Background Stroke is a major cause of long-term disability, often leading to cognitive and motor impairments. While physical rehabilitation is routinely provided to individuals following Stroke, access to cognitive rehabilitation remains limited. Extended Reality Serious Games (XR-SG) offer a promising method to deliver engaging and effective cognitive training embedded in motor response tasks. We recently developed the Autoadaptive REASmash, a non-immersive XR-SG designed to support the neurorehabilitation of visuospatial attention and distractor inhibition. The game adapts difficulty in correspondence to performance. Here, we examined the reliability of the XR-SG autoadaptive algorithm. Methods The autoadaptive REASmash’s algorithm modifies task difficulty in correspondence to performance, allowing the maintenance of a motivating success rate of 75-85%. Several parameters are adjusted in parallel, including: (i) target spatial position; (ii) target-distractor salience and distractor number; (iii) cues to attract attention to the target; and (iv) stimulus presentation time. Fifteen individuals with cortical-subcortical hemispheric stroke completed three 15-minute training sessions over one week, responding with their less affected hand. Results The intervention was well tolerated, and participants remained engaged throughout. Task complexity increased progressively within the first trials, stabilizing at the target success rate of 75-85%. When performance success rate stabilized, task complexity significantly differed across participants. Conclusions The observed variability in task complexity reflects the algorithm’s ability to adapt difficulty demands to each participant’s performance, maintaining the targeted success rate. By combining personalization, adaptivity, and accessibility, the Autoadaptive REASmash offers a valuable innovation to neurorehabilitation practice. Further research will investigate the clinical impact and long-term benefits on selected cognitive impairments such as spatial neglect.
Building similarity graph...
Analyzing shared references across papers
Loading...
Gregorio Sorrentino
Second European Symposium on Extended Reality (XR) Technologies
Building similarity graph...
Analyzing shared references across papers
Loading...
Sorrentino et al. (Wed,) studied this question.