Miguel Cienfuegos ^1,2* Abdeldjallil Naceri ³ Jonathan Maycock ⁴ Risto Kõiva ¹ Helge Ritter ^1,2 Thomas Schack ^1,2

• ¹ Bielefeld University, Bielefeld, Germany
• ² Center of Cognitive Interaction Technology, Bielefeld University, Bielefeld, North Rhine-Westphalia, Germany
• ³ Munich Institute of Robotics and Machine Intelligence, Technical University of Munich, Munich, Bavaria, Germany
• ⁴ Margin UG, Bielefeld, Germany

This study investigates the multifaceted nature of motor learning in a complex bimanual task by examining the interplay between mental representation structures, biomechanics, tactile pressure, and performance. We devised a novel maze game that exemplifies complex sequential coordination of vision and haptic control using two hands, requiring participants to maneuver a rolling sphere through a maze. We employed motion capture and innovative tactile sensors to obtain a precise multimodal picture of the entire interaction process, enabling us to analyze the essential aspects of the task and the progression of the underlying multimodal interaction patterns during learning. Our primary aims were to (1) investigate the effects of daily practice on task performance, biomechanics behavior, and tactile pressure, (2) examine the relationship between changes in mental representation structures and changes in skill performance in a complex movement, and (3) explore the interplay between biomechanics, tactile pressure, and mental representation structure in motor learning. Performance analysis revealed that after practice, motor skills improved, with groups formed based on initial maze performance-poor performers' group (PPG) and good performers' group (GPG), each of equal size-showing the GPG outpacing the PPG in maze navigation efficiency. Biomechanically, the GPG exhibited superior movement strategies, reflected in their peak velocity and fewer velocity peaks during task execution. Concurrently, tactile feedback, as evidenced by the pressure application patterns of the right-hand thumb, revealed that GPG players utilized more precise and focused pressure, suggesting enhanced motor control and task familiarity. On a cognitive level, both groups evolved 1 Cienfuegos et al.their mental representation structures of the maze game over time; however, the GPG group's posttest cognitive mapping indicated a more refined and structured understanding of the task. These results underscore the intertwined nature of biomechanical control, tactile feedback, and cognitive processing in motor skill acquisition. The findings resonate with established theories, such as the cognitive action architecture approach, emphasizing mental representation's significant role in motor action planning and execution. Moreover, the observed interplay between tactile feedback, biomechanics, and cognitive representation suggests that a holistic approach is essential for motor learning research, emphasizing the multifaceted nature of skill acquisition.