Hiromasa Kitazawa

Memory trace of motor learning shifts transsynaptically from cerebellar cortex to nuclei for consolidation

Adaptation of ocular reflexes is a prototype of motor learning. While the cerebellum is acknowledged as the critical site for motor learning, the functional differences between the cerebellar cortex and nuclei in motor memory formation are not precisely known. Two different views are proposed: one that the memory is formed within the cerebellar flocculus, and the other that the memory is formed within vestibular nuclei. Here we developed a new paradigm of long-term adaptation of mouse horizontal optokinetic response eye movements and examined the location of its memory trace. We also tested the role of flocculus and inferior olive in long-term adaptation by chronic lesion experiments. Reversible bilateral flocculus shutdown with local application of 0.5 microl-5% lidocaine extinguished the memory trace of day-long adaptation, while it very little affected the memory trace of week-long adaptation. The responsiveness of vestibular nuclei after week-long adaptation was examined by measuring the extracellular field responses to the electrical stimulation of vestibular nerve under trichloroacetaldehyde anesthesia. The amplitudes and slopes of evoked monosynaptic field response (N1) of week-long adapted mice were enhanced around the medial vestibular nucleus compared with those of control mice. Chronic flocculus or inferior olive lesions abolished both day and week-long adaptations. These results suggest that the functional memory trace of short-term adaptation is formed initially within the cerebellar cortex, and later transferred to vestibular nuclei to be consolidated to a long-term memory. Both day and week-long adaptations were markedly depressed when neural nitric oxide was pharmacologically blocked locally and when neuronal nitric oxide synthase was ablated by gene knockout, suggesting that cerebellar long-term depression underlies both acquisition and consolidation of motor memory.

Role of Primate Cerebellar Hemisphere in Voluntary Eye Movement Control Revealed by Lesion Effects

The anatomical connection between the frontal eye field and the cerebellar hemispheric lobule VII (H-VII) suggests a potential role of the hemisphere in voluntary eye movement control. To reveal the involvement of the hemisphere in smooth pursuit and saccade control, we made a unilateral lesion around H-VII and examined its effects in three Macaca fuscata that were trained to pursue visually a small target. To the step (3 degrees)-ramp (5-20 degrees/s) target motion, the monkeys usually showed an initial pursuit eye movement at a latency of 80-140 ms and a small catch-up saccade at 140-220 ms that was followed by a postsaccadic pursuit eye movement that roughly matched the ramp target velocity. After unilateral cerebellar hemispheric lesioning, the initial pursuit eye movements were impaired, and the velocities of the postsaccadic pursuit eye movements decreased. The onsets of 5 degrees visually guided saccades to the stationary target were delayed, and their amplitudes showed a tendency of increased trial-to-trial variability but never became hypo- or hypermetric. Similar tendencies were observed in the onsets and amplitudes of catch-up saccades. The adaptation of open-loop smooth pursuit velocity, tested by a step increase in target velocity for a brief period, was impaired. These lesion effects were recognized in all directions, particularly in the ipsiversive direction. A recovery was observed at 4 wk postlesion for some of these lesion effects. These results suggest that the cerebellar hemispheric region around lobule VII is involved in the control of smooth pursuit and saccadic eye movements.

Loss of adaptability of horizontal optokinetic response eye movements in mGluR1 knockout mice

Metabotropic glutamate receptor subtype 1 (mGluR1) plays an essential role in the cerebellar long-term depression (LTD). We examined the dynamic characteristics and adaptability of horizontal vestibulo-ocular reflex (HVOR) and optokinetic response (HOKR) eye movements in mGluR1 knockout mice. A mild difference was seen in the HOKR/HVOR dynamics between the wild-type and mGluR1(-/-) mice. Exposure to 1 h of sustained screen oscillation, which induced HOKR adaptation in wild-type mice, induced no change in mutant mice. These results suggest that the mGluR1 plays an essential role in the adaptation of HOKR, and LTD underlies the adaptation of ocular reflexes.

