Abstract

Traumatic brain injury (TBI) often results in long-term negative effects in attention, memory, perception, language, and executive functioning. Children and adolescents are the most vulnerable as TBIs are the leading cause of death and disability for this age group. Despite these high proportions and detrimental effects, few studies have utilized a developmentally appropriate, standardized measure to assess executive functioning within a pediatric TBI population. The current study compared children and adolescents who had sustained a TBI with a non-injured, matched control group on executive functioning using the Comprehensive Trail Making Test (CTMT). Data analyses revealed significant differences between groups on the CTMT Composite Index, each individual trail, and a combination of trails but with no within-group differences. Confirmatory factor analysis revealed that a single-factor model was a better fit for the present sample of TBI participants than the two-factor model evident within the CTMT standardization sample.

Introduction

Traumatic brain injuries (TBIs) are frequently referred to as “silent” due to the damaging, yet often invisible long-term effects demonstrated through memory loss, attentional disorders, and executive functioning deficits (Langlois, Rutland-Brown, & Thomas, 2006). In the United States alone, an estimated 1.4 million TBIs occur each year resulting in 50,000 deaths and 235,000 hospitalizations (Langlois et al., 2006). For children 14 years old and younger, TBIs are the leading cause of death and disability with an average occurrence of 475,000 brain injuries every year (Jankowitz & Adelson, 2006). A childhood TBI often results in life-long deficits. Persistent cognitive impairments affect basic psychological processes such as attention, perception, language, memory, and abstract reasoning (Farmer, Clippard, Luehr-Wiemann, Wright, & Owings, 1997). Often, the most frequently identified disrupted cognitive skill is “executive functioning.” Ylvisaker and DeBonis (2000) state that “executive system impairment is often the most debilitating disorder after TBI” (p. 35). However, as Maricle, Johnson, and Avirett (2010) note, there is no universally accepted definition of executive functioning, nor is there a mutually agreed upon list of the cognitive components which comprise executive functions. Numerous subdomains of cognition have been implicated in executive functioning, including set-shifting, problem solving, abstract reasoning, planning, organization, goal setting, working memory, inhibition, mental flexibility, initiation, attentional control, and behavioral regulation (Baron, 2004; Gioia, Isquith, & Guy, 2001).

Depending on the type of injury, the location of the frontal lobes is highly susceptible to the detrimental effects of a TBI. The bony protuberances at the base of the frontal lobes, and the frontal–temporal regions inside the skull, increase the vulnerability of the frontal lobes (Ylvisaker & DeBonis, 2000). An insult to the frontal lobes at any age can have significant consequences later in maturity given the protracted developmental process which continues through adolescence. Unfortunately, children may often “grow” into their deficits depending on what type of skill deficit is being examined. The individual's deficits may not be demonstrated until the injured brain area is needed to support a developmental transition (Ylvisaker & DeBonis, 2000). Given the importance of executive functioning on daily living skills, academic success, and social interactions, it is imperative to accurately assess these deficits in children and adolescents. However, of the measures specifically designed to address neuropsychological deficits after a brain trauma, few have been validated or standardized for use with children. A large majority of the test batteries used with children are downward extensions of measures designed for adult populations. These types of measures typically assess the presence or absence of an executive skill and rarely take into account the skill's developmental trajectory and how it influences the manifestation of the skill at different ages.

One widely used adult neuropsychological measure, which has seen multiple variations, is Trails A and Trails B. Reitan (1971) combined Trails A and B to comprise The Trail Making Test (TMT). The TMT is used to assess a number of neurocognitive abilities such as psychomotor speed, complex attention, visual scanning, and mental flexibility. It has been repeatedly demonstrated to be sensitive to brain injury in adults (Boll, Berent, & Richards, 1977; Jaffe et al., 1993; Reitan, 1955, 1971). Reitan and Wolfson (1992) followed up and published a TMT for use with children. However, as Neyens and Aldenkamp (1996) revealed that the child version did not demonstrate the same sensitivity as the adult version. Neyens and Aldenkamp (1996) found test–retest reliability coefficients of .33 (Part A) and .56 (Part B) for the child and adolescent version of the TMT. These coefficients were derived from a sample of 59 typically developing children between the ages of 4 and 12 who were tested a total of three times with 6-month intervals between testing sessions.

