Abstract
Within the field of cognitive rehabilitation after brain injury, rehabilitation training is constantly adjusted to match the skills and progress of the individual patients. As no two patients are alike in func- tional injury and recovery, it is a challenge to provide the right amount of training at the right level of difficulty at any given time. This study investigates whether a modified version of the artificial intelligence (AI) reinforcement method called the "actor-critic method" is able to detect response time patterns and subsequently con- trol the level of difficulty in a computer-based, cognitive training program. The efficacy of the AI logic was tested under the actual training conditions of a brain-injured patient. The results showed that the AI controlled training system was able to learn and adjust fast enough to control and adapt the level of difficulty of the training to match the changes in the patient's abilities over a three-week period.
| Translated title of the contribution | Anvendelse af kunstig intelligens til tilpasning af sværhedsgrad i computer-baseret kognitiv genoptræning |
|---|---|
| Original language | English |
| Journal | Journal of Cybertherapy and Rehabilitation |
| Volume | 4 |
| Issue number | 3 |
| Pages (from-to) | 387-397 |
| Number of pages | 11 |
| ISSN | 1784-9934 |
| Publication status | Published - 2011 |