In the absence of any environmental stimuli or behavioral tasks, the "resting" brain generates a continuous stream of electrical activity that can be recorded from the scalp, as electroencephalographic (EEG) waves. The intracellular correlate of this endogenous cerebral activity is characterized by background membrane voltage fluctuations (also known as "synaptic noise"), which are composed of a combination of excitatory and inhibitory synaptic potentials that reflect the ongoing activity of afferent networks 1,2. This spontaneous activity varies in frequency and amplitude with the different states of vigilance. Elucidating the impact of network activity on the excitability and responsiveness of single neurons is one of the major challenges of neurosciences 3,4.
Many experimental and computational studies have explored the functional impact of ongoing synaptic activity on the integrative properties of neurons. However, the role of the different neuronal parameters impacted by the background synaptic noise remains elusive. For instance, the mean level of membrane depolarization has been found positively 5,6 or negatively 7-9 correlated with the ability of sensory inputs to trigger action potentials. Moreover, whereas some investigations suggest that fluctuations of the membrane potential, resulting from a continuously varying stream of afferent synaptic inputs, strongly affect the responsiveness of single neurons by modulating the gain of their input-output relationship 3,10-13, others indicate that changes in membrane input conductance mediated by shunting inhibition are sufficient to modulate the neuronal gain regardless of the magnitude of membrane fluctuations 14,15. Finally, recent studies performed on awake animals stressed how the processing of sensory information in single neuron critically depends upon the state of vigilance and the current behavioral demand 16,17.
A straightforward strategy to elucidate the functional role of a given process in a highly interconnected system is to determine how its absence specifically alters the functioning of the system. This method has been extensively used in neuroscience research, for example using experimental lesions or inactivation of different brain areas 18-21, or pharmacological blockade of specific ion channels 22,23. Notably, it has been applied in vivo to unveil how functional connectivity and network dynamics affect single cell computation 24-27. However, to date local manipulations intended to block the firing of neurons and/or perturb their basic biophysical properties can be partially effective and are limited to relatively small brain volumes 28.
To overcome these limitations, we developed a new in vivo experimental approach in the rat to compare the electrophysiological properties of single neurons recorded in a given brain state, i.e., embedded in a particular network dynamic, to those obtained after complete suppression of the whole brain synaptic activity 29. In the control conditions, two distinct cortical dynamics could be generated. Sleep-like electrocorticographic (ECoG) patterns were induced by injection of moderate doses of sodium pentobarbital. Alternatively, fast ECoG waves of small amplitude comparable to the cortical activity underlying the waking state (waking-like pattern) could be produced by injection of fentanyl. Subsequently, while maintaining the same ECoG and intracellular recording, a complete silencing of endogenous brain electrical activity was obtained by systemic injection of a high dose of sodium pentobarbital, characterized by isoelectric ECoG and intracellular activities. Because the induction of such an extreme comatose could potentially have fatal consequences on biological functions, a careful and continuous monitoring of the physiological variables was essential. Therefore, we meticulously followed the heart beat frequency, the end-tidal CO2 concentration (EtCO2), the O2 saturation (SpO2) and core temperature of the rat throughout the experiments.
We evaluate single neurons properties during these different states using sharp microelectrodes, which are particularly suited for long and stable recordings in vivo. The procedure described here, can be combined with other electrophysiological and imaging approaches and could be extended to other animal models.