Adaptations of the original Trails A and B have been utilized in various forms within neuropsychological batteries such as the Delis–Kaplan Executive Function System (D-KEFS; Delis, Kaplan, & Kramer, 2001). The D-KEFS is designed to measure specific aspects of executive functioning through nine independent tests, one of which is a trails measure. However, validity evidence for the D-KEFS is scarce in regard to children and relies on the established validity of previous trails measures (Maricle et al., 2010). Given the low reliabilities of the adult and children's versions of the TMT, frequent use of adult measures for children, and limited validity studies with children, The Comprehensive Trail Making Test (CTMT; Reynolds, 2002) was designed to include children in the standardization and to off-set many of the psychometric deficiencies of the original TMT. The CTMT utilizes five trails instead of two and provides a nationally normed sample that includes children down to the age of 11. The five trails of the CTMT increase at various levels of difficulty to measure resistance to distraction, inhibition, task switching, and cognitive flexibility (Reynolds, 2002). On the CTMT, Trail 1 is equivalent to Trails A and Trail 5 is similar to Trails B of the original TMT. The CTMT has three additional trails which allows for increased specificity to help determine particular areas of deficit. According to Reynolds (2002), the CTMT is specifically useful for detecting “frontal lobe deficits, problems with psychomotor speed, visual search and sequencing, and attention; and impairments in set-shifting… ” (p. 5).

Despite the benefits of using the CTMT in assessing executive functioning, there have been relatively few studies using this measure after a childhood TBI. Armstrong, Allen, Donohue, and Mayfield (2008) assessed 60 children and adolescents to examine the sensitivity of the CTMT. The sample included 30 children who had sustained a TBI and 30 healthy matched controls with an age range of 11- to 19 years old. Results indicated that the TBI group performed ∼2 SD below the comparison group on all five trails and 2 SD below on the Composite Index. Given the small sample size, factor analysis of the CTMT was not possible in this study. Instead, the pattern of correlations between the trails of the TBI and control group samples was examined. A general pattern was noted with the TBI sample demonstrating significantly higher correlations than the control group. Armstrong and colleagues (2008) conclude that the CTMT maintains strong criterion validity and sensitivity, as well as good predictive discrimination, when used to assess older children and adolescents who have experienced a TBI. The authors note that additional research is needed to validate these results with a larger sample size, which would also permit a factor analysis to determine the underlying variables.

The purpose of the current study was to replicate and expand upon the Armstrong and colleagues (2008) research. In the standardization sample of the CTMT, a factor analysis identified two primary factors. The first factor included Trails 1 through 3 and was termed “Simple Sequencing.” The second factor was composed of Trails 4 and 5 and entitled “Complex Sequencing” (Reynolds, 2002). Armstrong and colleagues (2008) suggested that their preliminary findings warranted further examination of the CTMT factor structure within clinical populations. The current study examined whether a similar two-factor model, as evidenced in the standardization sample of the CTMT, was also present when using a TBI sample. Additional investigation examined whether there was a significant difference between the TBI group and the control group on the CTMT Composite Index score, as well as between groups on each individual trail and various groupings of trails.

Materials and Methods

Participants

One hundred and sixty children and adolescents between the ages of 11- and 19 years old participated in this study (Table 1). The sample included 80 children and adolescents who had sustained a TBI and 80 with no known history of head trauma. Data on the children and adolescents with TBI were obtained from the records of a pediatric specialty hospital in Dallas, Texas. The control group data were obtained from the standardization sample of the CTMT. There were a total of 116 men and 44 women with the clinical and control groups each containing 58 men and 22 women, respectively. Of the 160 participants, 60% were Caucasian (n = 96), 17.5% were African-American (n = 28), 13.1% were Hispanic American (n = 21), 1.3% were Asian-American (n = 2), and 1.3% were Native American (n = 2). Ethnicity was missing from 6.8% of the cases (n = 11) for the overall sample.

Table 1.

Demographic characteristics of the clinical and control groups

 Clinical group Control group 
Gender 
 Male 58 58 
 Female 22 22 
Age at testing 
 11–13 21 21 
 14–16 38 38 
 17–19 21 21 
Type of injury 
 Closed 73 — 
 Open — 
Cause of injury 
 Motor vehicle accident 35 — 
 Pedestrian hit by a car 11 — 
 Four-wheeled vehicle 11 — 
 Fall 10 — 
 Gunshot wound — 
 Skiing accident — 
 Other incidents — 
Total 80 80 
 Clinical group Control group 
Gender 
 Male 58 58 
 Female 22 22 
Age at testing 
 11–13 21 21 
 14–16 38 38 
 17–19 21 21 
Type of injury 
 Closed 73 — 
 Open — 
Cause of injury 
 Motor vehicle accident 35 — 
 Pedestrian hit by a car 11 — 
 Four-wheeled vehicle 11 — 
 Fall 10 — 
 Gunshot wound — 
 Skiing accident — 
 Other incidents — 
Total 80 80 

