Laminar-dependent effects of cortical state on auditory cortical spontaneous activity

Abstract

Cortical circuits spontaneously generate coordinated activity even in the absence of external inputs. The character of this activity depends on cortical state. We investigated how state affects the organization of spontaneous activity across layers of rat auditory cortex in vivo, using juxtacellular recording of morphologically identified neurons and large-scale electrophysiological recordings. Superficial pyramidal cells (PCs) and putative fast-spiking interneurons (FSs) were consistently suppressed during cortical desynchronization. PCs in deep layers showed heterogeneous responses to desynchronization, with some cells showing increased rates, typically large tufted PCs of high baseline firing rate, but not FSs. Consistent results were found between desynchronization occurring spontaneously in unanesthetized animals, and desynchronization evoked by electrical stimulation of the pedunculopontine tegmental (PPT) nucleus under urethane anesthesia. We hypothesize that reduction in superficial layer firing may enhance the brain's extraction of behaviorally relevant signals from noisy brain activity.

Keywords: cell-type; cortical circuit; ensemble recording; sensory cortex; slow oscillation.

Figure 1

State-dependent spontaneous activity across cortical layers under anesthesia. (A) Spontaneous shift of cortical states. The spectrogram (top) was computed from local-field potentials (LFPs) in the auditory cortex without auditory stimuli. Log-power was color-coded. The bottom panel shows a history of total LFP power at 0–7 Hz. (B) Multiunit activity (MUA) across cortical layers and LFPs during the desynchronized and synchronized state, with a schematic drawing of a 32 site linear electrode. LFP was taken from channel 16, likely located in infragranular layers. (C) Modulation index of spontaneous population activity in putative L2/3 (0–500 μm from cortical surface) and L5 (800–1100 μm). ^**p < 0.005 (t-test).

Figure 2

State-dependent and cell-type-specific spontaneous firing. (A) An example of pedunculopontine tegmental (PPT) nucleus stimulation and neural population activity in the auditory cortex. LFP, multiple single-units, spectrogram, and a history of low-frequency LFP power are shown. (B) Examples of morphologically identified pyramidal cells (PCs) and the effects of PPT stimulation on their spiking. Rasters show response of these cells to repeated PPT stimuli; lower graphs show peri-stimulus time histograms. (C) Modulation index of morphologically identified PCs and extracellularly classified single-units. L5slP, L5 slender PCs; L5tP, L5 thick PCs; sP and dP putative superficial and deep PCs, respectively; sFS and dFS, putative superficial and deep fast-spiking interneurons (FSs), respectively. ^*p < 0.05, ^**p < 0.01, and ^***p < 0.005 (signed-rank test). (D) Mean firing rate during the synchronized state plotted against that during the desynchronized state (left column) and modulation index (right column) for morphologically identified and putative PCs (top row), and putative FSs (bottom row). Solid lines in left panels correspond to equality; those in right panels are linear regression fits. Please note that linear fitting was done with a semi-logarithmic scale.

Figure 3

State-dependent burst firing. (A) An example of the distribution of inter-spike intervals (ISIs) for a morphologically identified L5tPC (see also Figure 2B) during the synchronized (blue) and desynchronized states (red). Inset: auto-correlogram. (B) Fraction of burst-like firing (≤20 ms ISIs) across cell-types during the synchronized (blue) and desynchronized states (red). ^**p < 0.005 and ^***p < 0.0005 (Kruskal–Wallis test with post-hoc rank sum test).

Figure 4

State-dependent and cell-type-specific spontaneous firing in unanesthetized animals. (A) Spontaneous shift of cortical states. The spectrogram (top) was computed from LFPs in the auditory cortex without auditory stimulation. The bottom panel indicates a history of total LFP power at 0–7 Hz. (B) MUA across layers and LFPs in the desynchronized (left) and synchronized states (right). (C) Modulation index of spontaneous population activity in putative L2/3 (0–500 μm from cortical surface) and L5 (800–1100 μm). ^*p < 0.05 (t-test). (D) Modulation index of extracellularly classified single-unit activity (SUA) in putative L2/3 and L5. ^*p < 0.05 and ^**p < 0.005 (signed-rank test). (E) Firing rate during the synchronized state plotted against that during the desynchronized state (left) and modulation index (right) for putative PCs and putative FSs. Solid line in left panel corresponds to equality; that in right panel is linear regression fit for putative PCs. Dotted line in right panel is linear fit for putative FSs. Linear fitting was done with a semi-logarithmic scale.