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

Abstract

The trail making test (TMT) is widely used to assess speed of processing and executive function. However, normative data sets gathered at different sites show significant inconsistencies. Here, we describe a computerized version of the TMT (C-TMT) that increases the precision and replicability of the TMT by permitting a segment-by-segment analysis of performance and separate analyses of dwell-time, move-time, and error time. Experiment 1 examined 165 subjects of various ages and found that completion times on both the C-TMT-A (where subjects connect successively numbered circles) and the C-TMT-B (where subjects connect circles containing alternating letters and numbers) were strongly influenced by age. Experiment 2 examined 50 subjects who underwent three test sessions. The results of the first test session were well fit by the normative data gathered in Experiment 1. Sessions 2 and 3 demonstrated significant learning effects, particularly on the C-TMT-B, and showed good test-retest reliability. Experiment 3 examined performance in subjects instructed to feign symptoms of traumatic brain injury: 44% of subjects produced abnormal completion times on the C-TMT-A, and 18% on the C-TMT-B. Malingering subjects could be distinguished from abnormally slow controls based on (1) disproportionate increases in dwell-time on the C-TMT-A, and (2) greater deficits on the C-TMT-A than on the C-TMT-B. Experiment 4 examined the performance of 28 patients with traumatic brain injury: C-TMT-B completion times were slowed, and TBI patients showed reduced movement velocities on both tests. The C-TMT improves the reliability and sensitivity of the trail making test of processing speed and executive function.

Conflict of interest statement

Competing Interests: The author has read the journal's policy and the authors of this manuscript have the following competing interests: DLW is affiliated with NeuroBehavioral Systems, Inc., the developers of Presentation software that was used to conduct these experiments. This in no way alters the authors' commitment to sharing data.

Figures

Fig 1. The C-TMT.

The display as seen by a subject midway between connecting circles 7 and 8 on the C-TMT-A. As each circle was correctly connected, its color changed from white to green (e.g., circles 1–7). The current path between circles was shown in white (e.g., between 7 and 8), but was replaced by a straight green line as each segment was completed.

Fig 2. C-TMT-A paths.

A. The path drawn by a subject on the C-TMT-A, containing an error shown by a red line connecting circles 4 to 6. The subject clicked outside circle 5, and then advanced to circle 6, which resulted in his return to circle 4 (white line). He then drew a circuitous connection to circle 5 through circle 7. No further errors were committed. B. The superimposed C-TMT-A paths of all Experiment 1 control subjects. Paths drawn by individual subjects are shown in blue. Erroneous connections are shown as straight red lines.

Fig 3. C-TMT-B paths.

A. A C-TMT-B path from a control subject containing an error shown in red. The subject drew a direct path from G to H (circling through 7) and was then returned to G (straight white line). No further errors were committed. B. The superimposed C-TMT-B paths of Experiment 1 control subjects. Paths drawn by individual subjects are shown in blue. Erroneous connections are shown as straight red lines.

Fig 4. C-TMT completion times for Trails A (top) and Trails B (bottom).

Shown as a function of age. The data are from Experiment 1 (blue diamonds), the first session of Experiment 2 (green triangles), malingering subjects in Experiment 3 (Mal, black crosses), and TBI patients in Experiment 4 (red squares). The linear age-regression functions from Experiment 1 are shown.

Fig 5. C-TMT completion time z-scores for Trails A (top) and Trails B (bottom).

Shown as a function of age. Regression functions from Experiment 1 were used to adjust Z-scores for age and computer use. Dashed lines show p

Fig 6. C-TMT velocity z-scores for Trails A (top) and Trails B (bottom).

Shown as a function of age. Regression functions from Experiment 1 were used to adjust Z-scores for age and computer use. Dashed lines show p

Fig 7. C-TMT-A completion-time z-scores as a function of the dwell-time/total time ratios.

The dashed horizontal line shows the p

Fig 8. Difference z-scores (C-TMT-B minus C-TMT-A) for subjects of different ages.

The dashed horizontal line shows the p