Of the 80 participants in the clinical group, all were considered moderate-to-severe head injuries, with 73 (91%) having incurred a closed head injury and 7 (9%) who had sustained an open head injury. Those in the TBI clinical sample varied in regard to the cause of the head injury. Thirty-five of the participants (44%) had been involved in a motor vehicle accident, 11 (14%) were pedestrians hit by a car, 11 (14%) were injured in an accident with a four-wheeled vehicle, 10 individuals (12%) sustained injuries due to a fall, 6 participants (7%) sustained injuries due to a gunshot wound, and 1 individual (1%) was involved in a skiing accident (Table 1). The remaining 6 participants (7%) sustained injuries due to other incidents. The average time between the injury and the neuropsychological evaluation was 20 months, with the median time between injury and evaluation being 11 months with a range of 5 to 115 months. Control group participants were drawn from the normative sample of the CTMT (Reynolds, 2002) and randomly matched, using a random number generator, on age and gender to the TBI clinical group.

Measures

The assessment of executive functioning in childhood and adolescence is a daunting task. Specifically within a TBI population, the importance of utilizing a measure that is standardized for children and adolescents is crucial due to the developmental shifts taking place during this time period. Use of the CTMT (Reynolds, 2002) allows for the measurement of executive control in the areas of selective and divided attention, mental set-shifting, inhibition of responses, and sequencing abilities.

Reliability and validity data on the CTMT is generally considered adequate. Gray (2006) noted that the reliability data for the CTMT are within acceptable limits; however, the reliability may not be high enough for diagnostic purposes. Servesko, Smith, and Edwards (2006) established convergent, divergent, and discriminant validity for the CTMT and suggested that the evidence generally supports the validity of the CTMT.

Trail 1 of the CTMT is similar to Trails A in the original version where the examinee is timed while connecting a line from the numbers 1 through 25. The nature of this task taps into an individual's ability to sustain attention while utilizing visual scanning and sequencing abilities (Moses, 2004). Trails 2 and 3 build on the original Trails A; however, the difficulty is increased due to additional distractor circles. This requires the examinee to resist and inhibit responses to the distractions. Reynolds (2002) states that both the simple (Trail 2) distracters and the complex (Trail 3) distracters increase the complexity of the selective and sustained attention necessary to be successful on this task. Trail 4 on the CTMT adds a further cognitive demand by utilizing numbers in both the numeric form (e.g., 2, 4) and the written word (e.g., two, four). The numbers are represented arbitrarily between the English word and the Arabic symbols which requires the examinee to inhibit responses and shift-set in a random pattern. Trail 5 necessitates switching between the numbers 1 through 13 and the letters A through L in an alternating sequence. Empty distractor circles are added to the visual array to increase the task-switching difficulty (Reynolds, 2002). The CTMT also differs from the original trails measure by increasing the complexity of the spatial arrangement. Rather than following a path that proceeds gradually out from the center, the CTMT varies the directionality of the target stimuli in their given sequence. This enhances the complexity of the CTMT to provide an assessment that more effectively measures executive functioning skills such as visual-spatial scanning, sequencing, cognitive flexibility, processing speed, and perceptual-motor integration (Moses, 2004).

The response time on each trail is recorded and errors are corrected as they occur. The types of errors are not scored; however, errors affect the overall performance, in that they require additional time for correction to take place. The results of the five trails combine to provide an overall Composite Index. Norms for each trail and the Composite Index are presented in the form of T-scores, which have a mean of 50 and a standard deviation of 10. Reynolds (2002) states that the individual trails T-scores are “age-corrected deviation scaled scores based on the cumulative frequency distributions of the raw scores” (p. 19). These scores are calculated based on the percentages linked to the raw scores in the standardization sample. Both the raw scores and the standard deviations were computed at 1-year intervals for the ages of 11- through 19 years old to account for developmental influences. The adult age ranges were based on 10-year increments (e.g., 20–29, 30–39, etc.). At each age interval, the raw scores were converted to T-scores to allow for consistency across age levels.

Procedures

The participants in the TBI sample had experienced a significant insult to their brain as confirmed by MRI or CT scans. Each individual was treated at a pediatric restorative facility post-injury. The CTMT was given as part of a comprehensive neuropsychological assessment which evaluated intellectual abilities, academic achievement, attentional processes, visual-motor skills, and behavioral regulation. The order of the assessment measures was dependent on the needs of the child taking into consideration factors such as fatigue, need for breaks, and the child's overall response to the assessment situation. The neuropsychological evaluation was completed by professionals trained to reliably administer the standardized procedures of each measure. The data for the TBI clinical group were collected from the individual records, coded and entered into a pre-established database. The data for the control group were randomly selected by the CTMT publisher after being matched to the clinical group on the factors of gender and age.

