Spatiotemporal patterns of an evoked network oscillation in neocortical slices: coupled local oscillators

Abstract

We have discovered an evoked network oscillation in rat neocortical slices and have examined its spatiotemporal patterns with voltage-sensitive dye imaging. The slices (visual and auditory cortices) were prepared in a medium of low calcium, high magnesium and with sodium replaced by choline to reduce the excito-toxicity and sodium loading. After slicing, the choline was washed out while normal calcium, magnesium, and sodium concentrations were restored. The oscillation was evoked by a single electrical shock to slices bathed in normal artificial cerebral spinal fluid (ACSF). The oscillation was organized as an all-or-none epoch containing 4-13 cycles at a central frequency approximately 25 Hz. The activity can be reversibly blocked by 6-cyano-7-nitroquinoxalene-2,3-dione (CNQX). 2-amino-5-phosphonopentanoic acid (APV), and atropine but not by bicuculline, indicating polysynaptic excitatory mechanisms. Voltage-sensitive dye imaging showed high-amplitude oscillation signals in superficial and middle cortical layers. Spatiotemporally, the oscillations were organized as waves, propagating horizontally along cortical laminar. Each oscillation cycle was associated with one wave propagating in space. The waveforms were often different at different locations (e.g., extra cycles), suggesting the co-existence of multiple local oscillators. For different cycles, the waves often initiated at different locations, suggesting that local oscillators are competing to initiate each oscillation cycle. Overall our results suggest that this cortical network oscillation is organized at two levels: locally, oscillating neurons are tightly coupled to form local oscillators, and globally the coupling between local oscillators is weak, allowing abrupt spatial phase lags and propagating waves with multiple initiation sites.

Figure 1

A. The stimulation-recording arrangement. The stimulation and recording electrodes were placed in temporal areas II–III, 1 to 2 mm apart. B. An example of the oscillations recorded from a local field potential electrode. C. The distribution of the number of oscillation cycles in each epoch (n= 240 epochs recorded from 14 slices, not counting the first spike). D. The power spectrum (FFT) of oscillations recorded from two animals (red and blue, 10 epochs each). The short red bar and blue bar shows the average peak frequency of each animal respectively (red, 25 Hz; blue, 26.6 Hz). Oc, occipital cortex; RF, rhinal fissure; PRh, perirhinal cortex; Ent, entorhinal cortex; Hip, hippocampus.

Figure 2

Electrical and optical signals of the oscillation. Traces 1 and 2, simultaneous recording of local field potential (LFP) and optical signal from layer III, about 1 mm lateral to the stimulation site. Trace 3, optical signal at 510 nm of light from the same detector during another trial. Trace 4, optical signal from the same detector after the tissue was bathed with bicuculline. The optical signals were from one photo-detector which received light from an area of 83 × 83 μm (20x objective).

Figure 3

Oscillation amplitude in different cortical layers. Voltage-sensitive dye signals (right traces) from 6 optical detectors imaging different cortical layers (left). The oscillation amplitude was high in Layer II-IV but low in deep layers. In contrast the first spike was seen in all the layers.

Figure 4

Horizontal propagation of different oscillation cycles. Left: Optical detectors, A and B, are placed in layer II–III about 1.2 mm lateral to the stimulating electrode (Stim). Right: Optical signals from the two detectors are shown as solid (A) and broken lines (B) respectively, during two trials of evoked oscillations. Arrows on each cycle mark the horizontal propagation direction of the cycle. The direction was determined by the onset time for the waves arriving at the two detectors. The propagating direction of the first spike (f) was same for both trials, coming from the stimulation site and propagating to the lateral. In trial 1, three cycles following the first spike (1–3) had the same propagating direction, medial to lateral. At the end of trial 1 there may be another cycle (s) with direction indeterminate due to small amplitude. In trial 2 (evoked 2 seconds later) two cycles (1–2) had the medial to lateral direction but the other two (r1 and r2) had reversed propagating direction. One cycle, S, appeared to occur simultaneously on both detectors. Signals were filtered between 3–80 Hz.

Figure 5

Initiation foci of the waves. Left: The orientation of the slice and stimulation site is illustrated on the top. The signal from one optical detector, d, is shown on the bottom. The images of three oscillation cycles (1,3,7) are shown on the right panel. Right: Pseudo-color images generated from the signals of all optical detectors, with signal amplitude colored according to a linear pseudo-color scale (top right, peak to red and valley to blue). All images are snap shot of 0.625 ms and the interval between images are 6.25 ms. Each row of the images started at the time marked by a vertical broken line in the trace on the left panel. The initiation site was determined as the detector with earliest onset time and marked as black crosses in the first images of each row. The first spike (1) initiated at the location of stimulating electrode and propagated as a wave to the lateral (top row). Two of the following oscillation cycles (3 and 7) initiated from two different locations. All three initiation foci are marked at the first image of the bottom row for comparison. This optical recording trial contains 1100 images (~0. 7 sec) and only a few of them are shown for clarity. Movies with this type of images are included in the supplemental material.

Figure 6

Distribution of the initiation foci. A. Spatial distribution of the initiation foci of 352 oscillation cycles (first spike excluded) in one representative slice. The location of each focus is labeled by a diamond shaped mark. The size of the diamond mark represents the repeating occurrence at the each location (numbers at the right). The first spike always started at stimulation site (the round dot). All the subsequent oscillation cycles were not initiated near the stimulation site. The majority of the initiation foci are located in temporal cortex III area and a large portion of them was clustered at a few locations (larger diamonds). B. Initiation foci of four example trials from the slice shown in A. Numbers (1–7) indicate the sequence of the cycles following the first spike. While in all trials the first spike initiated at the same location (F, the stimulation site), the subsequent oscillation cycles were apparently initiated from randomly distributed locations.

Figure 7

Waveform variations over space. A. The voltage sensitive dye signals at four locations in a coronal slice (1–4, left) are shown (1–4, right). The first spike (broken line F) is seen at all locations. Extra oscillation cycles are seen at locations 2 and 4 (arrowheads). Signals are filtered between 3 – 200 Hz. Vertical scale: 2×10⁻⁴ of resting light intensity. B. An example of oscillations in tangential slices. Left, The slice was sectioned parallel to the cortical surface, 500 μm thick including part of layer II–III and IV (location of the slice shown on the insert). Oscillations were seen in an area marked by the dashed line on about ~160 detectors. Traces from five detectors (1–5) are shown on the right. Right, At different locations (1 – 4) there are different number of cycles. Outside the boundary of the dashed line there was only the first spike but no subsequent oscillation cycles (trace 5 on the right). Trace AVG, an average of all the ~160 detectors show largely reduced oscillation amplitude, suggesting the whole population was not synchronized. Vertical scale: 2×10⁻⁴ of resting light intensity. Signals are filtered between 1 – 200 Hz.

Figure 8

Local field potential recordings of the oscillations under pharmacological manipulations. Blockade of NMDA receptors (A, B) and AMPA receptors (C) reversibly block the oscillations. Blocking GABA_A receptors (D) did not abolish the oscillations. Manipulation of muscarinic/nicotinic receptors with atropine (E) or carbachol (F) reversibly block the oscillations. All drugs were added to the perfusing ACSF. The perfusion rate was ~ 20 ml/min (28°C). n is the number of animals, for each animal 2 to 3 slices were tested.