Disruption of the two-state membrane potential of striatal neurones during cortical desynchronisation in anaesthetised rats

Abstract

In anaesthetised animals, the very negative resting membrane potential of striatal spiny neurones (down state) is interrupted periodically by depolarising plateaux (up states) which are probably driven by excitatory input. In the absence of active synaptic input, as occurs in vitro, potassium currents hold the membrane potential of striatal spiny neurones in the down state. Because striatal spiny neurones fire action potentials only during the up state, these plateau depolarisations have been perceived as enabling events that allow information processing through cerebral cortex-basal ganglia circuits. Recent studies have demonstrated that the robust membrane potential fluctuation of spiny neurones is strongly correlated to the slow electroencephalographic rhythms that are typical of slow wave sleep and anaesthesia. To further understand the impact of cortical activity states on striatal function, we studied the membrane potential of striatal neurones during cortical desynchronised states. Simultaneous in vivo recordings of striatal neurones and the electrocorticogram in urethane-anaesthetised rats revealed that rhythmic alternation between up and down states was disrupted during episodes of spontaneous or induced cortical desynchronisation. Instead of showing robust two-state fluctuations, the membrane potential of striatal neurones displayed a persisting depolarised state with fast, low-amplitude modulations. Spiny neurones remained in this persistent up state until the cortex resumed ~1 Hz synchronous activity. Most of the recorded neurones exhibited a low firing probability, irrespective of the cortical activity state. Time series analysis failed to reveal significant correlations between the membrane potential of striatal neurones and the desynchronised electrocorticogram. Our results suggest that during cortical desynchronisation continuous uncorrelated excitatory input sustains the membrane potential of striatal neurones in a persisting depolarised state, but that substantial additional input is necessary to impel the neurones to threshold. Our data support that the prevailing cortical activity state determines the duration of the enabling depolarising events that take place in striatal spiny neurones.

Figure 1. Electrode placement and neuronal labelling

A, diagrams of coronal sections of the rat brain (Paxinos & Watson, 1997) showing the localisation of the ECoG recording electrode (left) and the mesopontine region stimulation electrode (right). B, photomicrograph of one of the recorded neurones, labelled with neurobiotin, displaying the characteristic features of striatal spiny neurones (inset: dendritic spines).

Figure 2. Spontaneous cortical desynchronisation

A, representative segment of signal showing spontaneous shifts between synchronised and desynchronised ECoG patterns in a urethane-anaesthetised rat. Note the strong correlation between the high-amplitude V_m fluctuations of striatal neurones and the slow rhythmic activity of the frontal ECoG. The typical two-state membrane potential of striatal neurones was disrupted during periods of ECoG desynchronisation, in some instances for more than 30 s. B, all-points histograms depicting the profile of V_m values during representative synchronised (above) and desynchronised (below) ECoG epochs. The samples used for these plots were taken from epochs marked with an asterisk in A. C, the graphs depict ECoG and V_m power spectra, cross-amplitude and coherence spectra, and cross-correlograms, during different cortical activity states (left: synchronised ECoG; right: desynchronised ECoG). The spectra have been cut at 5 Hz to facilitate scrutiny of the ≈1 Hz components in the signals, which represent the slow oscillation in the ECoG and transitions between up and down states in striatal neurones. In each graph, each data series represents a single ECoG-V_m pair (n = 7 recording pairs), where each data series resulted from averaging information obtained from four signal epochs. Note the elevated coherence of the seven pairs when the ECoG displays the ≈1 Hz rhythm.

Figure 3. Desynchronisation induced by MPT stimulation

A, disruption of the slow frontal ECoG rhythm by electrical stimulation of the mesopontine tegmentum was accompanied by a change in the shape of the striatal neurone V_m, that shifted from the typical two-state profile to a low-amplitude high-frequency pattern. The arrow points to the stimulus artefact. In this experiment the mesopontine electrode was located in the side contralateral to the recording electrodes. B, all-points histograms depicting the profile of V_m values during representative synchronised (above) and desynchronised (below) ECoG epochs, taken from the signal depicted in A. C, the quantitative changes induced by mesopontine stimulation are evident in the spectra and cross-correlograms displayed below the signals (left: activity under a synchronised ECoG; right: activity under a desynchronised ECoG), which summarise findings from fourteen ECoG-V_m pairs. See Fig. 2 for details. The very low frequency components in the desynchronised ECoG power spectra are related to the stimulus artefact.

