Fixational eye movements, natural image statistics, and fine spatial vision

Area of Science:

• Neuroscience
• Computational Neuroscience
• Visual Neuroscience

Background:

• Perception and motor control are traditionally viewed as distinct fields.
• Humans actively seek sensory information, influencing perception.
• Natural visual input is dominated by self-generated motion, not static scenes.

Purpose of the Study:

• To investigate the impact of retinal image motion during fixation on visual input statistics and neural encoding.
• To explore the theory that unstable fixation is an efficient information acquisition strategy.
• To challenge the separation of perception and motor control in neuroscience.

Main Methods:

• Analysis of theoretical and experimental results on visual input statistics.
• Examination of the effects of eye movements on retinal image motion.
• Review of neural encoding mechanisms in response to dynamic visual stimuli.

Main Results:

• Incessant eye motion during fixation equalizes power across spatial frequencies in the retinal stimulus.
• This process leads to compact neural representations and emphasizes fine spatial detail.
• Unstable fixation may facilitate temporal multiplexing of visual information.

Conclusions:

• Motor control, specifically eye movement, significantly contributes to early visual representations.
• Perception and behavior are intricately linked, challenging traditional neuroscience paradigms.
• Active, unstable fixation is an efficient strategy for acquiring information from natural scenes.