An Automated Mobile Game-based Screening Tool for Patients with Alcohol Dependence

on YouTube ↗06m

" summary

An app captures gameplay performance, touch gestures and device motion while people play Tetris, Fruit Ninja and Unblock Puzzle and uses cleaned touch data plus sensor features to train classifiers on 40 alcohol-dependent patients and 40 matched controls. Random Forest on 3-minute Unblock Puzzle samples achieves 95% accuracy with 95% sensitivity and specificity and outperforms paper and EEG screens. The system runs quickly on smartphones and targets brief self-screening and relapse monitoring.

AI-written summary · about

" chapters (9)

1. 0:00Intro
2. 0:14Effects of Long-term Excessive Alcohol Use
3. 0:52Mobile Game-Based Approach for Alcohol Dependence Screening
4. 2:12Data Processing
5. 2:56Model Comparisons on Classification Performance over Five Difference Feature Sets
6. 3:50Model Classification Performance Comparisons between 10-Min and 3-Min Samples
7. 4:42Classification Performance in Terms of Sensitivity and Specificity
8. 5:08Key Findings
9. 5:39Implications