Abstract

We recorded from single units in the pretectal nucleus of the optic tract (NOT) of the nonhuman primate. Specifically, we examined units that are modulated during smooth tracking of a small laser spot against a dark background. We used a nonlinear optimization procedure to determine whether the unit responses of these parafoveal cells are better described by a model that incorporates retinal error motion parameters or by a model that incorporates eye motion parameters. Our main finding was that all the cells in our sample group were better fit with a three-component model that incorporated retinal error motion parameters of position, velocity and acceleration (average coefficient of determination = 0.84) than a model that used position, velocity and acceleration components of eye motion (average coefficient of determination = 0.68). Other analyses involved comparison of goodness of fit between the three-component retinal error model and two-component retinal error models that excluded position or acceleration related terms. We found that there was a statistically significant degradation in the fit when position and acceleration related terms were dropped from the retinal error based model (P<0.05). Unit data from experiments in which the laser spot was extinguished for a brief period of time during tracking showed that the unit response was decreased following the target blink. We conclude on the basis of this and previous experimental data and our dynamic modeling approach that the parafoveal cells in the NOT primarily encode retinal error motion. Further they encode position, velocity and acceleration components of retinal error that could be used by other downstream structures for synthesis of a smooth-pursuit eye movement.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call