, M. W., Callahan, T. S., & Franzen, M. D. ( 1992, November ). Factor analysis of the Wide Range Assessment of Memory and Learning (WRAML) scores in a clinical sample. Paper presented at the 12th annual meeting of the National Academy of Neuropsychology, Pittsburgh, PA.
Jordan, F. M., Cannon, A., & Murdoch, B. E. ( 1992 ). "Language abilities of mildly closed head injured children 10 years post-injury". Brain Injury, 6, 39-44.
Kaufmann, P. M., Fletcher, J. M., Levin, H. S., Miner, M. E., & Ewing-Cobbs, L. ( 1993 ). "Attentional disturbance after closed head injury". Journal of Child Neurology, 8, 348-353.
Klonoff, H., Clark, C., & Klonoff, P. S. ( 1993 ). "Long-term outcome of head injuries: A 23 year follow up study of children with head injuries". Journal of Neurology, Neurosurgery, and Psychiatry, 56, 410-415.
Knights, R. M., Ivan, L. P., Ventueyra, E. C. G., Bentivoglio, C., Stoddart, C., Winogron, W., Bawden, H. N. ( 1991 ). "The effects of head injury in children on neuropsychological and behavioral functioning". Brain Injury, 5, 339-351.
Levin, H. S., Culhane, K. A., Fletcher, J. M., Mendelsohn, D. B., Lilly, M. A., Harward, H., Chapman, S. B. , Bruce, D. A., Bertolino-Kusnerik, L., & Eisenberg, H. M. ( 1994 ). "Dissociation between delayed alternation and memory after pediatric head injury: Relationship to MRI findings". Journal of Child Neurology, 9, 81-89.
Levin, H. S., & Eisenberg, H. M. ( 1979 ). "Neuropsychological outcome of closed head injury in children and adolescents". Child's Brain, 5, 281-292.
Levin, H. S., Eisenberg, H. M., Wigg, N. R., & Kobayashi, K. ( 1982 ). "Memory and intellectual ability after head injury in children and adolescents". Neurosurgery, 11, 668-673.
Levin, H. S., High, W. M., Ewing-Cobbs, L., Fletcher, J. M., Eisenberg, H. M., Miner, M. E., & Goldstein, F. C. ( 1988 ). "Memory functioning during the first year after closed head injury in children and adolescents". Neurosurgery, 22, 1043-1052.
Lezak, M. ( 1995 ). Neuropsychological assessment, 3rd edition. New York: Oxford University Press.
Massagli, T. L., Jaffe, K. M., Fay, G. C., Polissar, N. L., Liao, S., & Rivara, J. B. ( 1996 ). "Neurobehavioral sequelae of severe pediatric traumatic brain injury: A cohort study". Archives of Physical Medicine and Rehabilitation, 77, 223-231.
Savage, R. C., & Wolcott, G. F. ( 1994 ). Educational dimensions of acquired brain injury. Austin, TX: PRO-ED.
Sheslow, D. V., & Adams, W. ( 1990 ). Manual for The Wide Range Assessment of Memory and Learning. Wilmington, DE: Jastak.
Teasdale, G., & Jennett, B. ( 1974 ). "Assessment of coma and impaired consciousness: A practical scale". Lancet, ii, 81-84.
Wechsler, D. ( 1981 ). Wechsler Adult Intelligence Scale-Revised. San Antonio: The Psychological Corporation.
Wechsler, D. ( 1987 ). Wechsler Memory Scale-Revised. San Antonio: The Psychological Corporation.
Wechsler, D. ( 1989 ). Wechsler Preschool and Primary Scale of Intelligence-Revised. San Antonio: The Psychological Corporation.
Wechsler, D. ( 1991 ). Wechsler Intelligence Scale for Children-III. San Antonio: The Psychological Corporation.
Williams, J., & Haut, J. S. ( 1995 ). "Differential performances on the WRAML in children and adolescents diagnosed with epilepsy, head injury, and substance abuse". Developmental Neuropsychology, 11, 201-213.
Winogron, H. W., Knights, R. M., & Bawden, H. N. ( 1984 ). "Neuropsychological deficits following head injury in children". Journal of Clinical Neuropsychology, 6, 269-286.
Yeates, K. O., Blumenstein, E., Patterson, C. M., & Delis, D. ( 1995 ). "Verbal learning and memory following pediatric closed-head injury". Journal of the International Neuropsychological Society, 1, 78-87.