2001-2005 - The function of the rostral and caudal reticular thalamic nucleus in the origin of SWD and sleep spindles

By: E.Y. Sitnikova

In this project an electroenchephalograpic study of paroxysmal brain activity was carried out in a genetic model of absence epilepsy, the WAG/Rij strain of rats. By means of quantitative electroenchephalograpic analysis, we examined mechanisms of absence epilepsy that were aimed to identify the origin of the SWD. Two primary hypotheses were put forward:

H 1 considered the anatomical-physiological organization of neuronal brain circuits and suggested that the rostral pole of the reticular thalamic nucleus (RTN) might contain a pacemaker for the occurrence of SWD;

H 2 considered functional aspects of pathophysiology of the SWD and predicted a relationship between sleep spindles and SWD.

Cortical topography of sleep spindles and SWD allowed us to assume that anterior sleep spindles may be transformed into SWD type I (a generalized seizure) and posterior sleep spindles may give rise to local occipital discharges, SWD type II. Since the topography of sleep spindles is poorly described, we started our project with a detailed analysis of two types of sleep spindles: anterior and posterior sleep spindles. Particular attention was paid to the role of the RTN in the two spindle types, because we expected that dysfunction of the RTN is the most probable reason why sleep spindles are transformed into SWD. Further and additional analysis of EEG signals was performed to establish a relation between sleep spindles and SWD, to check weather spindle activity is a precursor of SWD I, to examine those changes of thalamo-cortical activity that take place at seizure onset. We also tried to shed some light upon the nature of SWD type II.

In our original proposal we had proposed to study the effects of lesions in the RTN (in the rostral and caudal poles) in WAG/Rij and ACI rats. As expected, lesions in the rostral pole of the RTN in WAG/Rij rats diminished both SWD I and frontal sleep spindles (Hypotheses 1 and 2) )(van Luijtelaar et al., 2001.) Further experiments were designed to test the role of the RTN in occipital sleep spindles and SWD type II. Micro-lesions (1 mm in size) in the (ventral part of the) rostral pole of the RTN did not show the expected effects since the incidence of SWD I was increased. In fact, neurons in the RTN are subjected to inner GABA-ergic self-inhibition, out lesions might have been to small to destroy completely the functioning of the RTN but, vice versa, caused a self-disinhibition and enhanced oscillatory-promoting activity. Considering the topographical properties of the RTN, it’s laminar structure, functional non-homogeneity and technical difficulties of making precise damage in such a fine structure as the RTN, we decided to use a pharmacological way to manipulate the activity in the RTN. The Alpha-2 adrenergic agonist clonidine was chosen as a tool to aggravate thalamic activity (as was predicted by Buzsáki and co-authors [1991]): it aggravates the number of SWD I [van Luijtelaar, 1997] without a clear effect on number of type II SWD. Moreover, spectral analysis of thalamic EEG allowed us to report on cross-wise relation of SWD I and SWD II [Sitnikova and van Luijtelaar, 2005].

Oscillatory activity of the RTN can be modified considerably through cortical descending terminals. Recent investigations have questions the existence of a subcortical pacemaker of SWD I, more and more evidence has been put forward which suggest that the neocortex may trigger SWD [Meeren, 2002; Manning, 2004]. We attempted to check this idea and to put our project in line with modern studies, therefore more experiments were added to indicate the role of the cortex in SWD I. With microinjections in the somatosensory cortex, we proved that the SWD I are under control of the neocortex.

In the frame of the initial goal, we have performed detailed analysis of electroencephalographic properties of SWD and examined the possibility that normal sleep spindles could be immediate precursors of SWD. The similarities and dissimilarities between normal and paroxysmal oscillations were thoroughly studied here by means of cross-spectral, coherence and cross-correlation analyses.

Results

a. short summary

WAG/Rij rats have 2 types of SWD and two types of sleep spindles. The two types of sleep spindles can be discriminated based on spatial cortical and thalamic distribution: anterior sleep spindles are thalamically expressed in the RTN and ventral part of the VPM, posterior spindles in the dorsomedial part of the VPM, wlile the role of the RTN in posterior sleep spindles remains unclear. Moreover, the outcomes leave room for a non-thalamic origin of the posterior sleep spindles. Anterior and posterior sleep spindles also differ in amplitude and frequency characteristics.

In contrast to outcomes of previous studies in which large thalamic lesions (including substantial parts of the VPL and VPM) or when the lesions were restricted to the dorsal part of the RTN (Meeren, 2002; van Luijtelaar and Weltink, 2001), no decrease in the number of SWD was found when lesions were restricted to the ventral part of the rostral pole of the RTN.

b. what was in accordance with the hypothesis

A comprehensive EEG analysis was performed to reach the final aim of our study, i.e. to establish the role of the RTN in transitions from sleep spindles to SWD. Spectral, cross-spectral and coherent analyses were carried out on EEG’s derived from frontal, occipital cortical areas, specific thalamic nucleus (the ventroposteromedial complex, VPM) and the reticular thalamic nucleus (RTN). It was shown that anterior sleep spindles widely invade the frontal cortex, the VPM and the RTN. Thalamic expression of posterior sleep spindles is restricted to the dorsomedial part of the VPM, associations between occipital cortex and the RTN were more limited than that in anterior sleep spindles. This is in accordance to a hypothesis that two types of sleep spindles have different thalamic sources (Hypothesis 3). The RTN is involved in both spindle types, but different strategies (pathways) are used during the two spindle types to communicate with the cortex.

With spectral analysis we identified characteristics of paroxysmal pattern of EEG by putting forward two principles: the presence of a sharp spectral peak with mean frequency of discharges (suggesting that oscillations are regular) and high amplitude of first harmonic (suggesting the presence of rectangular component, i.e. spike). According to that principle, SWD type I is regarded as a pure paroxysm, sleep spindles are purely non-paroxysmal transients and SWD II bear some paroxysmal features.

The relationship between SWD I and anterior sleep spindles was more complex than initially expected. Anterior sleep spindles might be transformed to SWD I as a consequence of serious changes in thalamo-cortical machinery that leads to the appearance of the spike-component. Based on spectral analysis, it is concluded that the presence of the spike in the EEG changed the EEG pattern into a paroxysmal one. The spike-component had a high amplitude in the frontal cortex during SWD type I, but the spike was nearly absent in the thalamus. However, the thalamus is involved in the generation of the spike, because the thalamus strongly communicates with the frontal cortex in specific frequency range of 18-22 Hz, which corresponds to the first harmonic of the spike. Genesis of a paroxysmal element (spike) requires strengthening of the thalamo-cortical coherence and re-arrangement of overall pattern of thalamo-cortical associations. We observed that the central thalamo-cortical associations are inversed in transitions from wakefulness to sleep (when signals are in-phase synchronized) via anterior sleep spindles (when in-phase synchronization was nearly absent) towards SWD I (when signals are out-of-phase synchronized).

With the aid of spectral analysis we disproved our hypothesis about a relationship between posterior sleep spindles and SWD type II. Instead, we found evidence that SWD II is a local occipital paroxysm that resembles SWD I. Experiments with the alpha-2 agonist clonidine showed that noradrenergic system modulates both SWD I and SWD II in two different ways: clonidine injections highly enhanced SWD I, that corresponded to aggravatatoin of 9-14 Hz activity in the RTN, but SWD II showed a tendency to be reduced, that correspond to a reduction of 5-9 Hz activity in the occipital cortex, the VPM and the RTN. Considering the fact that the most significant effect of clonidine during SWD I and SWD II was found in the RTN, it was concluded that expression of both types of SWD might reflect a dysfunction of the RTN.