Synaptic Cooperativity Regulates Persistent Network Activity in Neocortex

Effect of an AMPA-receptor antagonist (GYKI) on up-states evoked by electrical stimulation of thalamocortical and intracortical pathways. A, The intensity was set so that, during control, electrical stimulation of the thalamus evokes up-states in somatosensory cortex (left), while stimulation within the cortex does not (right). Application of GYKI abolishes up-states evoked by thalamocortical stimulation but at the same time unmasks up-states evoked by intracortical stimulation. This effect is reflected in both intracellular and FP recordings. B, Population data of the effects of GYKI or d-AP5 on thalamocortical-evoked short-latency (2–8 ms) and long-latency (15–50 ms) FP responses. The long-latency response reflects the up-state. C, Population data of the effects GYKI or d-AP5 on intracortical-evoked short-latency (2–8 ms) and long-latency (15–50 ms) responses. *p < 0.01.

Histological verification of recorded cells and ChR2 expression. A, Identification of some of the cells studied. In some cases, during histological processing, the apical dendrite was truncated and lost by the resectioning (80 μm) of the 400 μm slices. B, Typical thalamocortical slice obtained from a Thy1-ChR2-eYFP-line18 mouse (1 resectioned 80 μm section is shown). EYFP, green; DAPI, blue. Note the robust eYFP fluorescence (ChR2 expression) of layer V cells and their apical dendrites, and the typical sparseness of expression in layer IV, indicating lack of ChR2 expression in thalamocortical fibers. C, Three successive slices from a CD-1 mouse cut in the thalamocortical plane (1 resectioned 80 μm section is shown) that had been injected with AAV-hSyn-ChR2-eYFP in somatosensory thalamus. Note the robust ChR2 expression of thalamocortical fibers in layer IV, and somewhat in layer VI, of somatosensory cortex. The inset strip in the middle slice shows a neurobiotin-filled cell (red) recorded in that slice overlapped with the ChR2 expression (green); this is cell9 reconstructed in A. EYFP, green; DAPI, blue; DyLight594, red. wm, white matter.

Activation of intracortical synapses that express ChR2 in slices from Thy1-ChR2-line18 mice, using different intensities of blue light, produces effects similar to intracortical electrical stimulation; up-states are evoked at low intensities but suppressed at high intensities. A, B, Examples from two different cells showing the effects of three different blue-light intensities on short-latency responses and up-states. Note that low-intensity blue light evokes up-states reliably, while increasing the intensity of the blue light leads to larger short-latency responses and abolishment of evoked up-states. The cell in A is cell13 reconstructed in Figure 3A. C, Effect of blue-light intensity on FP responses in somatosensory cortex; up-states are only evoked at low intensities. D, Population data showing the effect of intracortical blue-light intensity on short-latency (2–8 ms) and long-latency (15–50 ms) FP responses. The long-latency response reflects the up-state. *p < 0.01 versus low intensity.

Effect of GYKI on intracortical responses evoked by high-intensity blue-light stimuli in Thy1-ChR2-line18 slices. A, Typical example from a cell recorded at different V_m levels set by intracellular current pulses (500 ms; blue light occurs 100 ms after current pulse onset). The high-intensity blue-light stimulus evokes a sharp short-latency response curtailed by a robust IPSP that can lead to a rebound excitation. GYKI strongly suppresses the short-latency response (but does not abolish it; see below), but at the same time it unmasks an up-state during the period where previously the IPSP was evoked. B, Subsequent application of d-AP5 completely abolishes the unmasked up-state, leaving intact a short-latency response caused by direct activation of ChR2 channels in this cortical cell; it survived block of glutamate receptors and Na⁺ channels. The blue-light duration is clearly reflected in the depolarization it causes, indicating that the cell expresses ChR2. In A (top), the number of spikes evoked by the depolarizing current pulses (> 0 nA) are calculated (100 ms bin). Note the unmasking of spikes during GYKI. The cell is cell10 reconstructed in Figure 3A.

