Stability of thalamocortical synaptic transmission across awake brain states

Abstract

Sensory cortical neurons are highly sensitive to brain state, with many neurons showing changes in spatial and/or temporal response properties and some neurons becoming virtually unresponsive when subjects are not alert. Although some of these changes are undoubtedly attributable to state-related filtering at the thalamic level, another likely source of such effects is the thalamocortical (TC) synapse, where activation of nicotinic receptors on TC terminals have been shown to enhance synaptic transmission in vitro. However, monosynaptic TC synaptic transmission has not been directly examined during different states of alertness. Here, in awake rabbits that shifted between alert and non-alert EEG states, we examined the monosynaptic TC responses and short-term synaptic dynamics generated by spontaneous impulses of single visual and somatosensory TC neurons. We did this using spike-triggered current source-density analysis, an approach that enables assessment of monosynaptic extracellular currents generated in different cortical layers by impulses of single TC afferents. Spontaneous firing rates of TC neurons were higher, and burst rates were much lower in the alert state. However, we found no state-related changes in the amplitude of monosynaptic TC responses when TC spikes with similar preceding interspike interval were compared. Moreover, the relationship between the preceding interspike interval of the TC spike and postsynaptic response amplitude was not influenced by state. These data indicate that TC synaptic transmission and dynamics are highly conserved across different states of alertness and that observed state-related changes in receptive field properties that occur at the cortical level result from other mechanisms.

Figure 1.

Spontaneous firing rates and burst rates of single thalamic neurons change dramatically during alert and non-alert states. A, A transition from alert to non-alert states as defined by hippocampal EEG. Five seconds of theta activity are shown preceding the transition, followed by 5 s of HVIR. Vertical lines indicate action potentials from a simultaneously recorded LGN cell, with bursts indicated by asterisks. B, For all of our cells, the average of the power spectral density functions of hippocampal EEG during alert (black line) and non-alert (gray line) periods. The y-axis shows the percentage of power contained within each bin of 0.25 Hz. The distribution of peak frequencies during alert periods can be seen in B2. C, For both alert and non-alert states, we subtracted the total power in the 2–4 Hz range from the total power in the 5–7 Hz range. D, The spontaneous firing rates of LGNd (filled circles) and VB (open circles) neurons calculated during alert (x-axis) and non-alert (y-axis) periods. E, Burst frequency for both LGNd (filled circles) and VB (open circles) neurons during alert (x-axis) and non-alert (y-axis).

Figure 2.

The monosynaptic response generated in cortical layer 4 by single thalamic neurons is the same in alert and non-alert brain states. A, An example, in the alert state, of the spike-triggered LFP (left) and STCSD (right) depth profiles generated in an S1 barrel column by impulses of a VB thalamic neuron located in the topographically aligned VB thalamic barreloid. The vertical dashed lines indicate the time of the TC spike. Only spikes having a preceding interspike interval between 100 and 200 ms were used. B, The same measures as in A but taken in the non-alert state. For both the alert (C) and non-alert (D) states, a color map was applied to the STCSD depth profiles. Gain settings and color intensities for these spike-triggered LFPs and STCSD profiles are identical in both states.

Figure 3.

The monosynaptic response generated in cortical layer 4 by single thalamic neurons does not change with brain state but does change with preceding interspike interval. A–C, STCSD depth profiles generated in V1 during alert (left) and non-alert (right) brain states by spikes of a topographically aligned LGNd neuron. The cortical responses are shown that were generated by LGNd spikes with preceding interspike intervals that were short (5–20 ms; A), intermediate (100–200 ms; B), or long (500–3000 ms; C). The color gain settings are identical for each panel. D–F, For all of the VB and LGNd neurons studied, the amplitude of the monosynaptic responses generated in the cortex are shown for TC impulses that had preceding interspike intervals that were short (5–20 ms; D), intermediate (100–200 ms; E), or long (500–3000 ms; F). Response amplitudes did not differ between alert (x-axis) and non-alert (y-axis) states. In D and E, the mean amplitude of the monosynaptic response for each state is indicated by the asterisks.

Figure 4.

Synaptic dynamics at the TC synapse do not change with brain state. Data are shown for both visual (filled circles) and somatosensory (open circles) systems and in the alert (x-axis) and non-alert (y-axis) states. Points represent the degree of reduction (percentage) in postsynaptic responses that were generated by thalamic spikes with short preceding intervals (5–20 ms) versus spikes with long preceding intervals (500–3000 ms). The asterisks indicates the mean degree of synaptic depression within each state.

Figure 5.

TC activation of fast-spike inhibitory interneurons is also independent of behavioral state. A, Examples of cross-correlograms generated in the alert state (left) and non-alert state (right) for an LGN–V1 pair (top) and a VB–S1 pair (bottom). B, The TC synaptic efficacy of four somatosensory and three visual TC–cortical pairs during alert (x-axis) and non-alert (y-axis) periods. These state comparisons of synaptic efficacy are based on TC spikes with preceding interspike intervals of 100–200 ms.