Computation on Demand: Action-Specific Representations of Visual Task Features Arise during Distinct Movement Phases

Area of Science:

• Cognitive Neuroscience
• Neuroscience
• Human Motor Control

Background:

• Intention to act shapes feature processing, but its temporal dynamics and neural representation remain unclear.
• Existing research often uses artificial task feature combinations, potentially overemphasizing working memory's role.
• The dominant view suggests conjunctive neural representations for task features.

Purpose of the Study:

• Investigate the temporal evolution of feature representations during action planning.
• Examine the integration of features like shape and weight in neural representations.
• Test whether actions lead to sustained, unified feature representations.

Main Methods:

• Used electroencephalography (EEG) with a well-rehearsed object interaction task.
• Participants grasped or touched objects with varying shapes and weights.
• Applied multivariate analysis to EEG data to decode neural representations.

Main Results:

• Object shape representations were similar for grasping and non-grasping actions.
• Early visual cortex reactivated for shape during grasp planning, indicating feedback.
• Grasp-specific material/weight representations emerged during execution; integrated representations were transient.

Conclusions:

• The brain forms action-specific representations tailored to task demands.
• Neural representations are dynamically updated and integrated as needed, not fixed.
• Results challenge the notion of inevitably unified neural representations in goal-directed actions.