Spatiotemporal patterns of spindle oscillations in cortex and thalamus

Abstract

Spindle oscillations (7-14 Hz) appear in the thalamus and cortex during early stages of sleep. They are generated by the combination of intrinsic properties and connectivity patterns of thalamic neurons and distributed to cortical territories by thalamocortical axons. The corticothalamic feedback is a major factor in producing coherent spatiotemporal maps of spindle oscillations in widespread thalamic territories. Here we have investigated the spatiotemporal patterns of spontaneously occurring and evoked spindles by means of multisite field potential and unit recordings in intact cortex and decorticated animals. We show that (1) spontaneous spindle oscillations are synchronized over large cortical areas during natural sleep and barbiturate anesthesia; (2) under barbiturate anesthesia, the cortical coherence is not disrupted by transection of intracortical synaptic linkages; (3) in intact cortex animals, spontaneously occurring barbiturate spindle sequences occur nearly simultaneously over widespread thalamic territories; (4) in the absence of cortex, the spontaneous spindle oscillations throughout the thalamus are less organized, but the local coherence (within 2-4 mm) is still maintained; and (5) spindling propagation is observed in intact cortex animals only when elicited by low intensity cortical stimulation, applied shortly before the initiation of a spontaneous spindle sequence; propagation velocities are between 1 and 3 mm/sec, measured in the anteroposterior axis of the thalamus; increasing the intensity of cortical stimulation triggers spindle oscillations, which start simultaneously in all leads. We propose that, in vivo, the coherence of spontaneous spindle oscillations in corticothalamic networks is attributable to the combined action of continuous background corticothalamic input initiating spindle sequences in several thalamic sites at the same time and divergent corticothalamic and intrathalamic connectivity.

Fig. 6.

Local spindling prevents participation in global spindling. Multiunit recordings and LFPs were obtained from eight thalamic sites. Average rates were calculated (bin size, 0.1 sec) and depicted with the corresponding LFP for each channel (TC1–TC8). The two first spontaneous spindle sequences occurred simultaneously in LFPs from all electrodes. Local spindling in TC7 and TC8 prevented next global spindling to reach those electrodes (arrows).

Fig. 1.

Cortical spindles are coherent oscillations during natural sleep. In the top panel(HUMAN), spindles were recorded from six standard EEG derivations (indicated in the schematic at right,arrowheads) in a normal subject during sleep stage 2. Cross-correlations of individual spindle sequences (n = 15) were calculated between C3A2 and each one of the other channels. Averaged correlations (CROSS) showed rhythmicity at 14 Hz and central peak values between 0.7 and 0.9. Bottom panel (CAT) shows EEG from a chronically implanted naturally sleeping animal. EEG was recorded from six tungsten electrodes separated by 1 mm, inserted in the depth of the suprasylvian gyrus (Ssylv), represented by dots 1–6 in the scheme at right; also in the scheme are represented the ectosylvian (Ecto.) and the marginal (Marg.) gyri [anterior (Ant.) and posterior (Post.) are indicated]. The same procedure as for the human EEG was used to obtain the averaged cross-correlations depicted at right(CROSS), showing correlation at 14 Hz with central peaks between 0.75 and 0.9.

Fig. 2.

