Effects of reversible inactivation of the primate mesencephalic reticular formation. I. Hypermetric goal-directed saccades.

Single-neuron recording and electrical microstimulation suggest three roles for the mesencephalic reticular formation (MRF) in oculomotor control: 1) saccade triggering, 2) computation of the horizontal component of saccade amplitude (a feed-forward function), and 3) feedback of an eye velocity signal from the paramedian zone of the pontine reticular formation (PPRF) to higher structures. These ideas were tested using reversible inactivation of the MRF with pressure microinjection of muscimol, a GABA(A) agonist, in four rhesus monkeys prepared for chronic single-neuron and eye movement recording. Reversible inactivation revealed two subregions of the MRF: ventral-caudal and rostral. The ventral-caudal region, which corresponds to the central MRF, the cMRF, or nucleus subcuneiformis, is the focus of this paper and is located lateral to the oculomotor nucleus and caudal to the posterior commissure (PC). Inactivation of the cMRF produced contraversive, upward saccade hypermetria. In three of eight injections, the velocity of hypermetric saccades was too fast for a given saccade amplitude, and saccade duration was shorter. The latency for initiation of most contraversive saccades was markedly reduced. Fixation was also destabilized with the development of macrosaccadic square-wave jerks that were directed toward a contraversive goal in the hypermetric direction. Spontaneous saccades collected in total darkness were also directed toward the same orbital goal, up and to the contraversive side. Three of eight muscimol injections were associated with a shift in the initial position of the eyes. A contralateral head tilt was also observed in 5 out of 8 caudal injections. All ventral-caudal injections with head tilt showed no evidence of vertical postsaccadic drift. This suggested that the observed changes in head movement and posture resulted from inactivation of the caudal MRF and not spread of the muscimol to the interstitial nucleus of Cajal (INC). Evidence of hypermetria strongly supports the idea that the ventral-caudal MRF participates in the feedback control of saccade accuracy. However, development of goal-directed eye movements, as well as a shift in the initial position following some of the cMRF injections, suggest that this region also contributes to the generation of an estimate of target or eye position coded in craniotopic coordinates. Last, the observed reduction in contraversive saccade latency and development of macrosaccadic square-wave jerks supports a role of the MRF in saccade triggering.

Effects of reversible inactivation of the primate mesencephalic reticular formation. II. Hypometric vertical saccades.

Electrical microstimulation and single-unit recording have suggested that a group of long-lead burst neurons (LLBNs) in the mesencephalic reticular formation (MRF) just lateral to the interstitial nucleus of Cajal (INC) (the peri-INC MRF, piMRF) may play a role in the generation of vertical rapid eye movements. Inactivation of this region with muscimol (a GABA(A) agonist) rapidly produced vertical saccade hypometria (6 injections). In three of six injections, there was a marked reduction in the velocity of vertical saccades out of proportion to saccade amplitude (i.e., saccades fell below the main sequence). This was associated with a moderate increase in saccade duration. Inadvertent inactivation of the INC could not account for these observations because vertical, postsaccadic drift was not observed. Similarly, pure downward saccade hypometria, the hallmark of rostral interstitial nucleus of the medial longitudinal fasciculus (riMLF) inactivation, was always preceded by loss of upward saccades in our experiments. We also found a downward and ipsiversive displacement of initial eye position and evidence of a contraversive head tilt following piMRF injections. Saccade latency was shorter after two of six injections. Simulation of a local feedback model provided three possible explanations for vertical saccade hypometria: 1) a shift in the input to the model to request smaller saccades, 2) a reduction of LLBN input to the vertical saccade medium lead burst neurons (MLBNs), or 3) an increase in the gain of the feedback pathway. However, when the second hypothesis was coupled to a shortened duration of the saccade trigger (i.e., the discharge of the omnipause neurons), the physiological observations of piMRF inactivation could be replicated. This suggested that muscimol had targeted structures that provided both long-lead burst activity to the MLBNs in the riMLF and were critical for reactivation of the omnipause neurons. Evidence of markedly reduced vertical saccade amplitude, curved saccade trajectories, increased saccade duration, and saccades that fall below the amplitude/velocity main sequence in these monkeys closely parallels the oculomotor findings of patients with progressive supranuclear palsy (PSP).