Results

Data Analysis and Hypotheses

The initial research question sought to determine if there was a significant difference between the TBI group and the control group on the overall composite score of the CTMT. An analysis of variance (ANOVA) was utilized to compare the clinical group with the control group on the Composite Index. Secondly, the data were analyzed to determine if significant differences existed between the TBI group and the control group on each individual trail, as well as between groupings of trails (e.g., Trails 1–3 vs. Trails 4–5). This was examined through the use of a repeated-measures multivariate analysis of variance (MANOVA). Finally, confirmatory factor analysis was used to validate or refute whether the two-factor model (i.e., Trails 1–3 “Simple Sequencing” factor and Trails 4–5 “Complex Sequencing”) that was obtained with the entire standardization sample would remain consistent with a TBI cohort. An additional principal component analysis compared the factor structure of the matched control group to determine if it was consistent with the TBI group. The Varimax rotation method chosen was with Kaiser Normalization for an orthogonal factor matrix.

Results of the univariate ANOVA indicated a significant difference when comparing the clinical group and the control group on the overall composite score, F(1, 158) = 26.52, p < .0001. A repeated-measures MANOVA was utilized to examine the differences between groups on each trail. As depicted in Table 2, significant differences were evident across all five trails. Results revealed that all F-values were significant at the p < .0001 level. No significant main effect of trail was noted, F(4, 155) = 0.974, p = .424, as well as no significant interaction of group and trail, F(4, 155) = 0.357, p = .839. A significant effect with the control group was demonstrated, F(1, 158) = 29.306, p < .0001.

Table 2.

Means and SDs of the clinical and control groups on each individual CTMT Trail

 Control group (n = 80)
 
Clinical group (n = 80)
 
Univariate 
 Mean SD Mean SD F 
CTMT Trail 1 45.09 11.15 37.45 14.57 13.87* 
CTMT Trail 2 46.93 11.19 37.43 14.24 21.99* 
CTMT Trail 3 46.14 11.32 37.39 13.189 20.28* 
CTMT Trail 4 45.96 11.53 37.08 14.341 18.66* 
CTMT Trail 5 47.00 8.34 38.34 11.69 29.09* 
 Control group (n = 80)
 
Clinical group (n = 80)
 
Univariate 
 Mean SD Mean SD F 
CTMT Trail 1 45.09 11.15 37.45 14.57 13.87* 
CTMT Trail 2 46.93 11.19 37.43 14.24 21.99* 
CTMT Trail 3 46.14 11.32 37.39 13.189 20.28* 
CTMT Trail 4 45.96 11.53 37.08 14.341 18.66* 
CTMT Trail 5 47.00 8.34 38.34 11.69 29.09* 

Note: CTMT = Comprehensive Trail Making Test. *All F-values significant at the p < .0001 level. Means are from a repeated-measures multivariate analysis of variance. No significant main effect of trail, F(4, 155) = 0.974, p = .424. No significant interaction of group and trail, F(4, 155) = 0.357, p = .839.

The differences between groupings of trails were also analyzed with a repeated-measures MANOVA. The means of Trails 1–3 and Trails 4–5 for the clinical group and the control group were compared within each sample of individuals, as well as across groupings. As can be seen in Table 3, significant differences were evident between groups but not within each group. When the CTMT Trails 1–3 were considered as a set, significant differences were observed between the clinical and the control groups. This demonstrated that the control group consistently scored higher on the CTMT Trails 1 through 3, which measured simple sequencing abilities. Significance was also evident between groups when the CTMT Trails 4–5 were combined as a set. This illustrates the differences between groups on tasks which tap into executive functioning skills such as set-shifting and cognitive flexibility. However, when comparing the CTMT Trails 1–3 and Trails 4–5 set within the TBI group, no differences were noted, F(1, 79) = 0.122, p = .728. Nor were differences observed within the control group between the CTMT Trails 1–3 and Trails 4–5, F(1, 79) = 0.188, p = .666.

Table 3.

Univariate F and means (SD) of the control and clinical groups on the CTMT groupings of trails

 Mean SD Univariate
 
   F p 
CTMT Groupings of Trails     
 Trails 1–3   23.58 ≤.0001 
  Control group 46.05 9.51   
  Clinical group 37.42 12.73   
 Trails 4–5   28.57 ≤.0001 
  Control group 46.48 8.56   
  Clinical group 37.71 11.93   
 Mean SD Univariate
 
   F p 
CTMT Groupings of Trails     
 Trails 1–3   23.58 ≤.0001 
  Control group 46.05 9.51   
  Clinical group 37.42 12.73   
 Trails 4–5   28.57 ≤.0001 
  Control group 46.48 8.56   
  Clinical group 37.71 11.93   

