Neural control of gaze shifts :neuroanatomical and neurophysiological approaches

We used transneuronal retrograde transport and electrical simulation to study the anatomy of the oculomotor system and the phyiology of eye‐head coordination, respectively. To obtain a more comprehensive view of the oculomotor system, we injected rabies virus in the lateral rectus (LR) extraocular muscle of primates. We took advantage of the fact that the transfer of rabies virus through synapses is time dependent to distinguish the brain reions providing monosynaptic, disynaptic and polysynaptic input to the extraocular motorneurons of the abducens nucleus (Abd). We were particularly interested in the distribution of oculomotor related neurons in the frontal cortex and the elucidation ofgradients of SC neurons that might serve as a substrate of the spatiotemporal transformation. In our physiological study, the feline SC was electrically stimulated and the electromyographic (EMG) activity of several neck muscles was monitored to test the influence of the eye position on the decomposition of the gaze commnd issued by SC into the eye and head movement components. In the frontal cortex, numerous labeled neurons were found in the arcuate sulcus (AS), occupying both of its banks, the spur, the prearcuate convexity and a small part of the postarcuate cortex. All in all, this area was found to be much bigger than the low threshold frontal eye fields (FEF). In addition, clusters of labelled cells were found in the dorsomedial frontal cortex, occupying an area that corresponded well to the supplementary eye fields (SEF). Additional clusters were found in the principal sulcus and in parts of the cigulate and orbitofrontal cortices. These data show that a network of higher level neuronal structures work in tandem to control eye movements. In the brainstem, numerous regions were oligosynaptically connected with the Abd. Besides areas implicated in horizontal and vertical eye movements, substantial labeling was observed in the cholinergic system of the brainstem, in pretectal and accessory optic nuclei, in parabrachial and paralemniscal areas, and in the deep cerebellar nuclei. In the SC, we estimated the density of labeled cells along the horizontal meridian (representing purely horizontal saccades) and found it to increase from rostal to caudal locations. This rostrocaudal gradient could only partially serve as the substrate of the spatiotemporal transformation, as it was not as steep as the one we deduced from both experimental data and modeling arguments. It may be the case that several mechanisms work in parallel to implement this transformation. In the gaze decomposition study the amplitude of the electrically evoked phasic EMG responses of neck muscles was often found o increase when the eyes deviated in their pulling direction. This increase in amplitude was always accompanied by a decrease in response latency. However, the effect of eye position was not consistent since eye position accounted for 10‐50% of the variance of EMG amplitude and in several SC sites there was no correlation between the two variables