Cortical coherence of spindling under barbiturate anesthesia does not depend on intracortical horizontal connections. Spontaneous spindle oscillations were recorded from the depth (∼1 mm) of the suprasylvian cortex by means of eight tungsten electrodes (scheme, dots numbered 1–8) with interelectrode distances of 1 mm. Spindling sequences in the raw data (top panel, cortical leads 1–8) showed frequencies at 7–9 Hz, lasted 2–5 sec, and recurred every 2–7 sec almost simultaneously in all electrodes. Cross-correlations between electrode 1 and each of the others, for consecutive individual spindles (n = 15), were averaged (below; CROSS, Before cut; traces were displaced horizontally and vertically for clarity) and showed central peak values decreasing from 0.9 (1–2) to 0.5 (1–8). After cut shows averaged cross-correlations (n = 15) calculated after a deep cut between electrodes 4 and 5(black line on the scheme) that crossed from the marginal gyrus (Marg.) to the ectosylvian gyrus (Ecto.). The histology of the cut is shown (right to the scheme of recording electrodes) in a parasagittal section along the suprasylvian gyrus (anterior atleft; calibration bar in mm); some tracks of recording electrodes are also seen. After the transection, cross-correlations showed a similar decrease in central peak from 0.9 (1–2) to 0.5 (1–8). Correlations1–4 and 1–5 were flat because of the local lesion produced by the cut.

Fig. 3.

Spindles are synchronized between the cortex and the thalamus. Top panel shows spontaneous spindle sequences under barbiturate anesthesia, recorded by two bipolar electrodes located in the depth of the suprasylvian gyrus and separated by 15 mm (Cx1 and Cx2) and by six tungsten electrodes in the anteroposterior axis of the thalamus (Th1–Th6). The arrangement of recording electrodes is depicted in the schematic below; the thalamic electrodes penetrated through the marginal gyrus (dotted lines). A detail of two spontaneous spindle sequences, indicated by bar, is expandedbelow (arrow). Note that spindles occurred nearly simultaneously in all leads.

Fig. 4.

Coherent spindle oscillations over large thalamic territories. Spontaneous spindling was recorded under barbiturate anesthesia from eight thalamic areas (Th1–Th8) separated by 1 mm in the anteroposterior axis of the thalamus, from anterior planes 13 to 6 (3 mm lateral from the midline, depth +3). Spindles were simultaneous in all leads except for shorter spindle sequences that occurred almost exclusively in most rostral electrode 1 (local spindles). Below,Seq. power spectrum was calculated for each channel from contiguous windows of 0.5 sec. Total power between 7–14 Hz from each window was normalized to the highest value in each channel and displayed against time. The values of power increased and decreased together in all channels. Cross-correlations were computed between individual spindle sequences (n = 15) from channel 1 and the others and averaged (Averaged CROSS;1–1 is the autocorrelogram of channel 1). The value of the central peak of the seven averaged cross-correlations is plotted atright (Peak CROSS) and shows a decrease from 0.8 to 0.6.

Fig. 5.

Thalamic cells discharge spike-bursts grouped in spindle sequences in close time relation throughout the thalamus. Multiunit recordings from TC cells were simultaneously obtained from eight tungsten microelectrodes (TC1–TC8), separated by 1 mm in the anteroposterior axis of the thalamus (position of most anterior electrode is indicated). Spontaneous spindling activity is shown in1. Spindle sequence indicated by horizontal bar is expanded at right, in 2. Average rates were computed for each channel (bin size, 0.1 sec), and time of first burst in TC1 was used as marker for aligning the other traces. Fifteen consecutive spindle sequences are shown below for four channels (TC1,TC2, TC7, and TC8; abscissa scale is in seconds). Dotted line indicates time of the first burst from TC1 in each spindle sequence. Time dispersion of burst firing increased with the distance to the reference cell.

Fig. 7.

Decortication of left hemisphere. Nissl-stained coronal section showing hemidecortication and cut of the corpus callosum. The black spot on the medial wall of the right hemisphere is attributable to crystals of AgNO₃ used against bleeding during decortication. Al, Abl, Lateral and basolateral nuclei of amygdala; CL, centrolateral intralaminar nucleus; LG, lateral geniculate nucleus;VP, ventroposterior complex.

Fig. 8.