The latent factor structure within the TBI group and within the matched control group was also examined to determine if the two-factor model demonstrated within the entire CTMT standardization sample would remain true in the clinical sample. Analysis of the current sample revealed a single-factor solution. Table 4 displays the factor loading of each trail as evidenced through the principal component factor analysis. The single-component solution accounted for 76.35% of the overall variance. A forced factor analysis revealed that a two-factor solution accounted for only 9% more variance (84.85%). By adding a second component, the total variance accounted for was not significantly increased; therefore, a single-factor solution was a better fit for the sample of TBI participants. An additional analysis was conducted which examined the latent factor structure within the matched control group. As previously stated, analysis of the entire CTMT standardization sample yielded a two-factor model. This final analysis was conducted to determine if the matched control sample would yield a one-factor model, as was evident in the TBI group, or a two-factor structure similar to the CTMT standardization sample. Table 5 displays the factor loadings of each trail for the principal component factor analysis. The single-component solution accounted for 62.97% of the overall variance. A forced factor analysis revealed that a two-factor solution accounted for 14% more variance (76.82%). By adding a second component, the total variance accounted for was significantly increased; therefore, a two-factor solution was a better fit for the sample of control participants.

Table 4.

Principal component analysis of the TBI sample

CTMT trails Single-component factor 
Trail 1 .85 
Trail 2 .89 
Trail 3 .91 
Trail 4 .87 
Trail 5 .85 
Percentage of variance 76.35 
CTMT trails Single-component factor 
Trail 1 .85 
Trail 2 .89 
Trail 3 .91 
Trail 4 .87 
Trail 5 .85 
Percentage of variance 76.35 
Table 5.

Principal component analysis of the control sample

CTMT trails Component factors
 
 Factor 1 Factor 2 
Trail 1 .82 .27 
Trail 2 .84 .29 
Trail 3 .79 .30 
Trail 4 .35 .82 
Trail 5 .25 .88 
Percent of variance 62.97 13.87 
Cumulative percentage 62.97 76.82 
CTMT trails Component factors
 
 Factor 1 Factor 2 
Trail 1 .82 .27 
Trail 2 .84 .29 
Trail 3 .79 .30 
Trail 4 .35 .82 
Trail 5 .25 .88 
Percent of variance 62.97 13.87 
Cumulative percentage 62.97 76.82 

Discussion

The consequences of a pediatric TBI often persist throughout an individual's lifetime. Deficits are usually evidenced through memory difficulties, attentional issues, and executive dysfunction. Despite the prevalence and long-term detrimental effects, few studies have specifically addressed the executive functioning deficits in a pediatric TBI population using a standardized measure designed to account for the interaction of age and development. To this end, the present study was conducted to compare executive functioning in children and adolescents who have sustained a TBI with non-injured controls using the CTMT. The current investigation utilized a larger sample size and confirmed previous results (Armstrong et al., 2008) of significance on the CTMT Composite Index and individual trails between groups. This supports the conclusion that the CTMT demonstrates strong sensitivity and good overall utility for use with a child and adolescent TBI population.

When the CTMT trails measures were combined into groupings of Trails 1–3 and Trails 4–5, significant differences were noted between the clinical and the control groups. When examining within-group differences, no significance was demonstrated between the clinical and control samples on the combined mean of Trails 1–3 and Trails 4–5. The mean scores were slightly higher on Trails 4–5 versus Trails 1–3 for each group but not to a level of significance.

The current investigation extended the previous research by attempting to confirm the two-factor model that was reported in the CTMT standardization sample with a group of TBI participants. It was hypothesized that the two-factor model would also be present in the TBI group; however, this was not evidenced by the data. The analysis revealed that a single-component solution accounted for 76% of the overall variance. A forced factor analysis revealed that a two-factor solution accounted for only 9% more variance. By adding this second component, the total variance accounted for did not significantly increase. These results demonstrate that a single-factor solution was the best fit for the current TBI sample. This primary factor is likely due to the similarity between each trail measure. Although Trails 4–5 increases the demand on executive functioning due to the switching component, all five of the CTMT trails tap into planning abilities, sequencing, speed of processing, and visual scanning. It is possible that in a clinical TBI population, the CTMT is sensitive to the generalized brain dysfunction, but not necessarily the specific executive functions that the CTMT might be able to differentiate in a typically developing child or adolescent. Another hypothesis is that one specific executive function, such as processing speed, may be impaired in the clinical sample and this primary deficit may account for the low performance across all five trails.

