Unveiling Trail Making Test: visual and manual trajectories indexing multiple executive processes

Abstract

The Trail Making Test (TMT) is one of the most popular neuropsychological tests for executive functions (EFs) assessment. It presents several strengths: it is sensitive to executive dysfunction, it is easy to understand, and has a short administration. However, it has important limitations. First, the underlying EFs articulated during the task are not well discriminated, which makes it a test with low specificity. Second, the pen-and-paper version presents one trial per condition which introduces high variability. Third, only the total time is quantified, which does not allow for a detailed analysis. Fourth, it has a fixed spatial configuration per condition. We designed a computerised version of the TMT to overcome its main limitations and evaluated it in a group of neurotypical adults. Eye and hand positions are measured with high resolution over several trials, and spatial configuration is controlled. Our results showed a very similar performance profile compared to the traditional TMT. Moreover, it revealed differences in eye movements between parts A and B. Most importantly, based on hand and eye movements, we found an internal working memory measure that showed an association to a validated working memory task. Additionally, we proposed another internal measure as a potential marker of inhibitory control. Our results showed that EFs can be studied in more detail using traditional tests combined with powerful digital setups. The cTMT showed potential use in older adult populations and patients with EFs disorders.

Figure 1

(A) Experimental design and task validation. The trial begins with a mouse button press and continues until the mouse button is released or a maximum time of 25 s is reached. (B) Time to connect 12 items in order (completion criteria), and (C) Percentage of completed trials (PC) for both Part-A and Part-B. (D) Correlation between the Total IFS Score and the Completion Ratio (PC-B/PC-A). (E) Correlation between the Total IFS Score and the RT Ratio (RT-B/RT-A) (F) Hand trajectory in an A type trial. (G) Hand trajectory in a B type trial, with the same configuration. The colour bar represents the relative temporal evolution for each subject.

Figure 2

(A,B) Eye trajectory for all the subjects in a TMT-A (A) and TMT-B (B) trials. The colour bar represents the relative temporal evolution for each subject. Note that the yellow colour, representing the final fixations of the trial, are more consistently located around the last targets. (C–F) Individual eye movements characteristics. Distributions of saccade duration (C), saccade amplitude (D), fixation durations (E), and the number of fixations during the trial (F). Statistical significance of the difference between distributions was assessed with the Kuiper’s test: C: V = 0.01, p = 0.77; D: V = 0.01, p = 0.96; E: V = 0.02, p = 0.11; F: V = 0.17, p = 2.5*10–4.

Figure 3

(A) Phases of the TMT-Task, in terms of hand-eye interactions. This diagram classifies fixations according to the three phases where they occurred: exploratory, monitoring, and planning. (B) Boxplots of the median number of fixations in each of the three phases (exploratory, planning, and monitoring) for both parts (TMT-A and TMT-B). (C) Correlations between the RT Ratio (RT-B/RT-A) and the number of fixations ratio in each phase. (D) Boxplots of the median duration of fixations in each phase (exploratory, planning, and monitoring) for both parts (TMT-A and TMT-B). (E) Correlations between the RT Ratio (RT-B/RT-A) and the duration of fixations ratio in each phase.

Figure 4

(A) Diagram for identifying targets remembered. In (A.i) the target “2” is remembered as it was not looked right before reaching it with the cursor, while in (A.ii) the target “2” was not remembered as it was seen again right before reaching it with the cursor (B) Boxplots for the mean amount of Targets Remembered (TR) for each subject in both parts (TMT-A and TMT-B) (C) Correlation between the PC Ratio (PC-B/PC-A) and the TR Ratio (Targets Remembered in B/ Targets Remembered in A). (D) Experimental design of the Change Detection Task. Memory array: 4 or 6 coloured squares were shown on-screen during 150 ms, retention interval: only fixation cross through 900 ms, Test array: A single square appeared on screen with same colour and location for consistent types of trials and difference in one or both characteristics for inconsistent ones. (E) Correlation between Visual Working Memory Capacity (K_average) and the TR Ratio.

Figure 5

(A) Diagram showing the difference between the Correct detections and the False detections. (B,C) Spatial distribution of the paths explored by the hand when fixating a new item, for Correct (next in the sequence; B) and False (C) detections. 2-D paths were aligned and normalised so that the fixated item was at (0,1) (see Fig. S3). (D) Relative displacement towards the fixated item, in the direction of the item. Curves are aligned to the fixation onset. Red: TMT-A, Correct; Magenta: TMT-A, False; Blue: TMT-B, Correct; Cyan: TMT-B, False. (E) Difference between Correct and False detections for displacement. Black: TMT-A; Grey: TMT-B. (F) Area under the difference curves for the displacement. The area was calculated participant-by-participant between 50 and 300 ms.