Strong spatiotemporal coherence of thalamic spindle oscillations is lost after ipsilateral decortication. Eight thalamic foci (Th1–Th8), corresponding to the electrodes indicated in the scheme of Figure 7, were recorded simultaneously after complete ipsilateral decortication. Parts indicated by bars 1 and 2, expandedbelow, show examples of lack of temporal coordination between spindle sequences (in 1) but also epochs with well organized, simultaneous spindle sequences in many different thalamic foci (in 2; see also such coherent spindle sequences in the right part of the top panel). At bottom, Seq. power spectrum was calculated as in Figure 4. Overall, changes in total power were no longer simultaneous among different thalamic channels. However, asterisks (in the top panel with LFPs, as well as in power spectrum, showing the same spindle sequence) indicate one of the spindle sequences that were synchronized throughout the thalamus.

Fig. 10.

Cortical stimulation with low intensity triggers spindle sequences that travel away from the stimulating site. LFPs were recorded from eight tungsten electrodes (Cx1–Cx8) inserted in the depth (∼1 mm) of the suprasylvian cortex. Stimuli were applied through bipolar electrodes situated 4 mm in front and 4 mm behind the electrode array (see similar location of recording cortical electrodes in the scheme of Fig. 2 and similar location of stimulating cortical electrodes in the schematic with cortical recording electrodes of Fig.3). Synchronized spindling was triggered by high intensity stimulation in the posterior part of the suprasylvian gyrus (top panel; initial response expanded at right, with sweeps displaced horizontally). Low intensity stimulation at the same posterior site triggered spindling that propagated to more anterior foci (second panel; initial response, expanded atright, barely reached electrode Cx1). Anterior stimulation also gave rise to full synchrony of evoked spindles on high intensity volleys (third panel), whereas propagation was from anterior to posterior sites by using low intensity stimuli (bottom panel).

Fig. 9.

Spike-bursts from rostral RE cells may precede those from TC cells during spontaneous spindling after decortication. Multiunit recordings from the RE nucleus and six thalamic foci (TC1–TC6) were recorded simultaneously with 0.4 mm interelectrode distances. Spike-bursts from three spontaneous spindles are shown in which firing of RE cells consistently preceded TC firing. The first spindle sequence (bar) is expanded below(arrow). Average rates were computed (bin size, 0.1 sec), and first burst from RE cells at each spindle sequence was taken as time 0 to align the rate meters from the other cells. (Right column depicts RE, TC4, andTC6. Average includes rates from RE and all TC cells.) Fifteen consecutive spindle sequences show that spike-bursts from RE cells preceded those from TC cells. A tendency for increased delay with distance suggested that thalamic propagation may occur in decorticated conditions.

Fig. 11.

Intracortical horizontal connections are not implicated in spindling propagation. Same experimental conditions as in Figure 10. A deep coronal cut was performed between electrodes 4 and 5, which caused an important decrease in amplitude in those leads. Propagating spindle sequences were still observed in both directions (posterior-to-anterior and anterior-to-posterior).

Fig. 12.

Evoked spindle sequences display propagation in the thalamus when low intensity stimulation is applied to cortex. Multiunit recordings of TC cells were obtained from eight locations (TC1–TC8) separated by 1 mm in the anteroposterior axis of the thalamus (position of the most anterior electrode is indicated). Cortical stimulation is marked by anarrow. 1, Low intensity stimulation in the anterior suprasylvian gyrus generated a spindle sequence that propagated from anterior to posterior at a velocity of 1.5 mm/sec.2, Posterior cortical stimulation at low intensity generated a mirror image response.

Fig. 13.

Cortical stimulation synchronizes spindling oscillations throughout the thalamus. Multiunit recordings were obtained from the same positions as in Figure 12, and the average rates (bin size, 0.1 sec) were calculated. Left column depicts examples from three consecutive stimuli. Right columnshows averaged peristimulus histograms calculated from the rate meter data (n = 15). Low intensity cortical stimuli generated propagating spindle sequences when applied through the posterior (top panel, stimuli represented bydotted lines) or anterior (second panel) electrodes. Increasing the intensity of cortical stimulation triggered simultaneous spindle sequences in the eight thalamic electrodes.