Interception of virtual throws reveals predictive skills based on the visual processing of throwing kinematics

Area of Science:

• Human motor control
• Action prediction
• Human-computer interaction

Background:

• Predicting action outcomes is vital for social interactions and performance in dynamic environments like sports.
• Understanding how humans predict complex actions, especially in real-time, is key to fields ranging from robotics to skill acquisition.

Purpose of the Study:

• To quantitatively assess the predictive abilities of non-trained adults intercepting thrown balls in immersive virtual reality.
• To investigate how visual information (complete action vs. ball flight) and individual motor styles influence action prediction accuracy.

Main Methods:

• Participants intercepted thrown balls in an immersive virtual reality environment.
• Performance was measured under conditions with complete visual information of the throw and occluded ball flight.
• The influence of the thrower's individual motor style on prediction was analyzed.

Main Results:

• Participants demonstrated effective prediction of action outcomes, even without prior training.
• Performance improved when the entire throwing action was visible compared to only the ball flight.
• Predictive accuracy and movement direction were adaptable even when ball flight was occluded.
• Individual motor styles of the thrower significantly impacted prediction performance.

Conclusions:

• Humans possess an innate ability to predict complex, full-body actions and their environmental consequences.
• Online prediction of actions is utilized to optimize motor performance in interactive tasks.
• This suggests a functional understanding of how common human actions translate into environmental changes.