Activation of thalamocortical synapses that express ChR2 in slices from mice injected with AAV-hSyn-ChR2-eYFP in somatosensory thalamus, using different intensities of blue light, produces effects similar to intracortical electrical or blue-light stimulation but different from electrical stimulation of the thalamus; up-states are evoked at low intensities and suppressed at high intensities. A, Example of thalamocortical ChR2 expression (eYFP) in the somatosensory (note the barrels in layer IV) cortex of a typical thalamocortical slice from a mouse injected with AAV-hSyn-ChR2-eYFP in the somatosensory thalamus. EYFP, green; DAPI, blue. B, Different effects of application of blue light into the thalamus or into the somatosensory cortex on FP responses recorded in cortex. Blue light in thalamus results in small short-latency responses that evoke up-states (like thalamocortical electrical stimulation). Application of blue light in cortex of the same slice, which recruits only thalamocortical synapses regardless of whether their axons reach the thalamus in the slice, evoked up-states only at low intensities. However, increasing the intensity recruits many more thalamocortical synapses, which lead to sharper short-latency responses that abolish evoked up-states. C, D, Examples from two different cells showing the effects of three different blue-light intensities on short-latency responses and up-states. Note that low-intensity blue light evokes up-states reliably, while increasing the intensity of the blue light leads to larger short-latency responses and abolishment of evoked up-states. The cell in D is cell9 reconstructed in Figure 3A. E, Effect of blue-light intensity on FP responses in somatosensory cortex; up-states are only evoked at low intensities. F, Population data showing the effect of intracortical blue-light intensity on short-latency (2–8 ms) and long-latency (15–50 ms) FP responses. The long-latency response reflects the up-state. *p < 0.01 versus low intensity.

If synaptic cooperativity is increased, GYKI also unmasks up-states in thalamocortical pathways. A, B, Effect of GYKI on thalamocortical responses evoked by electrical stimulation of the thalamus (A) and by high-intensity blue-light stimuli in slices from mice injected with AAV-hSyn-ChR2-eYFP in somatosensory thalamus (B). Responses in A and B are from the same cell, in which electrical and blue-light stimulation alternated. Evoked responses are recorded at different V_m levels set by intracellular current pulses. High-intensity electrical stimulation of the thalamus evokes an up-state during control, while high-intensity blue-light stimulus evokes a sharp short-latency response but no up-state. GYKI abolished the up-state evoked by electrical stimulation of the thalamus and the short-latency response evoked by blue-light pulses. At the same time, GYKI unmasks an up-state evoked by blue light. Subsequent application of d-AP5 completely abolishes the unmasked up-state (gray trace). C, The same effects are observed in FP recordings evoked by high-intensity blue-light stimuli. D, E, Population data of the effects of GYKI and subsequent application of d-AP5 on thalamocortical-evoked short-latency (2–8 ms) and long-latency (15–50 ms) FP responses. The long-latency response reflects the up-state. *p < 0.01 GYKI versus control.

GYKI unmasks an NMDA-mediated slow excitation and relieves shunting inhibition evoked by intracortical electrical stimulation. A, In slices that evoke up-states, it is difficult to study what GYKI unmasks to trigger up-states because the response is overwhelmed by the unmasked up-state. B, To overcome this problem, we used slices that do not produce spontaneous or evoked up-states but that otherwise appear to have normal short-latency responses. Under these conditions, GYKI unmasked a slow NMDA-mediated excitation not apparent during control (see arrows); it was abolished by d-AP5 (Fig. 9). C, Population data of the effect of GYKI on spikes evoked by positive current pulses (> 0 nA) before (−100–0 ms bin) and after the intracortical stimulus (0–400 ms bins) in slices that do not produce up-states. During control, the IPSP evoked by the intracortical stimulus shunts the spikes evoked by the current pulse for the period after the intracortical stimulus, but not before the stimulus. GYKI relieved this effect; the number of spikes increased during the period after the stimulus but not before. *p < 0.01 GYKI versus control per 100 ms bin.