Ocular drift shakes the stationary view on pattern vision.

Area of Science:

• Neuroscience
• Computational Vision
• Visual Perception

Background:

• Spatial vision traditionally assumes static visual input, ignoring involuntary eye movements like ocular drift.
• Ocular drift continuously alters visual input, emphasizing high spatial frequencies.
• The precise role of ocular drift in processing visual features, such as edges, is not fully understood.

Purpose of the Study:

• To investigate the influence of ocular drift on edge sensitivity in noisy visual stimuli.
• To evaluate whether incorporating ocular drift into a mechanistic model of spatial vision improves predictive accuracy.
• To understand how standard spatial vision models handle dynamic visual input.

Main Methods:

• Examined the effect of drift-induced stimulus power shifts on empirical data.
• Developed and tested a drift-enhanced mechanistic model of spatial vision.
• Compared the performance of the drift-enhanced model against the original model and a simpler single-channel model.

Main Results:

• Drift-induced shifts in stimulus power improved predictions of empirical data, aligning with human contrast sensitivity.
• The original spatial vision model outperformed the drift-enhanced version, suggesting inherent compensation mechanisms.
• A simpler single-channel model benefited from drift but performed poorly without it.

Conclusions:

• Standard spatial vision models may implicitly compensate for ocular drift, mimicking dynamic input processing.
• Current model architectures might favor stationary assumptions, potentially leading to self-confirming theories.
• Accurate modeling of visual feature extraction may necessitate incorporating the dynamic nature of visual input and reassessing existing frameworks.