We tested the prediction of a neuromimetic model of the saccade generators in which the rapid velocities of saccades are due an abrupt increase in the firing rate, the burst, in reciprocally innervating EBNs and IBNs.3,13
The model also predicted that a pause in the external inhibition is critical for the high firing rate during the burst, a mechanism known as PIR. Simulation of the pathological enhancement of PIR, by increasing Ih
, revealed high-frequency back-to-back saccades (i.e., saccadic oscillations). The model also predicted that enhanced Ih
would increase the frequency of saccadic oscillations, without affecting the amplitude (). In contrast, enhanced IT
would increase the amplitude of saccadic oscillations but decrease their frequency ().3
Our strategy for testing these model predictions was to determine if a selective blocker of IT
reduced the amplitude of physiological saccadic oscillations and increased their frequency. We also determined whether nonselective blockade of Ih
, by propranolol,7,8
would reduce the amplitude of oscillations. Our findings were consistent with both model predictions; the ethosuximide reduced the amplitude but increased the frequency of saccadic oscillations in both the normal subjects, and propranolol reduced the amplitude of oscillations in the patient with mSOLT.
The model also predicted that the frequency of oscillations is less sensitive to changes in IT than is the amplitude. depicts that, in the model, for the given change in IT, the amplitude of simulated saccadic oscillations is affected more than the frequency. Consistent with this prediction, the amount of change in the amplitude of oscillations was larger and reached statistical significance as compared to the amount of change in the frequency in both subjects.
There was an inter-axis variability in the effects of ethosuximide in both subjects. This variability could be explained, in part, by differences in the expression profile of CaV3 ion-channels (conducting IT
) at anatomically separate brainstem sites of the burst neurons controlling vertical, torsional, and horizontal saccades. The burst neurons for horizontal saccades are found in caudal pons, while those for vertical and torsional movements are located in rostral mesencephalon.14
The conductance-based model predicted that mSOLT is caused by an inherited abnormality causing increased membrane excitability or reduced external inhibition of the pontine burst neurons.3
Our results provided support for this model, since non-selective blockade of Ih
with propranolol could theoretically reduce the excitability of the burst neurons, and, in turn, account for the reduction of the amplitude of saccadic oscillations in the mSOLT patient.
According to the model predictions, blockade of Ih
should reduce PIR and therefore reduce saccade velocity.3,15
However, in both our normal subjects, there was no change in saccade velocity. In our model, the strength of PIR required to simulate oscillations had to be supra-physiological; however, to simulate slow saccades, PIR must be significantly less than physiological. Therefore, it is possible that at the given doses, ethosuximide and propranolol might reduce PIR, but not enough to affect saccade velocity.
The recordings in the healthy subjects were done with scleral search coils, under closed eyelids. Does mechanical friction due to lid closure affect the eye velocity or induce coil slippage? We had addressed this issue in an independent study where we had shown that the acceleration and velocity of the eye movement during vestibulo-ocular reflex evoked by head impulses (the mechanism that is not physiologically affected by eye closure) were the same during open and closed eyelids.16
In the same study, the peak velocities or accelerations of the eye movements evoked by vestibulo-ocular reflex during head impulses are of comparable speed to those of saccades.16
Therefore, it is unlikely that eye closure induced mechanical hindrance to the movement of the eye or of the search coil mounted on it.
To summarize, the conductance-based model emphasized that increased excitability or disinhibition of the pontine burst neurons can cause saccadic oscillations. The results of the current experiments are in accord with the specific model predictions and in particular with the hypothesis that Ih and IT conductance determines the frequency and amplitude of the saccadic oscillations. There are, however, caveats. The drugs could have effects on other structures related to saccade generation that might influence the frequencies and amplitudes of saccadic oscillations. Electrophysiological experiments investigating the effects of selective ion channel blockers on saccades, injected locally at the location of the burst neurons or omnipause neurons, are needed to further test these ideas.