Effects of reversible pharmacological shutdown of cerebellar flocculus on the memory of long-term horizontal vestibulo-ocular reflex adaptation in monkeys

The adaptation of the horizontal vestibulo-ocular reflex (HVOR) provides an experimental model for motor learning. Two studies, using cats and mice, respectively, have recently suggested pharmacologically that the memory of adaptation is located multiply in the cerebellum and brainstem. Here, we examined the effects of acute cerebellar flocculus shutdown on the adaptation in four monkeys. Two hours of 0.11Hz-10° turntable oscillation while viewing a stationary checked-patterned screen through the left-right reversing prism decreased the HVOR gains by 0.16, and 3 days of prism wearing combined with 2h of daily turntable oscillation decreased the HVOR gains by 0.27. Injections of lidocaine into bilateral flocculi did not affect the nonadapted HVOR gains, but depressed the visual suppression of the HVOR. They recovered the HVOR gains decreased by 2h of training, but very little affected the HVOR gains decreased by previous 2 days of training. Injections of control Ringers solution did not affect the gains adapted by 2h or 3 days of training. These results are consistent with the previous studies, and suggest that the memory trace of adaptation of the HVOR initially resides in the flocculus but later resides, presumably, in the vestibular nuclei in the monkey.

Role of primate cerebellar lobulus petrosus of paraflocculus in smooth pursuit eye movement control revealed by chemical lesion

The primate lobulus petrosus (LP) of the cerebellar paraflocculus receives inputs from visual system-related pontine nuclei, and projects to eye movement-related cerebellar nuclei. To reveal a potential involvement of LP in oculomotor control, we lesioned LP unilaterally by local injections of ibotenic acid in three Macaca fuscata. We examined the effects of lesion on eye movements evoked by step (3 degrees )-ramp (5-15 degrees/s) moving target. To step-ramp moving target, the monkeys showed an initial slow eye movement and later a small catch-up saccade, which was followed by the post-saccadic pursuit nearly matching to the velocity of the ramp target motion. After LP lesioning, the velocity of post-saccadic pursuits in the ipsiversive and down-ward directions decreased by 20-40% in all three monkeys. These deficits lasted for at least 1 month, and some recovery was observed. In the amplitudes of catch-up saccades, no consistent changes were seen among the three monkeys after LP lesioning. These results suggest an involvement of LP in the primate smooth pursuit eye movement control.

Subdural applications of NO scavenger or NO blocker to the cerebellum depress the adaptation of monkey post-saccadic smooth pursuit eye movements.

We examined pharmacologically whether cerebellar long-term depression (LTD) may play a role in the adaptation of smooth pursuit eye movements in two Macaca fuscata, which were trained to pursue a target moving in the horizontal plane in a 3° step-10 deg/s ramp mode. The monkeys showed small catch-up saccades followed by 6–8 deg/s post-saccadic pursuit movements. Adaptation of the post-saccadic pursuit velocity was induced by repetition of acceleration of the target to 20 deg/s after the catch-up saccades. Injections of 0.1 mM hemoglobin or 20 mM NG-monomethyl-L-arginine solution into the subdural space above the paraflocculus-flocculus scarcely affected the post-saccadic pursuit velocity, but markedly depressed its adaptation. These observations suggest that cerebellar LTD may underlie the adaptation of smooth pursuit.

Adaptive modifications of post-saccadic smooth pursuit eye movements and their interaction with saccades and the vestibulo-ocular reflex in the primate

Adaptation of the horizontal smooth pursuit eye movement was examined using step-ramp moving target paradigms in chronically prepared Macaca fuscata. Monkeys were trained to pursue a small target which moved in the horizontal plane in a 3 degrees step-10 degrees/s ramp or a 0 degrees step-10 degrees/s ramp mode for 300-400 ms. When the target moved from central fixation point in a step-ramp mode, the monkeys usually responded with an initial pursuit eye movement (latency, 100-120 ms) which reached to a nearly constant velocity of 4 degrees/s in 50 100 ms, followed by a 1-3.5 degrees catch-up saccade (latency, 170-230 ms). The catch-up saccade was followed by a 7-8 degrees/s post-saccadic pursuit. The post-saccadic pursuit velocity was measured 40-90 ms after the end of the catch-up saccade. When the target velocity was doubled (20 degrees/s) for 100-200 ms immediately after the onset of the catch-up saccade, the post-saccadic pursuit velocity increased by 40%. When the target velocity was decreased by half (5 degrees/s) immediately after the onset of the catch-up saccade for 150 ms, the post-saccadic pursuit velocity decreased by 30%. These increases or decreases of post-saccadic pursuit were observable within just 50-150 trials. The adaptation of post-saccadic pursuit occurred independent of the position of the target. The amplitude and latency of the catch-up saccade also increased correspondingly when the post-saccadic pursuit velocity was adaptively increased. Adaptation of smooth pursuit did not affect the dynamics of reflex eye movements, including the horizontal vestibulo-ocular reflex (HVOR) gain and phase measured by 0.33 Hz-10 degrees (peak-to-peak) turntable oscillations in darkness. Conversely, adaptation of the HVOR gain induced by a 2 h sustained oscillation of the turntable and screen in reversed direction at 0.33 Hz-10 degrees affected little the velocity of post-saccadic pursuit or the amplitude of the catch-up saccade. These results suggest that different neural mechanisms are respectively involved for the adaptation of horizontal smooth pursuit and HVOR in the primate.

