Pii: S0306-4522(96)00573-8

Spindling activity characterizes the EEG of animals and humans in the early stages of resting sleep. Spindles are defined as waxing and waning rhythmic waves at 7–14 Hz that recur periodically every 3–10 s. Spindling originates in the thalamus, but a role for the cerebral cortex in triggering and synchronizing thalamic spindles was shown by stimulation of the contralateral cortex avoiding antidromic activation of thalamocortical axons and by diminished coherency of thalamic spindles after hemidecortication. Spontaneous spindles under barbiturate anesthesia are waxing and waning but under ketamine–xylazine anesthesia or when evoked by strong stimuli spindle waves are almost exclusively waning, i.e. they start with maximum amplitude and then decrease progressively. Waxing and waning of spindles has been ascribed to progressive entrainment of units into the oscillation followed by a progressive desynchronization. Therefore, exclusively waning spindles would be produced by an initial high synchrony in the corticothalamic network. Such a situation is observable upon strong stimulation or, spontaneously, when spindles are paced by the slow cortical oscillation and preceded by a strong corticothalamic drive. We have conducted experiments in naturally sleeping cats to verify the occurrence of two patterns of spindle oscillations and to test the role of the cortex in synchronizing and shaping spindles. We have found that indeed two types of spindles (waxing and waning or mostly waning) occur in naturally sleeping animals. We also demonstrate that during cortical spreading depression spindles are less synchronous and only of the waxing and waning type. As cortical activity recovers, waning spindles reappear and are preceded by electroencephalogram deflections which are related to corticothalamic depolarizing inputs. Our results strongly support the hypothesis of the role of the cerebral cortex in shaping and synchronizing thalamically generated spindles. ? 1997 IBRO. Published by Elsevier Science Ltd. In order to present spindling in the EEG of the unanesthetized naturally sleeping cats, it is useful to describe first the patterns of spindles under ketamine–xylazine anesthesia, in which spontaneous activity in corticothalamic network is high, and under barbiturate anesthesia, in which the background activity is considerably lower. Under ketamine–xylazine anesthesia (10–15 mg/ kg; 2–3 mg/kg, i.m.), spindles are preceded by a depth-positive cortical EEG wave that ends with a sharp negative wave followed by a waning spindle sequence, usually at the upper frequency limit of spindling (13–14 Hz) (Fig. 1, Ketamine). Under barbiturate anesthesia (25–35 mg/kg), spindling is waxing and waning, and its frequency is lower with increasing the barbiturate doses (Fig. 1, Barbiturate). During natural slow wave sleep, two types of spindle pattern were observed, one with a characteristic waxing and waning pattern (Fig. 1, Natural sleep, right panel), while in the other spindles were preceded by an EEG biphasic complex (depthpositive, depth-negative) followed by a spindle sequence lacking the initial waxing feature (Fig. 1, Natural sleep, left panel). To test the hypothesis that the sharp depthnegative EEG deflections of the cortical slow oscillation trigger and synchronize spindle sequences in the thalamus thus changing their shape, cortical depression was induced by placing a drop of concentrated potassium acetate (3 M) in the vicinity of recording electrode 1 (n=6). The potassium solution induced a complete flattening of the local EEG that spread slowly (12 mm/min) towards electrode 8 (Fig. 2, top panel, spreading depression). The progressive spread of cortical depression was reflected as a complete flattening of the EEG trace recorded by the corresponding tungsten electrode, with a progression from the middle to the posterior suprasylvian gyrus (Fig. 2, a–d). After a short period (50–80 s) of almost complete absence of cortical activity (Fig. 2, d), spindling activity reappeared, and it was first visible *To whom correspondence should be addressed. Pergamon Neuroscience Vol. 77, No. 4, pp. 933–936, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 PII: S0306-4522(96)00573-8