It is possible that the differences seen between the one-factor results of the current study versus the two-factor model in the standardization sample could be a function of the differing sample size. The 80 participants used for the current analysis is on the low end of the recommended sample size for a confirmatory procedure. The same analysis was repeated using the matched sample control group to determine if the data would reveal a single-component structure, as in the TBI sample, or a two-factor solution evidenced in the entire CTMT standardization sample. Results demonstrated that a single component accounted for 63% of the overall variance. When adding an additional factor, 14% more of the overall variance was accounted for with a two-factor solution. Factors are considered “pure” when the items load onto a component at .45 or higher, and are below .33 on other factors. As can be seen in the current analysis, factor 1 has loadings above that level for Trail 1 (.82), Trail 2 (.84), and Trail 3 (.79). Factor 2 is composed of Trail 4 (.80) and Trail 5 (.88). These high loadings demonstrate that a two-factor model is a better fit for the matched sample control group. This result coincides with the factor structure evidenced in the CTMT standardization sample.

Another potential cause for the differing results could be the basal and ceiling effects of the test measure used. On the CTMT, the lowest possible T-score that an individual could obtain is 17. However, six individuals had results indicating scores below this basal point. Although this hindrance likely did not affect the results of the comparisons between the clinical and the control groups, it may have influenced the comparisons within the TBI group. Greater differences between the groupings of Trails 1–3 versus Trails 4–5 in the clinical sample may have been evident if a lower basal was possible with the CTMT.

One possible limitation of the current study was the lack of distinction between moderate and severe TBI. GCS scores were not available for all participants in the clinical group. Injury severity has been shown to be predictive of long-term outcomes (Anderson, Morse, Catroppa, Haritou, & Rosenfeld, 2004). Those with a moderate TBI may present a different profile of executive functioning than individuals with a severe TBI. By distinguishing between the levels of mild, moderate, and severe brain injury, future research could provide more specific cognitive profiles of strengths and weaknesses post-injury.

Another possible moderating variable that the study did not control for, and which could potentially have influenced participants' performance, was medication. Although the type and amount of medication prescribed was not available, it is possible that the subjects' ability to remain on task could have been mediated due to medication and thus resulted in an elevated score. This interaction effect could inflate the results due to a medication rather than measuring a true executive functioning deficit.

An additional limitation to this study includes the participants' age at injury. As discussed previously, the importance of developmental and contextual factors cannot be minimalized. Anderson and Yeates (2010) state that the age at injury and the previous skills attained are crucial factors when determining future recovery. The present research examined children and adolescents between the ages of 11- and 19 years old due to the age of those standardized with the CTMT. Future research should expand the age range of participants by utilizing children younger than age 11 and examining the developmental trajectory of executive functioning after a head injury (Taylor & Alden, 1997). In addition, this research could be extended utilizing other measures of executive functioning such as the D-KEFS (Delis, Kaplan, & Kramer, 2001) to expand the age range and allow for a broader developmental profile. An additional expansion could include other subpopulations using the CTMT as executive dysfunction is noted in other disabilities such as autism. While the CTMT showed strong reliability and validity in its development (Gray, 2006; Moses, 2004; Reynolds, 2002) with an LD population, a gifted subgroup, and a CVA sample, additional external research is needed to validate these results. Two outside studies have supported the CTMT's validity (Servesko et al., 2006) and its overall utility and specificity (Armstrong et al., 2008); yet additional research is needed to confirm these results with other subpopulations.

It would also be beneficial to have additional research related to the interaction effects of development and the time of injury. As has been noted in several studies previously (Anderson, 2002; Blair, Zelazo, & Greenberg, 2005; Diamond, Carlson, & Beck, 2005; Espy et al., 2001; Zelazo, Muller, Frye, & Marcovitch, 2003), the developmental trajectory of executive functioning exhibits a stair-stepped path. The maturational course varies depending on the specific skill being examined. Future research could include a wider age range of individuals and use a measure that could accurately tap into the emerging skills at each developmental phase. Longitudinal research in this area would also be a valuable asset to the current knowledge base. Given the rapid initial improvements that often take place following a TBI, as well as the deficits that may emerge post-injury, long-term studies are needed to chart these changes. Depending on the age when the injury occurred, as well as severity and location of the injury, the cognitive and behavioral sequelae may differ greatly (Brookshire, Levin, Song, & Zhang, 2004). The time period since the injury is also important as children often “grow” into their deficits as they mature (Anderson & Yeates, 2010). As additional educational and environmental demands are placed on the child, developmental milestones may not be met due to an earlier injury.

The effect of gender is another variable that may also play a role developmentally with children and adolescents. Gender in adults does not appear to influence their performance on the trails tasks (Hamdan & Hamdan, 2009; Tombaugh, 2004; Zalonis et al., 2010), although limited research is available for children and adolescents. Given that there were no gender differences noted in the CTMT standardization sample and no differences in adult populations, it is not likely that children and adolescents would demonstrate gender differences on trails tasks. In this study, the clinical sample of girls was deemed too small to effectively determine significant results based on gender. In the future, research that includes all of these influential factors such as age at injury, potential gender differences, as well as severity and location of injury, over a long period of time will better be able to parcel out the variables that best predict positive and negative outcomes.