Acute effects of tetrahydrobiopterin on the dynamic characteristics and adaptability of vestibulo-ocular reflex in normal and flocculus lesioned rabbits

The acute effects of (6R)-5,6,7,8-tetrahydro-L-biopterin (R-THBP) on the dynamic characteristics of horizontal vestibulo-ocular reflex (HVOR) were examined in chronically prepared alert pigmented rabbits. The HVOR gain was measured by 10 degrees-0.1 Hz sinusoidal oscillation of the turntable in darkness. In control rabbits, intramuscular injection of R-THBP (10 mg/kg) induced an increase of HVOR gain by 0.1-0.2, which lasted for 1-3 h. Sustained 10 degrees-0.1 Hz sinusoidal oscillation of the turntable and screen in the reversed direction induced an adaptive increase of HVOR gain by 0.2 in 3 h, which was not affected by applications of R-THBP. In flocculus lesioned rabbits, the HVOR gain reduced by 23%, and applications of R-THBP no longer induced a HVOR gain increase. These experimental observations suggested that the R-THBP affected on the HVOR dynamics within the cerebellar flocculus or its related neural areas.

Difference of climbing fiber input sources between the primate oculomotor-related cerebellar vermis and hemisphere revealed by a retrograde tracing study

The cerebellar flocculus-paraflocculus complex, vermal lobule VII (V-7) and hemispheric lobule VII (H-7) are involved in learning-dependent smooth pursuit eye movement control. To locate the sources of climbing fiber inputs to the H-7 and V-7, we injected retrograde tracers and examined the locations of retrogradely labeled neurons in the inferior olive in 4 monkeys. After the injection of cholera toxin B (CTB) into the H-7, retrogradely labeled neurons were observed abundantly in cell group d, i.e., dorsal cap, of the caudal medial accessory olive (MAO) and ventral lamella of principal olive (PO). After injections of fast blue (FB) into the V-7, retrogradely labeled neurons were observed mainly in cell group b of MAO, but rarely in cell group d or PO. Cell group d is known to receive inputs from the nucleus optic tract (NOT) and project climbing fibers to the flocculus and ventral paraflocculus, and cell group b is known to receive inputs from the superior colliculus. These results suggest that the three oculomotor cerebellar areas may use different visual signals for the control of smooth pursuit: the flocculus-paraflocculus complex and H-7 receive visual climbing fiber inputs derived mainly from the NOT via cell group d, while the V-7 receive visual climbing fiber inputs derived mainly from the superior colliculus via cell group b.

Dynamic characteristics and adaptability of reflex eye movements of Fyn-kinase-deficient mice

Fyn-kinase is expressed widely in the entire brain, including the cerebellum. Fyn-kinase-deficient mice are known to exhibit hypersensitivity to ethanol. To evaluate the cerebellar functions of Fyn-kinase, we examined the dynamic characteristics of the horizontal optokinetic response (HOKR) and vestibulo-ocular reflex (HVOR) and its adaptability in Fyn-kinase-deficient mice. The HOKR was induced by sinusoidal oscillation of a checkered screen and the HVOR was induced by sinusoidal oscillation of a turntable in darkness. The HOKR gains of mutant mice were higher than those of the wild-type mice, and the HVOR phases of mutant mice were less advanced than those of the wild-type mice. However, no difference was noted in the adaptability of the HOKR induced by 1 h of sustained screen oscillation between the mutant and wild-type mice. The cerebellar functions appear to be unaffected by Fyn-kinase knockout.