Yoichi Oda

Properties of cerebello-precerebellar reverberating circuits

Intracellular recordings were made from neurons of the red nucleus (RN) in cats where the cerebellar cortical effects were removed by chronic ablation of the intermediate part of the anterior lobe of the cerebellum. A prolonged depolarization could be elicited by stimulating the nucleus interpositus (IP) of the cerebellum, nucleus reticularis tegmenti pontis (NRTP) and the nucleus reticularis paramedianus (PMRN). This prolonged depolarization was abolished after cooling the inferior and middle cerebellar peduncles and persisted after ablation of the cerebral sensorimotor cortex. The prolonged depolarization was also recorded intracellularly from IP neurons. It was concluded that the prolonged depolarization set up in RN neurons is due to the repetitive discharges of IP neurons which produces tonic bombardment onto RN cells. The mechanisms of the repetitive discharges of IP neurons are considered to be due to impulse reverberation via the IP. The dynamic properties of the reverberating circuits were characterized by regenerative behavior. Above and below threshold, there were two states, the excited state where many constituent neurons were active, and the resting state where all neurons were inactive. It was found that cats with chronically stimulated cerebral peduncle (CP), and tested in an acute experiment, showed sometimes effective for inducing the prolonged depolarization and repetitive discharges of RN neurons by stimulation of IP. The prolonged depolarization thus produced could be reversibly abolished by cooling the middle and inferior cerebellar peduncles. The possible constituent neurons of the reverberating circuits were investigated in light of previous physiological investigations of stimulating the NRTP, PMRN, nucleus reticularis lateralis (LRN), nucleus olivaris inferior (IO) and recording EPSPs in RN cells. The RN cells receive axon reflex activation from NRTP and PMRN, and disynaptic excitation from NRTP, PMRN, LRN and IO. Based on these and other available data, the components of the cerebello-precerebellar reverberating circuits are discussed.

Simple and complex spike activities of purkinje cells during locomotion in the cerebellar vermal zones of decerebrate cats

SummaryIn walking cats decerebrated at the premammillary level, single neurone activity of Purkinje cells (P-cells) with long corticofugal axons was recorded in the cerebellar vermis. The P-cells (N = 145) were identified as they showed spontaneous simple and complex spikes and also antidromic activation from Deiters nucleus. These P-cells were classified into 6 groups according to the receptive fields of the climbing fibre responses (CFRs) which were evoked by electrical stimulation in each limb at the radial and sciatic nerve bundles. One group designated as forelimb units received the CFRs from both forelimbs and from neither hindlimb. According to previous studies, this group of P-cells is thought to make inhibitory connections with Deiters neurones projecting to the ipsilateral cervicothoracic spinal cord.For the forelimb units, two types of discharge patterns for simple spikes were found in relation to limb movements during locomotion. Type I cells showed one peak in their firing rate in the late swing (E1) or early stance (E2) phase of the ipsilateral forelimb. Type II cells showed two peaks and two valleys during one step cycle: one peak was in the E1 phase, the other in the late stance (E3) or early swing (F) phase; each of the two valleys followed the peak. Complex spikes of the forelimb units occurred more frequently in the E1 phase than during the other phases. The increased activity of simple and complex spikes of the forelimb units in the E1 phase is suggested to have a functional significance in preparing the appropriate floor reaction forces that appear upon touchdown of the ipsilateral forelimb.