The current investigation utilized a larger sample size to confirm previous results (Armstrong et al., 2008) that the CTMT is appropriate for use with the pediatric TBI population for detecting executive functioning deficits in the areas of sequencing, visual scanning, planning, and cognitive flexibility. Additionally, this research enhanced the previous results by demonstrating that a one-factor model was a better fit for a sample of moderate-to-severe TBI participants. This differs from the two-factor model demonstrated in the CTMT standardization sample. The evidence for the one-factor model with the TBI sample is likely due to the executive functioning demands necessary across all five CTMT trails. One primary executive functioning deficit may be affecting performance on all trails or the CTMT could be measuring an overall brain dysfunction as a result of the TBI. Results indicate that the CTMT is effective for use with children and adolescents who have sustained a TBI in contrast to typically developing children and adolescents to measure deficits including planning abilities, sequencing, speed of processing, and visual scanning.

Given that TBIs are the leading cause of death and disability in children and adolescents (Jankowitz & Adelson, 2006), the need for accurate assessments like the CTMT to measure deficits is crucial in planning and educating the individual post-injury. A child may return to school and an adult to the workplace with no visible effects of an injury. However, all too often underlying deficits remain to cause difficulties. Others may not understand the differences in personality or the emerging deficits that can become evident as the injured individual attempts to return to “normal” life. Additional emotional struggles may arise as head-injured individuals often remember what they were like before the incident and are frustrated with the current lack of functioning. These cognitive, social, and emotional difficulties speak to the importance of designing and validating appropriate assessment measures for children and adolescents who have sustained a TBI.

Conflict of Interest

None declared.

References

Anderson
P.
Assessment and development of executive function (EF) during childhood
Child Neuropsychology
 , 
2002
, vol. 
8
 
2
(pg. 
71
-
82
)
[PubMed]
Anderson
V. A.
Morse
S. A.
Catroppa
C.
Haritou
F.
Rosenfeld
J. V.
Thirty month outcome from early childhood head injury: a prospective analysis of neurobehavioral recovery
Brain
 , 
2004
, vol. 
127
 (pg. 
2608
-
2620
)
Anderson
V. A.
Yeates
K. O.
Anderson
V. A.
Yeates
K. O.
Introduction: pediatric traumatic brain injury
Pediatric Traumatic Brain Injury: New Frontiers in Clinical and Translational Research
 , 
2010
New York
Cambridge University Press
Armstrong
C. M.
Allen
D. N.
Donohue
B.
Mayfield
J.
Sensitivity of the Comprehensive Trail Making Test to traumatic brain injury in children and adolescents
Archives of Clinical Neuropsychology
 , 
2008
, vol. 
23
 
3
(pg. 
351
-
358
)
Baron
I. S.
Neuropsychological Evaluation of the Child
 , 
2004
New York
Oxford University Press
Blair
C.
Zelazo
P. D.
Greenberg
M. T.
The measurement of executive function in childhood
Developmental Neuropsychology
 , 
2005
, vol. 
28
 
2
(pg. 
561
-
571
)
Boll
T. J.
Berent
S.
Richards
H.
Tactile-perceptual functioning as a factor in general psychological abilities
Perceptual and Motor Skills
 , 
1977
, vol. 
44
 (pg. 
535
-
539
)
[PubMed]
Brookshire
B.
Levin
H. S.
Song
J. X.
Zhang
L.
Components of executive function in typically developing and head-injured children
Developmental Neuropsychology
 , 
2004
, vol. 
25
 
1
(pg. 
61
-
83
)
[PubMed]
Delis
D
Kaplan
E.
Kramer
J. H.
Delis-Kaplan Executive Function System Examiner's Manual
 , 
2001
San Antonio, TX
The Psychological Corporation
Diamond
A.
Carlson
S. M.
Beck
D. M.
Preschool children's performance in task switching on the dimensional change card sort task: separating the dimensions aids the ability to switch
Developmental Neuropsychology
 , 
2005
, vol. 
28
 
