From the Journals

Robot-assisted, gamelike tool effective for classifying ADHD



A novel robot-assisted, gamelike test accurately classified ADHD type in elementary school–aged children, according to Mun-Taek Choi, PhD, and associates.

A total of 326 children in the third and fourth grades were included in the study, 35 of whom had been diagnosed with ADHD and 26 of whom were at risk. For the 10- to 12-minute test, participants followed a robot on a path across a numbered mat while stimuli were shown on a TV with both images and sound, and completed a task at each numbered square, reported Dr. Choi, of Sungkyunkwan University, Suwan, South Korea, and associates. The study was published in the Journal of Intelligent & Robotic Systems.

Inattentive and hyperactive-impulsive behavior was measured by the number of omission and commission errors. Response time and task completion time contributed to the measure of inattentive and hyperactive-impulsive behavior. Working memory deficits were measured as deviations in the prescribed route.

The tool was able to identify the children with ADHD with a high degree of accuracy, up to a maximum of 97%. This figure improved over the course of the study as the tool learned more, indicating that generalization errors were not a serious issue for the tool, the investigators noted.

“Unlike conventional questionnaire-based tests, the robot-assisted test increases the accuracy of ADHD diagnosis by directly reflecting the quality of children’s behavior during the activity game with the robot involved in the action. Since the test obtains behavioral patterns and levels using robotic sensing technologies, it can reliably determine the three key elements of ADHD diagnosis: hyperactivity, inattentive behavior, and working memory,” the investigators wrote. Ultimately, Dr. Choi and associates wrote, the tool could help clinicians diagnose childhood ADHD.

The study was funded by the South Korean Ministry of Trade, Industry, & Energy. No disclosures were reported.

SOURCE: Choi M-T et al. J Intell Robot Syst. 2018 Jun 19. doi: 10.1007/s10846-018-0890-9.

Next Article: