Carryover and spatial bias effects in Trail Making Test Part A sequential visual search

Congratulations to Dr. Kenneth Heilman on the publication of “Carryover and spatial bias effects in Trail Making Test Part A sequential visual search,” which appears in the August edition of Journal of Vision.



The Trail Making Test (TMT) was designed for the U.S. Army 1944 Individual Test Battery. Norms availability promotes widespread adoption. Part A of two-part TMT (TA) assesses complex attention with a timed task to connect, sequentially, 26 scattered numbered circles in numeric order. The present study manipulated spatial and temporal characteristics of that scattering to better separate influences on task performance of implicit learning or memory, and of spatial attentional bias. Forty healthy young adults (25 female; age 20.38 ± 1.41; years education 14.1 ± 1.38) were tested individually on 8 TA trials shown on an iPad with trials in different randomized orders. In each trial, the screen contained 26 locations with numeric labels. In half of the trials (scrambled), locations of non-target, task irrelevant labels shuffled on each click. Clicking the prior search target also caused the label for the new search target to appear in its to-be-found location rather than in some other location. In the other half of the trials (static), pairing of labels with locations did not change so that incidental learning was possible. One trial each of scrambled and static manipulations had the same label-location target pairs. Scrambled non-target label re-locations substituted for persisting target label locations in the static trials. Four location arrangements were initial scatter, left-right reversed, up-down reversed, and both left-right and up-down reversed, symmetries unreported by participants. Mean inter-target interval (ITI) was shorter for static than for scrambled trials (1.13 ± 0.12; F(1,319) = 131.2, p < 0.001), consistent with implicit memory for non-targets speeding static searches. Post-hoc, paired ITI was shorter on left half-screen than on right half-screen (0.206 ± 0.043; t(3999) = 4.74. p < 0.001), consistent with “pseudo-neglect” spatial bias speeding left-sided searches. Upper/lower half-screen lacked significant search speed influence. How age changes these effects is under investigation.