Task-dependent modulation of the sensorimotor transformation for smooth pursuit eye movements

Area of Science:

• Neuroscience
• Computational Neuroscience
• Vision Science

Background:

• Smooth pursuit eye movements are crucial for maintaining gaze on moving objects.
• Understanding how the brain transforms visual motion into eye movements is a key challenge in neuroscience.

Purpose of the Study:

• To investigate the transformation of retinal image velocity into smooth pursuit eye velocity.
• To examine how stimulus and task parameters influence the selectivity of pursuit eye movements.
• To develop a computational model explaining the observed pursuit behaviors.

Main Methods:

• Monkeys' eye movements were measured while viewing one or two moving targets.
• Experiments manipulated target features (color, motion direction) and attentional cues.
• A simple recurrent network model was developed to simulate pursuit behavior.

Main Results:

• When presented with two identical targets, pursuit eye velocity was the vector average of individual target responses.
• Attentional cues (target color) biased pursuit but did not eliminate the influence of distractors.
• Radial image motion promoted averaging, while nonradial motion favored a winner-take-all response.

Conclusions:

• The sensorimotor interface for pursuit employs a flexible decision rule, varying between vector averaging and winner-take-all.
• Computational elements like selection bias, competitive inhibition, and response normalization are critical for modeling smooth pursuit.