Inhibitory long-term potentiation underlies auditory conditioning of goldfish escape behaviour

Long-term potentiation (LTP), the increase in synaptic strength evoked by high-frequency stimulation, is often considered to be a cellular model for learning and memory. The validity of this model depends on the assumptions that physiological stimuli can induce LTP in vivo and that the resulting synaptic modifications correlate with behavioural changes. However, modifiable synapses are generally embedded deep in complex circuits. In contrast, the goldfish Mauthner (M)-cell and its afferent synapses are easily accessible for electrophysiological studies, and firing of this neuron is sufficient to trigger fast escape behaviour in response to sudden stimuli,. We have previously shown that tetanic stimulation can induce LTP of the feedforward inhibitory synapses that control the excitability of the M-cell,. Here we report that natural sensory stimulation can induce potentiation of this inhibitory connection that resembles the LTP induced by afferent tetanization. Furthermore, comparable acoustic stimulation produced a parallel decrease in the probability of the sound-evoked escape reflex. Thus we demonstrate for the first time, to our knowledge, a behavioural role for the long-term synaptic strengthening of inhibitory synapses.

Classical Conditioning Mediated by the Red Nucleus in the Cat

We have attempted to develop a behavioral and neuronal model for classical conditioning in the corticorubrospinal system. A conditioned stimulus (CS) was applied to the cerebral peduncle (CP) in cats which had lesions that interrupted the corticofugal fibers caudal to the red nucleus. The unconditioned stimulus (US) was an electric shock to the skin of the forelimb that produced flexion of the limb. After pairing of the CS and US in close temporal association, an initially ineffective stimulus to the cerebral peduncle was found to give rise to the flexion of the elbow. Extinction of the conditioned response was achieved by applying the CS alone or by reversing the sequence of the stimuli (US-CS: backward pairing). Furthermore, the US alone did not produce an increase in the effectiveness of the CS stimulus. Finally, pairing the fixed CS stimuli with the US at random intervals did not produce any increase in performance in response to the CS. In these respects, the observed behavioral modification has the features of associative conditioning. Because the thresholds for and the strength of elbow flexion induced by stimulation of the nucleus interpositus of the cerebellum were identical in the experimental and control animals, the interpositorubrospinal system cannot be the site of the plastic change. Since the conditioned response is most probably mediated by the corticorubrospinal system, it is likely that a modification of the corticorubral synapses underlies this behavioral change.

Formation of functional synapses in the adult cat red nucleus from the cerebrum following cross-innervation of forelimb flexor and extensor nerves: I. Appearance of new synaptic potentials

We investigated the effects of cross-innervating the peripheral forelimb flexor and extensor nerves of adult cats on the time course of corticorubral EPSPs. Red nucleus neurons were identified by antidromic invasion from C1 or L1 spinal segments as innervating the upper spinal segments (C-cells) or sending axons to the lumbosacral cord (L-cells). In C-cells, a fast-rising component, superimposed on the slow-rising corticorubral EPSPs induced by the cerebral sensorimotor cortex or the cerebral peduncle (CP) stimulation, was noted. The mean time-to-peak of this component in cross-innervated cats operated more than two months earlier was 1.9 +/- 0.9 ms (n = 160), shorter than in normal cats (3.6 +/- 1.4 ms, n = 100). The same value in cats cross-innervated less than two months before was 2.7 +/- 1.0 ms (n = 53). The mean time-to-peak of CP-EPSPs from L-cells was 2.9 +/- 0.9 ms (n = 115). The fast-rising component had a latency of 0.96 +/- 0.19 ms (n = 122), and it was mediated by fibers with conduction velocities of less than 20 m/s. The projective area of the fast-rising component is organized somatotopically. Since it is more sensitive to membrane hyperpolarization than slow rising corticorubral EPSPs, it is mediated by synapses located more proximally than the corticorubral synapses of normal cats. The time course of facilitation by preceding cerebral peduncle stimulation of the nucleus interpositus (IP)-induced RN population responses was measured. It was characterized by a rapid, followed by a slower, rise time in the RN region where C-cells are concentrated. In contrast, the L-cell region was characterized by a slow rise time. In cats subjected to self-union of the peripheral flexor and extensor nerves, the majority of C-cells had CP-EPSPs with a time-to-peak within the normal range. Our results suggest that after cross-innervation sprouting and formation of functional synapses occur on the proximal portion of the soma-dendritic membrane of red nucleus neurons.

