The plot is a running average histogram of the firing frequency (10 s bins) of a LRn-p neurone, before, during and after bath perfusion of DA (30 μm). Raster plots of spike time stamp are shown above, during control (upper) and DA (30 μm) perfusion (lower).

The left panel shows the spike time autocorrelogram of a HR neurone in control conditions. Horizontal dotted lines in the autocorrelogram indicate 99% confidence limits (Abeles, 1982). The plot on the right is a running average histogram of the firing frequency (10 s bins) of the whole experimental session, before, during and after bath perfusion of DA (30 μm). Raster plots of spike time stamp are shown on top, during control (left) and DA (30 μm) perfusion (right).

A, B and C are typical LRp neurones responding with prominent inhibition, weak inhibition or no change to DA (30 μm), respectively. In each row, left panels are spike time autocorrelograms of the corresponding LRp neurone, in control conditions. Horizontal dotted lines in the autocorrelograms indicate 99% confidence limits (Abeles, 1982). The plots on the right are running average histograms of the firing frequency (10 s bins) of the whole experimental session, before, during and after bath perfusion of DA (30 μm). Raster plots of spike time stamp are shown on top, with start of DA (30 μm) perfusion indicated by a vertical arrow.

A, B and C are typical LRp neurones responding with prominent inhibition, weak inhibition or no change to DA (30 μm), respectively. In each row, left panels are spike time autocorrelograms of the corresponding LRn-p neurone, in control conditions. Horizontal dotted lines in the autocorrelograms indicate 99% confidence limits (Abeles, 1982). The plots on the right are running average histograms of the firing frequency (10 s bins) of the whole experimental session, before, during and after bath perfusion of DA (30 μm). Raster plots of spike time stamp are shown on top, with start of DA (30 μm) perfusion indicated by a vertical arrow.

A, typical HR (a), LRp (b) and LRn-p (c) neurones. In each panel, the plots on the left are spike time autocorrelograms, while plots on the right are inter-event interval histograms. Horizontal dotted lines in the autocorrelograms indicate 99% confidence limits (Abeles, 1982). Raster plots of spike time stamp are shown on top of each row. B, histograms of the control firing frequency (a) and autocorrelation probability (b) of the three neuronal populations, showing separate relative distributions and distinct peaks. Panels in c are scatter plots of the basal firing rate vs. autocorrelation probability (left) and vs. coefficient of variation (right), showing fairly distinct clusters, corresponding to the three populations.

Aa and Ba, autocorrelation vs. time plots of spike time stamps recorded by two separate electrodes. Autocorrelation probability is expressed with a colorimetric scale and measured across the whole recording session, at consecutive 10 s intervals. The autocorrelation peaks, measured every 10 s, give rise to continuous horizontal bands over the blue background, reflecting the maintenance of firing regularity. Firing inhibition by DA (30 μm) is expressed by the disappearance of the autocorrelation signal, which slowly recovers at washout. Note the perfect match between the two autocorrelogram vs. time plots, most likely reflecting action potentials of the same neurone, recorded by the two separate electrodes. Spike waveforms detected by the two electrodes are shown in Ab and Bb (black line indicates upper and lower spike template limits). Note the similarity of the two spike waveforms in Ab, as opposed to the different waveforms in Bb. Ac and Bc, crosscorrelation probability expressed with a colorimetric scale and measured across the whole recording session, at consecutive 30 s intervals. Firing synchronicity is shown by the horizontal band around 0 ms.

A, cumulative plots of the change in firing frequency mediated by 30 μm DA, expressed as percentage of control. The thick line is relative to all cells, thin line to LRp cells, dashed line to LRn-p cells and dotted line to HR cells. The effect of DA on the firing rate was separated into prominent inhibition (0–59%; striped box) and weak inhibition (60–89%; grey box), judged by the presence of a shoulder in the rising slope of the cumulative curves, indicated by the black arrow. Additionally, DA effect was defined as no change (90–110%; white box) and excitation (>110%; black box). B, pie plots of the proportion of cells responding with prominent inhibition, weak inhibition, no change or excitation, in all the neurones and in the HR, LRp and LRn-p subpopulations. C, cumulative histograms (mean ±s.e.m.) of the basal firing rate (left) and autocorrelation probability (right) of the HR (top row), LRp (middle row) and LRn-p (bottom row) neurones, further separated according to their response to 30 μm DA. The number of cells is indicated in parentheses in each column. Significant differences are shown above the corresponding histograms (*P < 0.05; **P < 0.01; ***P < 0.001; t test for unpaired data).

