BibTex
@inproceedings{Lara-Garduno:2019:10.20380/GI2019.10,
author = {Lara-Garduno, Raniero and Igarashi, Takeo and Hammond, Tracy},
title = {3D-Trail-Making Test: A Touch-Tablet Cognitive Test to Support Intelligent Behavioral Recognition},
booktitle = {Proceedings of Graphics Interface 2019},
series = {GI 2019},
year = {2019},
issn = {0713-5424},
isbn = {978-0-9947868-4-5},
location = {Kingston, Ontario},
numpages = {8},
doi = {10.20380/GI2019.10},
publisher = {Canadian Information Processing Society},
}
Abstract
The decades-old Trail-Making Test has established itself as an effective and versatile cognitive testing tool. However, its reliance on a paper-and-pencil method of administration severely limits its capabilities for quantitative evaluation. We evaluate input data via sketch recognition algorithms on a digitized version of the Trail-Making Test, as well as exploring the viability of a novel touch-based version designed around the basic concepts of the traditional paper-and-pencil exam. Two quantitative studies helped evaluate the viability of these digital examinations: the first builds a normative data set that shows a stable increase in test completion times with a lower overall skewness, while the second indicates the new examination’s capability to perform behavioral classification based on participant nationality by correctly classifying input behavior with up to 96% accuracy.