While we fixate our gaze, our eyes undergo smooth random motion called fixational drift. We are typically unaware of fixational drift, yet this motion is highly consequential for high acuity vision. In the past few decades, fixational drift has been studied extensively in the context of its influence on visual coding and inference. However, the mechanisms that are responsible for the generation of the motion have remained unknown. In particular, it has been unclear whether fixational drift arises from peripheral sources, such as noise in the ocular muscles or in the motoneurons that directly innervate them, or from central sources within the brain. The new work by Nadav Ben-Shushan, Nimrod Shaham, Mati Joshua, and Yoram Burak shows that fixational drift is correlated with neural activity, and identifies the origin of the motion in central neural circuitry, upstream of the ocular motoneurons. In addition, the work proposes that the main drive for the motion arises in a memory circuit in the brainstem, the oculomotor integrator, that maintains a memory of the desired eye position during fixation. Ben-Shushan et al propose that noise within this neural circuit drives continuous random drifts in the represented memory, and these drifts accumulate over time and drive the eye motion. To test this hypothesis, the researchers developed a detailed theoretical model of noisy neural dynamics within the oculomotor integrator and of the response dynamics of the ocular muscles, which is based on the parameters of the primate and human oculomotor system. The model explains very well the salient statistical features of the motion, and thus provides compelling evidence in support of the hypothesis.

Image description: Left, schematic figure showing the main ingredients in the theoretical model. Here, the components that control the horizontal position of the eye are illustrated. The oculomotor integrator is a neural network in the brainstem, responsible for holding the eyes still between saccades. The output of the oculomotor integrator network feeds into motoneurons in the brainstem that in turn drive the muscles, and the muscles control the horizontal potion of the eye. Right: an example of a trajectory of the eye (horizontal position) generated by the model (red), compared to a measured eye trajectory (blue). To facilitate the comparison, noise was added to the simulated trajectory to mimic measurement errors arising from the apparatus that was used to generate the blue trace.