Grasping of Real-World Objects Is Not Biased by Ensemble Perception

Area of Science:

• Cognitive Neuroscience
• Motor Control
• Visual Perception

Background:

• The visual system creates summary representations of objects, influencing individual object perception.
• Motor behavior typically relies on precise object dimensions and resists perceptual biases.
• The role of ensemble perception in guiding actions, especially with unconstrained dimensions, is less understood.

Purpose of the Study:

• To investigate if ensemble statistics of surrounding objects influence grip aperture and orientation during reaching-to-grasp tasks.
• To determine if visuomotor behavior is biased by ensemble representations when a target's orientation is not explicitly defined.

Main Methods:

• Participants reached-to-grasp a central target surrounded by an ensemble display.
• Grip aperture and orientation were measured during the grasping action.
• Manual estimations of ensemble size and orientation were collected.
• Support vector machine classification was used on kinematic and electromyography data.

Main Results:

• Grasping actions (grip aperture and orientation) were not biased by the ensemble statistics of surrounding objects.
• Participants could accurately perceive and estimate the average size and orientation of the ensemble.
• Machine learning models could classify ensemble statistics from perceptual data but not from grasping data.

Conclusions:

• Visually guided grasping behavior is robust and not influenced by ensemble perception, even with unconstrained dimensions.
• Ensemble representations do not bias motor actions toward the average, maintaining accurate interaction with real-world objects.