Cerebellar Control of Locomotion Investigated in Cats: Discharges from Deiters' Neurones, EMG and Limb Movements during Local Cooling of the Cerebellar Cortex

Publisher Summary This chapter investigates the neuronal mechanisms of the cerebellar control of locomotion, using a stepping preparation where the 4 limbs of cats are decerebrated at the thalamic level. Any electrical stimulation to induce locomotion is avoided as it might give artificial signals to the neuronal circuitry concerning coordinative control of stepping. In the stepping preparations, the activity of Deiters neurones projecting to the ipsilateral lumbosacral cord (L-Deiters neurones), the EMGs of hindlimb muscles, and limb movements are analyzed before, during, and after local cooling of the cerebellar cortex. During each cycle of quadrupedal stepping, most L-Deiters neuroncs show two peaks of impulse discharges—the first (A peak) in the stance phase and the second (B peak) in the swing phase of the ipsilateral hindlimb. The A peak often starts rising shortly before the ipsilateral hindlimb is placed. When movements of either hind- or forelimbs are stopped, this frequency modulation in Deiters neurones is greatly depressed. It is indicated that coordinated movements of fore- and hindlimbs are essential in building up the frequency modulation in L-Deiters neurones.

Formation of functional synapses in the adult cat red nucleus from the cerebrum following cross-innervation of forelimb flexor and extensor nerves

SummaryWe investigated the effects of cross-innervating the peripheral forelimb flexor and extensor nerves of adult cats on the time course of corticorubral EPSPs. Red nucleus neurons were identified by antidromic invasion from C1 or L1 spinal segments as innervating the upper spinal segments (C-cells) or sending axons to the lumbosacral cord (L-cells).In C-cells, a fast-rising component, superimposed on the slow-rising corticorubral EPSPs induced by the cerebral sensorimotor cortex or the cerebral peduncle (CP) stimulation, was noted. The mean time-to-peak of this component in cross-innervated cats operated more than two months earlier was 1.9 ± 0.9 ms (n = 160), shorter than in normal cats (3.6 ± 1.4 ms, n = 100). The same value in cats cross-innervated less than two months before was 2.7 ± 1.0 ms (n = 53). The mean time-to-peak of CP-EPSPs from L-cells was 2.9 ± 0.9 ms (n = 115).The fast-rising component had a latency of 0.96 ± 0.19 ms (n = 122), and it was mediated by fibers with conduction velocities of less than 20 m/s. The projective area of the fast-rising component is organized somatotopically. Since it is more sensitive to membrane hyperpolarization than slow rising corticorubral EPSPs, it is mediated by synapses located more proximally than the corticorubral synapses of normal cats.The time course of facilitation by preceding cerebral peduncle stimulation of the nucleus interpositus (IP)-induced RN population responses was measured. It was characterized by a rapid, followed by a slower, rise time in the RN region where C-cells are concentrated. In contrast, the L-cell region was characterized by a slow rise time.In cats subjected to self-union of the peripheral flexor and extensor nerves, the majority of C-cells had CP-EPSPs with a time-to-peak within the normal range.Our results suggest that after cross-innervation, sprouting and formation of functional synapses occur on the proximal portion of the soma-dendritic membrane of red nucleus neurons.

Synaptic plasticity in the red nucleus and learning

Pairing of the stimulus to the cerebral peduncle (CP) with that to the forearm skin leads cats to flex their forearms within a 10-day training period in response to stimulus to CP, which was initially ineffective. Behavioral study and extracellular unit analysis suggested that the cellular mechanism for this conditioning lies at the corticorubral (CR) synapses. Since formation of new CR synapses occurs in parallel with the recovery from behavioral deficits after brain damage and peripheral nerve cross-innervation, we explored the possibility that the formation of new CR synapses underlies conditioning. We investigated the time course of the CR excitatory postsynaptic potentials (EPSPs) as well as the distribution of the CR synapses on the somadendritic membrane of the red nucleus neurons and compared them with those observed in control animals. In conditioned animals, the times-to-peak of the CR EPSPs were significantly shorter than those in control animals. Electron microscopic studies demonstrated that more CR synapses make contact with large, i.e. proximal, dendrites and somata of red nucleus neurons in conditioned cats than in control ones. These results support the view that the formation of new synapses on the proximal dendrites and soma underlies classical conditioning in the cat.

