Gestalts at threshold could reveal Gestalts as predictions

Area of Science:

• Cognitive Neuroscience
• Computational Vision
• Psychophysics

Background:

• The
• Gestalts as predictions
• hypothesis posits that Gestalt phenomena arise from predictive processing mechanisms.
• Understanding mid-level vision requires clarifying the roles of higher-order object-selective areas and feedback connections.

Purpose of the Study:

• To investigate the robustness of Gestalt phenomena under varying saliency conditions, particularly below perceptual threshold.
• To explore the relationship between Gestalt perception and neural processing, specifically the engagement of object-selective areas.
• To assess the utility of classic psychophysics in examining hierarchical predictive processing in Gestalt perception.

Main Methods:

• Two psychophysical experiments were conducted manipulating contrast and configurality.
• Performance was assessed across the psychometric function, including conditions below perceptual threshold.
• Accuracy gains and saliency associated with Gestalt stimuli were measured relative to control conditions.

Main Results:

• Gestalt phenomena remained robust against saliency manipulations, even below threshold, showing accuracy gains near chance performance.
• Peak differences between Gestalt and control conditions occurred during response saturation, mirroring neuronal contrast response profiles.
• Gestalt perception was associated with steeper psychometric functions, indicating enhanced discriminability.

Conclusions:

• The findings support the
• Gestalts as predictions
• hypothesis by demonstrating the influence of Gestalt phenomena even in sub-threshold conditions.
• Results suggest differential engagement of object-selective areas and percept-based processing in Gestalt perception.
• Nonlinearities in psychometric functions point to top-down modulation of early visual cortex, consistent with predictive coding frameworks.