Burst spiking of a single cortical neuron modifies global brain state

Abstract

Different global patterns of brain activity are associated with distinct arousal and behavioral states of an animal, but how the brain rapidly switches between different states remains unclear. We here report that repetitive high-frequency burst spiking of a single rat cortical neuron could trigger a switch between the cortical states resembling slow-wave and rapid-eye-movement sleep. This is reflected in the switching of the membrane potential of the stimulated neuron from slow UP/DOWN oscillations to a persistent-UP state or vice versa, with concurrent changes in the temporal pattern of cortical local field potential (LFP) recorded several millimeters away. These results point to the power of single cortical neurons in modulating the behavioral state of an animal.

Fig. 1

UP/DOWN and persistent-UP states observed in simultaneous whole-cell and LFP recordings in rat cortex. (A) Left, schematic illustration of recording configuration. Whole-cell and LFP electrodes were separated by 0.3–6 mm. Right, a whole-cell recorded pyramidal neuron in somatosensory cortex. (B) UP/DOWN (left) and persistent-UP (right) states observed in a visual cortical neuron. (C) LFP recorded 2 mm from the patch recording in (B). (D) Distributions of membrane potentials (E_m) during UP/DOWN (left) and persistent-UP (right) states. Data are from 3 cells (marked by different colors). (E) LFP power spectra from the same experiments as in (D), each normalized by the mean power at 0.5–1.5 Hz during UP/DOWN state. Insets: power spectra on log-log scales. Gray shadings, low and high frequency ranges for computing L/H power ratio.

Fig. 2

Switch from UP/DOWN to persistent-UP state induced by single-neuron burst spiking. (A) State switch indicated by change from bimodal to unimodal E_m distribution (color coded, computed in 20-s windows). Blue bar: burst spiking period. Insets above, sample whole-cell recording traces during periods marked by arrows. (B) State switch indicated by change in LFP power spectrum (color coded, log scale, computed in 20-s windows) recorded ~1 mm from whole-cell electrode. Insets, sample LFP traces. (C) L/H power ratio (between 0.5–4 Hz and 20–60 Hz) for LFP in (B). Dashed red lines: averaged values 3 min before and 3 min after burst spiking; distance between these lines was used to measure burst-induced L/H ratio change. (D–F) Similar to (A–C), from another experiment; LFP was recorded ~5 mm from patch electrode.

Fig. 3

Switch from persistent-UP to UP/DOWN state induced by single-neuron spiking. (A) State switch indicated by change from unimodal to bimodal E_m distribution. (B) State switch indicated by change in LFP power spectrum (recorded ~5 mm from whole-cell electrode). (C) L/H power ratio, for LFP in (B). (D–F) Similar to (A–C), from another experiment; LFP was recorded ~1 mm from patch electrode.

Fig. 4

Time course and frequency dependence of the switch of cortical state. (A) Average LFP L/H power ratio for 15/38 experiments with significant ratio changes induced by spiking at 50–100 Hz (solid symbols in C). Upper, average of 9 experiments showing significant UP/DOWN to persistent-UP transitions; lower plot, average of 6 experiments showing persistent-UP to UP/DOWN transitions. L/H ratio of each experiment was normalized by its mean before spiking. Error bar, ±SEM. (B) Distributions of time for spiking-induced switch (upper) and spontaneous reversal (lower), for the same 15 experiments as in (A). Gray stripe, burst spiking period. (C) |Z-score| for burst-induced LFP L/H ratio change vs. burst frequency. Each point represents one experiment performed on one whole-cell recorded neuron in visual (circle) or somatosensory (square) cortex. Solid symbols, experiments with significant ratio changes (p < 0.05); “0 Hz”, control experiments with no spiking during the 3-min sham period. Compared to “0 Hz”, Z-score distribution was significantly different for 25 Hz (p = 0.010, two-sample Kolmogorov-Smirnov test) and 50–100 Hz (p = 0.0014) but not for 5 Hz (p = 0.20). Probability of burst-induced transition (percentage of solid symbols) is significantly higher for 25 Hz (46%, p < 0.001, bootstrap (15)) and 50–100 Hz (39%, p < 0.001), but not for 5 Hz (16%, p = 0.11).