JMIR Publications

ABSTRACT

Background:

The Processing Speed Index (PSI) of the Korean Wechsler Intelligence Scale for Children - Fifth Edition (K-WISC-V) is highly correlated with ADHD symptoms and serves as an important indicator of cognitive function. However, restrictions on how frequently PSI testing can be performed prevent PSI determination on a daily or short-term basis. Therefore, an accessible and objective technique for predicting PSI scores is needed to help improve monitoring and intervention for children with ADHD.

Objective:

To overcome the limitations of K-WISC-V and collect PSI scores in the short term, this study aimed to develop a machine learning-based technique to use data collected from serious games to predict the PSI scores of children with ADHD.

Methods:

Sixty children (6-12 years old) with ADHD were recruited. The participants completed an initial PSI assessment using K-WISC-V followed by 25 min of engagement with serious game content. Data from the game sessions were used to train machine learning models, and the models’ performance in predicting PSI scores was evaluated using the root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percent error (MAPE), with K-fold cross-validation (k=4) applied to ensure robustness.

Results:

Sixty children (6-12 years old) with ADHD were recruited. The participants completed an initial PSI assessment using K-WISC-V followed by 25 min of engagement with serious game content. Data from the game sessions were used to train machine learning models, and the models’ performance in predicting PSI scores was evaluated using the root mean squared error (RMSE), mean absolute error (MAE), and mean absolute percent error (MAPE), with K-fold cross-validation (k=4) applied to ensure robustness.

Conclusions:

The PSI prediction model using serious games data shows promise for providing an objective assessment tool for ADHD. By accurately predicting PSI scores, this approach could complement traditional subjective evaluations, allowing for continuous tracking of patient status and optimized treatment planning. The use of serious games offers a novel, engaging way to gather meaningful cognitive data outside traditional clinical settings, potentially improving accessibility and adherence. However, the results remain to be validated in larger, more diverse populations, and the long-term feasibility of using serious games in clinical and educational settings needs to be explored. Clinical Trial: The initial phase of this study (IRB No. 2021-10-080) was approved by the Institutional Review Board of Kye Myung University Hospital, certified by KOLAS in South Korea. The subsequent phase was registered with the Korea Clinical Research Information Service (CRIS Registration No. KCT0009862), accredited by WHO. Further details are provided in the Methods section.