2
(pg. 
689
-
729
)
Epsy
K. A.
Kaufmann
P. M.
Glisky
M. L.
McDiarmid
M. D.
New procedures to assess executive functions in preschool children
The Clinical Neuropsychologist
 , 
2001
, vol. 
15
 (pg. 
46
-
58
)
Farmer
J. E.
Clippard
D. S.
Luehr-Wiemann
Y.
Wright
E.
Owings
S.
Bigler
E. D.
Clark
E.
Farmer
J. E.
Assessing children with traumatic brain injury during rehabilitation: promoting school and community reentry
Childhood Traumatic Brain Injury
 , 
1997
Austin, TX
Pro-Ed
(pg. 
33
-
62
)
Gioia
G. A.
Isquith
P. K.
Guy
S. C.
Simeonsson
R. J.
Rosenthal
S. L.
Assessment of executive functions in children with neurological impairment
Psychological and Developmental Assessment: Children with Disabilities and Chronic Conditions
 , 
2001
New York
Guilford Press
Gray
R.
Review of comprehensive trail making test
Journal of Psychoeducational Assessment
 , 
2006
, vol. 
24
 
1
(pg. 
88
-
91
)
Hamdan
A. C.
Hamdan
E. M. L. R.
Effects of age and education level on the TMT in a healthy Brazilian sample
Psychology and Neuroscience
 , 
2009
, vol. 
2
 (pg. 
199
-
203
)
Jaffe
K. M.
Fay
G. C.
Polissar
N. L.
Martin
K. M.
Shurtleff
H. A.
Rivara
J. M.
, et al.  . 
Severity of pediatric traumatic brain injury and neurobehavioral recovery at one year: a cohort study
Archives of Physical Medicine and Rehabilitation
 , 
1993
, vol. 
74
 (pg. 
87
-
95
)
Jankowitz
B. T.
Adelson
P. D.
Pediatric traumatic brain injury: past, present and future
Developmental Neuroscience
 , 
2006
, vol. 
28
 (pg. 
264
-
275
)
Langlois
J. A.
Rutland-Brown
W.
Thomas
K. E.
Traumatic Brain Injury in the United States: Emergency Department Visits, Hospitalizations, and Deaths.
 , 
2006
Atlanta, GA
Centers for the Disease Control and Prevention, National Center for Injury Prevention and Control
Maricle
D. E.
Johnson
W. L.
Avirett
E.
Miller
D. C.
Best practices in assessing and intervening in children with executive function disorders
Best Practices in School Neuropsychology
 , 
2010
NY
Wiley & Sons
Moses
J. A.
Comprehensive Trail Making Test review
Archives of Clinical Neuropsychology
 , 
2004
, vol. 
19
 (pg. 
703
-
708
)
Neyens
L. G. J.
Aldenkamp
A. P.
Stability of cognitive measures in children of average ability
Child Neuropsychology
 , 
1996
, vol. 
2
 
3
(pg. 
161
-
170
)
Reitan
R. M.
The relation of the Trail Making Test to organic brain damage
Journal of Consulting Psychology
 , 
1955
, vol. 
19
 
5
(pg. 
393
-
395
)
Reitan
R. M.
Trail Making Test results for normal and brain-damaged children
Perceptual and Motor Skills
 , 
1971
, vol. 
33
 (pg. 
575
-
581
)
[PubMed]
Reitan
R. M.
Wolfson
D.
Neuropsychological Evaluation of Older Children.
 , 
1992
Tucson, AZ
Neuropsychology
Reynolds
C. R.
Comprehensive Trail Making Test: Examiner's Manual.
 , 
2002
Austin, TX
Pro-Ed
Servesko
A. M.
Smith
S.
Edwards
J. A.
Validity of the Comprehensive Trail Making Test.
 , 
2006
Paper presented at the National Academy of Neuropsychology
San Antonio, TX
Taylor
H. G.
Alden
J.
Age-related differences in outcomes following childhood brain insults: an introduction and overview
Journal of the International Neuropsychological Society
 , 
1997
, vol. 
3
 (pg. 
1
-
13
)
[PubMed]
Tombaugh
T. N.
Trail Making Test A and B: normative data stratified by age and education
Archives of Clinical Neuropsychology
 , 
2004
, vol. 
19
 (pg. 
203
-
214
)
Ylvisaker
M.
DeBonis
D.
Executive function impairment in adolescence: TBI and ADHD
Topics in Language Disorders
 , 
2000
, vol. 
20
 
2
(pg. 
29
-
57
)
Zalonis
I.
Christidi
F.
Karaizou
E.
Triantafyllou
N. I.
Kapaki
E.
Paraskevas
G.
, et al.  . 
Reference data for derived Trail Making scores in Greek healthy population
Annals of General Psychiatry
 , 
2010
, vol. 
9
 
(Suppl. 1)
pg. 
S145
 
Zelazo
P. D.
Muller
U.
Frye
D.
Marcovitch
S.
The development of executive function in early childhood
Monographs of the Society for Research in Child Development
 , 
2003
, vol. 
68
  
(Serial No. 274).
[PubMed]