Fig 1.
Study design and main experimental parameters.
A: Position of the recording and reference electrodes in the rat skull modified from Paxinos and Watson. B: Time course of the experiments showing the recording sessions with respect to LPS administration. C: Resting state connectivity showing the frequency peaks defined by the debiased weighted phase lag index (dwPLI). Shown are the dwPLI pairs between all channels averaged over all baseline recordings at each frequency. The connectivity bands identified are denoted with top bars. D: Sickness behavior at baseline, 3, 6, 24 and 48 h after LPS administration. Kruskal-Wallis multiple comparison test with respect to baseline: * p = 0.02; ** p = 0.0011; **** p < 0.0001. E-F: Cytokine expression in the cerebral cortex and hippocampal formation of rats in saline treated or 24 and 48 h after LPS administration. Rats receiving saline injections (n = 6) were not used in recording experiments. LPS resulted in a strong response for interleukin-1 beta (IL-1β) (E) and a modest change in interferon gamma (IFN-γ) (F). * p = 0.05; **** p < 0.0001.
Fig 2.
LPS: Spectral power and connectivity bands.
Resting state power profile viewed across the full analyzed spectrum (1 to 200 Hz) (A) and constrained to 40 Hz (B) allowing better visualization of the lower frequencies. C: Power was averaged over all electrodes for each condition to yield a global measure of spectral power. Delta and theta power were increased 6 h after LPS administration (squares), while gamma power and high frequency oscillations (HFO) were reduced at this time point. Power in all affected frequency bands returned to baseline 24 h following LPS. Symbols (*) and lines indicate the 95% HDI contrast between any two conditions obtained from the posterior of the hierarchical Bayesian analysis excluded 0 (n = 11). D-F: Graphical visualization of the effect of LPS on resting state connectivity using the debiased weighted phase lag index (dwPLI). Each trace represents an electrode pair. Note the reductions in the alpha and beta-gamma connectivity bands 6 h after LPS and in the beta-gamma and HFO bands 24 h after LPS.
Fig 3.
Resting state connectivity reductions elicited by LPS. The debiased weighted phase lag index (dwPLI) was used as measure of connectivity between channels. Connectivity topography graphs for each of the 3 frequency bands analyzed (alpha, 5–15 Hz; beta-gamma 20–80 Hz and HFO 150–200 Hz) denoting the connection 'strength' between electrodes separated by intra- (A) and inter- (C) hemispheric connections. Baseline connectivity is denoted in red, with the strength of each connection represented by line width. Changes in connectivity between 6 or 24 h (contrast) with baseline are represented with red (increases) and blue (decreases) with the width of the line denoting the magnitude of change. Note that baseline effects are expected to be positive due to properties of the dwPLI. For analysis, connectivity was averaged over all channel pairs either within (B) or between (D) hemispheres to yield an average measure of inter- vs intra-hemispheric connectivity. No hemisphere by time interactions were detected indicating similar effects of LPS across hemispheres. Data and analyses are presented for both hemispheres for completeness. LPS reduced intra- and inter-hemispheric alpha connectivity at 6 h and returned to baseline at 24 h. Beta-gamma connectivity was reduced at 6 h (intra- and inter-hemispheric) and remained reduced 24 h (inter-hemispheric) after LPS. By contrast, HFO connectivity (intra- and inter-hemispheric) were reduced at 24 h after LPS administration.
Fig 4.
LPS increases phase-amplitude coupling.
Resting state phase-amplitude coupling increased following LPS. The modulation index (MI) was used as a measure of phase-amplitude coupling between theta (3–8 Hz) phase and gamma power (30–100 Hz), and theta phase and HFO (150–200 Hz) power. The MI was averaged over all electrodes for each condition to yield a global measure of the MI. The theta-gamma MI was increased 6 h after LPS administration (squares). Interestingly, the theta-HFO MI was increased only at the 24 h time point relative to both baseline and 6h post-LPS. Symbols (*) and lines indicate the 95% HDI contrast between any two conditions obtained from the posterior of the hierarchical Bayesian analysis excluded 0 (n = 11).