Figure 5. Quantitative assessment of the time difference between the recovery of ECoG synchronisation and two-state V_m transitions after MPT stimulation

The arrows indicate the artefact produced by MPT stimulation, which is followed by an episode of ECoG desynchronisation and disruption of two-state transitions in the striatal neurone. The action potentials were truncated at −20 mV. A sigmoid function was adjusted to the variance curves depicted above each signal (see Methods for details). The centre of each sigmoid function is indicated with a dot. In the example depicted, the ECoG reached a steady ≈1 Hz rhythm 6 s before the V_m of the striatal neurone resumed a stable two-state behaviour.

Figure 7. Results from 6-OHDA-lesioned rats and ECoG suppressions

A, mesopontine region stimulation produced ECoG desynchronisation with concomitant disruption of the two-state V_m of a striatal neurone in a 6-hydroxydopamine-lesioned rat. In this experiment the mesopontine region electrode was placed in the same side of the recordings and lesion. B, tyrosine hydroxylase immunohistochemistry revealed a severe destruction of the nigrostriatal projection in 6-hydroxydopamine-lesioned rats. Calibration bar: 1.5 mm. C, the stepping test demonstrated a severe akinesia of the forelimb contralateral (contra) to the 6-hydroxydopamine-induced lesion (* P < 0.05, Wilcoxon signed-rank test). D, a synchronised ECoG displaying short ‘suppressions’ (arrows). Note the correspondence between the duration of suppressions and down states. E, a typical burst-suppression ECoG in a deeply anaesthetised rat. Up states in the striatal neurone occurred concomitantly with outburst of cortical activity.

Figure 4. Membrane potential during up states, down states and persistent depolarisations

Box plots showing the median V_m (circles), 25th and 75th percentiles (bar limits), and range (error bars), during up states, down states and the sustained depolarisations (desync) observed during spontaneous (right; n = 7) or induced (left; n = 14) ECoG desynchronisation. The V_m during down states differed significantly from values of up states and persistent depolarisations (* P < 0.05, Tukey's test after a significant one-way ANOVA for repeated measures (induced desynchronisation) or Dunn's method after a significant Kruskal-Wallis ANOVA by ranks (spontaneous desynchronisation)). There were no significant differences between the corresponding V_m values of spontaneous and induced desynchronisations (Kruskal-Wallis ANOVA by ranks).

Figure 6. Special features of desynchronisation induced by MTP or sensory stimulation

A, episodes of ECoG desynchronisation evoked by electrical stimulation of the MPT (arrows). Increasing stimulation current produced episodes of desynchronisation of increasing length, associated with persistent up states of increasing duration. B, a short duration sensory stimulus (pressure applied to the tail) produced a long lasting episode of ECoG desynchronisation and concomitant disruption of two-state transitions in a striatal neurone (above). In another rat (below), a brief pressure applied to the tail, that failed to desynchronise the ECoG, did not produce noticeable changes in the V_m of the recorded striatal neurone, while a sustained pressure produced a brief episode of ECoG desynchronisation associated to disruption of two-state transitions in the striatal neurone. The vertical lines in the recordings are artefacts introduced by the stimulation device. The action potentials were truncated at −20 mV.

Figure 8. Correlation between the ECoG and V_m at high frequencies

During episodes of ECoG desynchronisation, the V_m of striatal neurones displayed low-amplitude high-frequency modulations that varied in frequency within short time-windows (above). The graphs depict time series analysis of four disjoined 500 ms epochs (thin lines) of a single ECoG-V_m pair (two of the epochs were obtained from the signal segment displayed above). Thick lines are averages of the four epochs. Note that for individual epochs coherence might attain high values (i.e. peaks at 15 and 34 Hz with coherence > 0.8) but the mean coherence did not reach significance (coherence > 0.66; Halliday et al. 1995). The mean cross-correlation displayed a weak modulation at 15 Hz and no evidence of 34 Hz oscillation.