EFFECT OF EYE POSITION WITHIN THE ORBIT ON ELECTRICALLY ELICITED SACCADIC EYE-MOVEMENTS - A COMPARISON OF THE MACAQUE MONKEYS FRONTAL AND SUPPLEMENTARY EYE FIELDS

1. We quantitatively compared the effects of eye position within the orbit on saccadic eye movements electrically elicited from two oculomotor areas of the macaque monkey's frontal lobe-cortex: the frontal eye field (FEF) and the supplementary eye field (SEF). 2. The effect of eye position on electrically elicited saccades was studied by delivering 70-ms trains of intracortical microstimulation while the monkeys fixated a spot of light. Tests of different fixation points located across a rectangular array were randomly intermixed. Complete experiments were carried out on 38 sites in three FEFs of two monkeys and 59 sites from three SEFs of the same two monkeys. Stimulation currents for the array experiments were usually 10-20 muA above the site threshold; the average current used was 36 muA for FEF and 49 muA for SEF. 3. The magnitude of effect of the initial eye position on the elicited saccade's dimensions was quantified at each site by computing the linear regression of saccadic eye movement displacement on the eye position within the orbit when stimulation was applied. This computation was done separately for the horizontal and vertical axes. We call the resulting pair of regression coefficients ''orbital perturbation indexes.'' Indexes of 0.0 represent elicited saccades that do not change their trajectory with different initial eye positions (constant-vector saccades), whereas indexes of - 1.0 represent elicited saccades that end at the same orbital position regardless of initial eye position (goal-directed saccades). 4. The effect of eye position varied across sites. In both FEF and SEF, the orbital perturbation indexes were distributed between approximately 0.0 and -0.5, with the horizontal and vertical indexes highly correlated across sites. 5. The average orbital perturbation indexes were small for both eye fields and were not significantly different. The mean horizontal indexes were -0.13 and -0.16 for SEF and FEF, respectively. The mean vertical indexes were -0.16 and -0.13. Neither SEF versus FEF difference was statistically significant. 6. In both SEF and FEF, sites yielding larger-amplitude saccades generally had larger orbital effects than sites yielding smaller saccades. This relationship accounted for the majority of the variability of the orbital perturbation indexes across sites in both SEF and FEF. 7. These results indicate that SEF and FEF are not distinguished from each other by the orbital dependence of their electrically elicited saccades. Thus they do not confirm the previously hypothesized dichotomy that FEF codes constant-vector saccades and SEF codes goal-directed saccades. 8. Instead, these data indicate a vectorial coding of saccades in both eye fields, with comparable orbital perturbation of elicited saccade vectors, especially for larger saccades, in both structures. The overall pattern of these perturbations seems to reflect a failure of cerebellar and brain stem circuits to adequately compensate for orbital mechanics, especially for the fact that centrifugal eye movements require greater muscular force than centripetal movements. We hypothesize that intracortical microstimulation in SEF and FEF often fails to adequately engage these compensatory circuits. 9. We also observed that the thresholds of elicited saccades often vary systematically as a function of initial eye position, and we postulated a model for orbital perturbation of saccade dimensions on the basis of this threshold variability. In this model, the elicited saccade vector obtained at a given initial eye position reflects a summation across the saccade vector representations excited by the stimulation, with each representation being weighted by its threshold at that eye position. Under this model, the degree of orbital perturbation varies with the heterogeneity of saccade direction and size representation in the neighborhood of the stimulating electrode. We suggest that this threshold effect, operating additively with inadequate cerebellar compensation for orbital mechanics as described above, could account for most of the orbital perturbation observed in FEF and SEF.