Analysis of Errors on the Trail Making Test

Abstract

This study was designed to determine whether the number and/or types of errors on the Trail Making Test differentiated head-injured and normal control subjects. Errors on Part B were categorized into two types of shifting errors (from number to letter and from letter to number) and two types of sequencing errors (number and letter). Subjects consisted of 133 head-injured patients and 145 normal controls. Analysis showed that the frequency of errors on Parts A and B did not differ significantly between the groups nor did the percentage of subjects making errors. Total shifting and sequencing errors also did not differentiate between the two groups. Although head-injured subjects were more likely than controls to err in shifting from letters to numbers, this finding was of no apparent clinical usefulness. The discriminative value of time scores was confirmed.

Administration and interpretation of the Trail Making Test

“…Subject connects the wrong circle Cross out the erroneous line and provide immediate feedback, either in terms of ascending connections for part A or alternation for part B these parts [35][36][37][38] ; however, neither error scores [38][39][40] nor the interactions between parts A and B contribute to diagnostic sensitivity of injury severity 38 ; time to completion provides the most parsimonious and sensitive score. TMT results shortly after injury, as part of a larger battery of tests and demographic features (e.g., age, duration of loss of consciousness) are valid prognostic indicators of functional recovery in traumatic brain injury patients, although they do not account for a large proportion of variance alone [41][42][43] .…”

“…The effect of errors is realized on the total time to complete the test, as the examiner stops the examinee and returns him or her to the last correct response. Although some studies have found that patient populations, such as those with head injury, are more prone to make errors in shifting from letters to numbers, these findings are of limited clinical value 5 .…”

Trail Making Test: Normative data for Turkish elderly population by age, sex and education

“…There are some investigators claiming that, periods of time to complete Parts A and B are able to differentiate participants with brain damage from the healthy participants, while the error analyses (alignment and pairing errors) are inadequate to differentiate these two groups [26]. However, some other authors debate that the number and type of mistakes recorded during the test could obviously predict cognitive disturbances [6,27].…”

The Effects of Aging, Malingering, and Traumatic Brain Injury on Computerized Trail-Making Test Performance

“…Mild TBI (mTBI) can prolong TMT completion times in the chronic phase, particularly on Trails B, even in patients with no neuroimaging evidence of brain damage [55]. The increased completion times are not generally accompanied by increases in error frequency [28, 56]. However, prolongations in both Trails A and Trails B can be seen in more complicated mTBI cases [51, 54].…”

“…However, comparisons of C-TMT-B minus C-TMT-A completion times and C-TMT-B/C-TMT-A ratios between TBI patients and controls failed to reach significance for either completion times or velocities. We also found no differences between TBI patients and controls in error rates on either test [56]. …”