Aa, photograph of a midbrain slice placed over an array of planar multi-electrodes, detectable in transparency through the slice. The schematic drawing below highlights electrode position in relation to substantia nigra pars compacta (SNc), the substantia nigra pars reticulata (SNr) and the medial terminal nucleus of the accessory optic tract (MT). The dotted arrow is placed medially to the SNc and points towards the frontal direction, along the fronto-caudal axis. Scale bar, 300 μm. Ab, electrical activity detected from each of the 64 electrodes from the same slice shown in a. B, spikes acquired from a single channel were sorted by principal component analysis. Amplitudes of the first three components (PC1, PC2 and PC3) are then displayed in a three-dimensional plot, giving rise to spatially separated clusters, indicative of separate spike waveforms. The four clusters of the three-dimensional plot were obtained with the spike waveforms indicated by the corresponding letters (black line indicates upper and lower spike template limits).

A, autocorrelation vs. time plots of spike time stamps recorded from the same SNc slice with separate electrodes of neurone 1, neurone 2 and neurone 3. The different profiles of autocorrelation probability and sensitivity to 30 μm DA clearly show that they are different neurones. Crosscorrelation probability plots obtained from the corresponding pairs indicated the occurrence of synchronised firing in 1 vs. 2 and in 1 vs. 3, but not in 2 vs. 3. Crosscorrelograms were measured from the whole recording time (1 ms bin size). Horizontal dotted lines in the crosscorrelograms indicate 99% confidence limits (Abeles, 1982). B, mapping of the neuronal distribution in the same horizontal slice, in relation to their corresponding position in the 64 channels MEA grid. The grey area indicates MT position, while the two black dotted lines define the borders of the crescent-like SNc area. Neurones are colour-coded according to their classification into HR, LRp and LRn-p neurones (left panel) or sensitivity to DA (30 μm), as prominent inhibition, weak inhibition, no change or excitation (right panel). Black double-arrowed lines are used to link neurones functionally connected by a significant proportion of crosscorrelated spikes.

Aa, c and e, autocorrelation vs. time plots of spike time stamps recorded from the same SNc slice, with separate electrodes. Autocorrelation probability is expressed with a colorimetric scale and measured across the whole recording session, at consecutive 10 s intervals. The autocorrelation peaks, measured every 10 s, give rise to continuous horizontal bands over the blue background, reflecting the maintenance of firing regularity. DA (30 μm) effect on neuronal firing is expressed in Aa, c and e (right panel) by the shift or disappearance of the autocorrelation signal, which slowly recovers at washout, while no effect of DA was evident in Ae (left panel). The panels in Ab, d and f are crosscorrelation probability plots obtained from the corresponding pairs on the left, measured from the whole recording time (1 ms bin size). Horizontal dotted lines in the crosscorrelograms indicate 99% confidence limits (Abeles, 1982). Ba, b and c, crosscorrelation probability expressed with a colorimetric scale and measured across the whole recording session, at consecutive 30 s intervals, refers to the pairs of neurones shown in Aa, c and e, respectively. Firing synchronicity is shown by the horizontal band around 0 ms. Note the increase in crosscorrelation probability associated with DA (30 μm) during drug perfusion (a and c) or when spikes recovered at drug washout (b). Bd, superimposed histogram (mean ±s.e.m.) and scatter plot of the crosscorrelation probability measured from all pairs of neurones (n= 53) during 1 min recording in control conditions and in DA, or when firing inhibition started to recover. The two populations were significantly different (***P < 0.001; t test for paired data).