The thalamo-cortical complex network correlates of chronic pain

Figures

• 

  Figure 1. Explanation of the experimental setup. (A) The illustration of the stereotactile coordinate in the Paxinos’s atlas (reprint permission n. 1982084; Paxinos, G., Watson, C. The Rat Brain in Stereotaxic Coordinates (Third Edition), 172 (Elsevier, 2006)). The two black lines represent the electrode matrices illustrated in (B). (C) A typical time series of the raw extracellular recordings where the functional connections were extracted by means of the functions in (D). (E) A sample of a binary matrix extracted by recordings that corresponds to the graph in (F).

• 

  Figure 2. Basilar features of the spiking activity. (A) The plots show the distribution of firing rate (spike/s) for the recorded neurons of VPL thalamus (left) and for the S1 cortex (right). While in VPL there is a statistical significance between the normal and neuropathic animals, in S1 there is only significance among the neuropathic classes. (B) The cross-correlation estimated between the neurons from VPL (left), S1 (center) and VPL versus S1 (right). Cross-correlation is unable to discriminate control from neuropathic rats. CR represents the control, PI the peripheral injury, the SC the sciatic constriction (Bennet-Xie model) and the SL (Seltzer model) animals. Distributions are visualized as boxplot according to the standard parameters.

• 

  Figure 3. Comparison among the experimental conditions. (A) Average graph weights (obtained from LFP) did not statistically differ from normal to neuropathic models. Similarly, graph weights from spiking activity produce the same result. CR represents the control, PI the peripheral injury, the SC the sciatic constriction (Bennet-Xie model) and the SL (Seltzer model) animals. Lines outgoing the central balls indicate the standard deviation.

• 

  Figure 4. Node degree distribution along the experimental conditions and recorded signal types. The node degree histograms of the graphs obtained from the LFP (above) and spiking activity (below). CR represents the control, PI the peripheral injury, the SC the sciatic constriction (Bennet-Xie model) and the SL (Seltzer model) animals.

• 

  Figure 5. Analysis of betweenness centrality (BC). In the LFP (left), the BC is mainly distributed in S1 in normal animals while in SL the balancing is reversed, in SC appears equidistributed an in PI is barely maintained. A similar trend is shown also in the spiking activity (right). CR represents the control, PI the peripheral injury, the SC the sciatic constriction (Bennet-Xie model) and the SL (Seltzer model) animals.

• 

  Figure 6. A comparison between the power spectral distributions of the four EEG electrodes recorded from normal rats (left side) and CP rats (right side). Frequencies are divided into four EEG bands: Delta (1–4 Hz), Theta (4–8 Hz), Alpha (8–12 Hz), Beta (13–25 Hz). The statistical analysis did not reveal any significant difference between the whole two distributions (P = 0.320, two-sample Kolmogorov-Smirnov test) and, by compare pair-wisely, each band (P > 0.1 in all ranksum tests).

• 

  Figure 7. Samples of the spiking activity functional graphs for each experimental condition. First column reports the raw weighted adjacency matrices while the second column indicates the respective binarized versions. The third column represents the graphs as plotted by the “CommunityGraphPlot” function (Mathematica, http://www.wolfram.com). Remarkable differences are appreciable both among CP conditions and between CR and CP condition. CR represents the control, PI the peripheral injury, the SC the sciatic constriction (Bennet-Xie model) and the SL (Seltzer model) animals.