Formation of new corticorubral synapses as a mechanism for classical conditioning in the cat

An electron microscopic quantitative study of corticorubral synapses was performed in the cat which acquired classical conditioning. Conditioned stimulus was applied to the cerebral peduncle and the unconditioned stimulus was an electrical shock to the forelimb skin. The proportion of corticorubral synapses contacting with somata and proximal dendrite was increased after conditioning. It was suggested that collateral sprouting and the formation of new synapses underlie classical conditioning.

Electrophysiological evidence for formation of new corticorubral synapses associated with classical conditioning in the cat

The present study was performed to clarify whether or not structural plasticity of synaptic connections underlies classical conditioning mediated by the red nucleus (RN) in the cat. Conditioned forelimb flexion is established by pairing electrical conditioned stimuli (CS), applied to corticorubral fibers at the cerebral peduncle (CP), with a forelimb skin shock (the unconditioned stimulus, US), but not by applying the CS alone or by pairing the CS and US at random intervals. In our previous study, it was shown that the firing probability of rubrospinal neurons (RN neurons) in response to the CS was well correlated with acquisition of the conditioned forelimb flexion and that the primary site of neural change underlying establishment of the conditioned forelimb flexion was suggested to be at corticorubral synapses. In the present study, we investigated corticorubral excitatory postsynaptic potentials evoked by CP stimulation (CP-EPSPs), in order to identify the neuronal mechanism underlying establishment of classical conditioning. In normal cats, CP-EPSPs had a typical slow-rising phase, which has been attributed to the distal location of corticorubral synapses on the dendrites of RN neurons. In contrast, in animals that received paired conditioning, subsequent CP stimulation evoked potentials with a fast-rising time course. In control groups of cats that received CS alone, CS randomly paired with the US, or only the same surgical operations as the conditioned animals, most of the CP-EPSPs displayed slow-rising EPSPs that similar to those observed in normal cats. The mean time from onset to peak of the potentials in the conditioned animals was significantly shorter than that seen in other groups. Therefore, the appearance of a fast-rising potential correlates well with acquisition of the conditioned forelimb flexion. The amplitude of the fast-rising potential was gradually changed with stimulus intensity. It had a short onset latency following CP stimulation (0.9 ms), which was similar to that of the slow-rising EPSP in normal cats. It followed high-frequency stimulation up to 100 Hz. These results suggest that the newly appearing, fast-rising potential was a monosynaptically evoked EPSP. Fast-rising EPSPs were also induced by stimulation of the sensorimotor cortex (SM). Since the SM-EPSP was occluded by the CP-EPSP, the SM cortex is, at least in part, a likely source of fast-rising EPSPs. Fast-rising SM-EPSPs were also observed at the unitary level. The SM-EPSPs in the conditioned animals exhibited somatotopical representation in their cortical origin, as has been described in normal cats. The electrotonic length was calculated from the voltage transient responses to current steps injected into the RN neurons. There was no concomitant change in the electrotonic length following the classical conditioning. Furthermore, the fastrising EPSPs were often observed as if they were superposed on the slow-rising EPSPs that were observed in normal animals. These observations suggest that the appearance of fast-rising EPSPs is due to the formation of new corticorubral synapses on the somata or the proximal dendrites of the RN neurons, and not as a result of a reduction in the electrotonic length of the RN neurons. The present study provides further evidence that this type of structural plasticity of synaptic connections underlies establishment of the classically conditioned forelimb flexion.