A genetic algorithm-based method to modulate the difficulty of serious games along consecutive robot-assisted therapy...

Area of Science:

• Neurorehabilitation engineering
• Human-computer interaction
• Game design

Background:

• Optimizing therapy intensity in neurorehabilitation is crucial for limb movement recovery and patient motivation.
• Existing adaptive therapy solutions require session-specific recalibration.
• A dynamic, multi-session adaptation method is needed.

Purpose of the Study:

• To propose and evaluate a dynamic adaptation method for neurorehabilitation games that operates across multiple sessions without recalibration.
• To enhance therapy intensity adaptation using a genetic algorithm to maximize user movement and motivation.
• To investigate the impact of adaptive game design on user psychophysiological states.

Main Methods:

• A genetic algorithm was employed to adapt game parameters, aiming to maintain a target score and motivate user movement.
• The adaptation method was tested with two serious games and a rehabilitation robot over five sessions.
• User psychophysiological state was assessed using the Self-Assessment Manikin (SAM) test and physiological signals (cardiorespiratory, galvanic skin response).

Main Results:

• The genetic algorithm successfully identified game parameters that maximized user movement in both evaluated games.
• High fidelity was observed in maintaining the target score rate in one game.
• Personalization increased across sessions, indicated by growing game parameter dispersion.
• User emotional and physiological states remained consistent despite increased game difficulty.

Conclusions:

• Genetic algorithm-based adaptation effectively maximizes neurorehabilitation therapy performance across sessions without recalibration.
• Game design significantly influences the effectiveness of the adaptation process.
• Adaptive game design, as implemented, does not adversely affect users' emotional or physiological well-being.