The hypothesis that nucleus reticularis thalami (RE) is the generator of spindle rhythmicity during electroencephalogram (EEG) synchronization was tested in acutely prepared cats. Unit discharges and focal waves were extracellularly recorded in the rostral pole of RE nucleus, which was completely disconnected by transections from all other thalamic nuclei. In some experiments, additional transections through corona radiata created a triangular island in which the rostral RE pole survived with the caudate nucleus, putamen, basal forebrain nuclei, prepyriform area, and the adjacent cortex. Similar results were obtained in two types of experiments: brain stem-transected preparations that exhibited spontaneous spindle sequences, and animals under ketamine anesthesia in which transient spindling was repeatedly precipitated during recording by very low doses of a short-acting barbiturate. Both spindle-related rhythms (7- to 16-Hz waves grouped in sequences that recur with a rhythm of 0.1-0.3 Hz) are seen in focal recordings of the deafferented RE nucleus. The presence of spindling rhythmicity in the disconnected RE nucleus contrasts with total absence of spindles in cortical EEG leads and in thalamic recordings behind the transection. Oscillations within the same frequency range as that of spontaneous spindles can be evoked in the deafferented RE nucleus by subcortical white matter stimulation. In deafferented RE cells, the burst structure consists of an initially biphasic acceleration-deceleration pattern, eventually leading to a long-lasting tonic tail. Quantitative group data show that the burst parameters of disconnected RE cells are very similar to those of RE neurons with intact connections. In the deafferented RE nucleus, spike bursts of RE neurons recur periodically (0.1-0.3 Hz) in close time-relation with simultaneously recorded focal spindle sequences. The burst occurrence of deafferented RE cells is greatly reduced after systemic administration of bicuculline. The preservation of both spindle-related rhythms in the disconnected RE nucleus, together with our recent experiments showing abolition of spindle oscillations in thalamic nuclei after lesions of RE nucleus (24), demonstrate that RE nucleus is the generator of spindle rhythms.

The effects of depriving thalamic relay and intralaminar nuclei from their reticularis thalami (RE) inputs were investigated in acute and chronic experiments on cat. In acutely prepared animals, two (frontal and parasagittal) thalamic transections were made; extracellular and intracellular recordings were performed in RE-disconnected thalamic nuclei. In chronic experiments, the RE nuclear complex was lesioned by means of kainic acid injections; the activity of RE-deprived thalamocortical neurons was extracellularly studied during wakefulness and synchronized sleep. Two features distinguish RE-deprived nuclei from normal thalamic nuclei: absence of spindle-wave rhythmicity and all-burst activity of neurons. The abolition of spindle-related rhythms (sequences of 7- to 14-Hz waves recurring periodically with a rhythm of 0.1-0.2 Hz) in RE-disconnected thalamic nuclei and ipsilateral neocortical areas contrasted with normal spindling rhythmicity in contralateral EEG leads. Spontaneously occurring, rhythmic, long-lasting inhibitory postsynaptic potentials (IPSPs), as observed in intact preparations, were no longer observed in RE-disconnected thalamic neurons. The remaining inhibitory events consisted of short-duration IPSPs. The possibility that RE nucleus is a pacemaker for spindling rhythms, imposing them through inhibitory projections to target thalamic areas, is supported by our concurrent experiments that indicate RE neurons preserve their rhythmicity after disconnection from their major (cortical and thalamic) input sources. RE-deprived thalamocortical neurons exclusively exhibit high-frequency spike bursts whose intrinsic structure is identical to that of intact thalamic relay cells. Instead of the spindle-related sequences of bursts seen in normal animals, the bursts of RE-disconnected thalamocortical neurons are single events, with a dramatic rhythmicity at 1-2 Hz. The presumed mechanism of this rhythmicity is the periodic activation of a low-threshold somatic conductance whose deinactivation is brought about by temporal integration of short-lasting IPSPs. It is known that high-frequency spike bursts of thalamic relay neurons result from hyperpolarization of cell membrane. We blocked the underlying inhibitory events by bicuculline and reversibly changed the all-burst activity of RE-disconnected neurons into a tonic mode. Since the only activity of RE-deprived thalamocortical neurons consists of burst discharges, we hypothesize that local-circuit GABAergic neurons are released from inhibition after RE disconnection or lesion.

This study tested the hypothesis that inhibitory actions are exerted by reticularis thalami (RE) neurons upon thalamocortical neurons. The RE neurons were recorded in the rostral pole and lateral districts of the nucleus, and were activated monosynaptically by cortical volleys. Thalamocortical neurons were identified antidromically in intralaminar and ventrolateral nuclei. During sleep with EEG synchronization, prolonged spike barrages of RE neurons extended over the whole spindle sequences. This result suggests that RE neurons are depolarized throughout spindle oscillations, whereas thalamocortical neurons show, simultaneously, long hyperpolarizations and short rebounds. During waking, parallelism rather than reciprocity was found between RE and thalamocortical neurons. Spontaneous discharge rates almost doubled in RE neurons on arousal from sleep, and the probability of cortically evoked short-latency discharges increased. The increase in spontaneous firing rates of RE neurons during natural arousal is consistent with their short-latency synaptic excitation by stimulating the rostral brain stem reticular formation after chronic degeneration of passing fibers. We suggest that RE cells inhibit GABAergic local-circuit cells, in addition to inhibiting thalamocortical neurons, and that different ratios of inhibitory effects are exerted by RE neurons upon these two cell classes during waking and sleep. We further suggest that, upon arousal, disinhibition of thalamocortical neurons (via the local-circuit neurons) outweighs direct inhibition of the thalamocortical neurons.

SUMMARY1. Unit discharges were extracellularly recorded from antidromically identified thalamocortical neurones of ventralis lateralis (v.1.) and centralis lateralis (c.l.) nuclei as well as from reticularis thalami (re.) neurones during wakefulness and electroencephalogram-synchronized sleep of the behaving cat. Various parameters of sleep-related discharge bursts were analysed.2. Statistical analyses revealed striking similarities between motor relay (v.1.) and intralaminar (c.l.) neurones. More than 60 % of bursts consist of three to five spikes at 250-400 Hz. The defining feature of bursts in all cortically projecting neurones is a progressive increase in the duration of successive interspike intervals. 3. As in thalamocortical cells, all re. neurones change their tonic discharges in waking to bursting firing in sleep, regardless of the increased or decreased firing rates from wake to sleep in individual neurones. The bursts of re. neurones are essentially different from those of thalamocortical cells. In re. neurones, burst structure consists of an initial progressive decrease in duration of interspike intervals, followed by an increase in duration of successive intervals, eventually leading to a long-lasting tonic spike train at about 100 Hz. In contrast with bursts of thalamocortical neurones, only 6 % of re. bursts are shorter than 50 ms; the total duration of the burst extends between 50 ms and 1'5 s. Population periburst histograms show the beginning of a decline in firing probability about 1-5 s prior -to burst onset and an increased firing probability persisting for 300-350 ms after burst onset.4. The different electrophysiological properties underlying the burst structure of cat's thalamocortical and re. neurones are discussed, with emphasis on dissimilar aspects of re. bursts in unanaesthetized and barbituratized preparations. Various factors that may account for the transition from tonic mode in waking to bursting mode in sleep are envisaged.

No abstract