Observability constrains expertise-dependent kinematic readout in action prediction

Area of Science:

• Cognitive Neuroscience
• Motor Control
• Human Movement Analysis

Background:

• Human movement exhibits variability, complicating action prediction from kinematic data.
• Perceptual expertise typically improves anticipation, but its modulation by kinematic variability is not well understood.

Purpose of the Study:

• To investigate how kinematic observability influences expert performance in action prediction.
• To determine if expertise benefits are modulated by trial-to-trial kinematic variations.

Main Methods:

• Introduced 'kinematic observability' as a measure of movement instance distance from the optimal decision boundary.
• Employed an expert-novice paradigm in a table-tennis action prediction task.
• Sampled trials based on objective evidence strength, not body part masking.

Main Results:

• Experts outperformed novices on high-observability trials, aligning readout weights with optimal encoding.
• The expert advantage decreased under low observability conditions.
• Experts showed a criterion shift towards prior expectations when kinematic evidence was weak.

Conclusions:

• Expertise enhances anticipatory action prediction when kinematic observability is high.
• The benefits of expertise are reduced with weak kinematic evidence.
• Experts adjust their decision criteria based on kinematic observability, relying more on prior expectations when cues are ambiguous.