High Conductance in Dendritic Function

introduction

Most of our understanding of dendritic function has come from studies in isolated preparations like brain slices.This approach has been very successful in defining the basis of dendritic excitability and identifying subunits in neurons. These in vitro recordings have not only shown the diversity of voltage-gated ion channels in dendrites, but have also mapped their distributions and revealed how their densities change during development.However, the baseline conditions in brain slices are often very different from those in the intact brain. In vivo neocortical neurons receive continuous excitatory and inhibitory postsynaptic potentials (EPSPs/IPSPs). Such an intense background activity arises from the cortical presynaptic neurons, spontaneously spiking at low rates, and it induces the postsynaptic membrane voltage to fluctuate, resulting in an irregular spike emission. Under such conditions, the biophysical awake cortical properties of the neurons are substantially altered compared to those of a neuron at rest. So observations in vitro may not directly transfer to the in vivo situations.

In this essay I'm going to argue that the research of Wolfart et al (2005) is compatible with the findings of Destexhe, Rudolph & Pare (2003).

The findings of Destexhe et al (2003)

Destexhe et al (2003) compared the electrophysiological properties of neocortical neurons in vivo and in vitro. Measurements of cortical neurons in virto reveal characteric low input resistance, depolarization, rapid and random membrane potential fluctuations and spontaneous firing (~10Hz). The cerebral cortex is very much interconnected with 5000 to 60000 synaptic contacts per pyramidal neuron. Deshexthe et al reviewed that neuron behaviour can be explained by large opposing excitatory and inhibitory synaptic currents induced by the synaptic bombardment. This intense synaptic bombardment is leading to a 'high-conductance state' that differs markedly from the conditions measured in cortical slices in vitro. Intracellularly, this state is characterized by a high conductance, due to sustained synaptic inputs, and a considerable amount of noise, due to the seemingly random nature of these inputs. The firing rate of the neuron in the high-conductance state is not so much determined by the magnitude of the conductance as by the balance between the excitatory and inhibitory background.

In vivo recordings date back to the first intracellular recordings of central neurons in the middle of the last century, and were obtained in motor neurons of the cat spinal cord (Brock, Coombs & Eccles). Since then, such recordings have been obtained in nearly all cortical regions. These recordings were usually performed in deeply anaesthetized animals, most commonly under barbiturate anaesthesia. Barbiturates depress cortical excitability, leading to an electroencephalographic (EEG) pattern like that of slow-wave sleep. In brain slices and anaesthetized animals cells have a high input resistance ('low conductance state'), are hyperpolarized and show little spontaneous activity. This is in contrast to activated states, during which the EEG shows low-amplitude fast activity ('desynchronized EEG'), associated with asynchronous and irregular firing. Intracellular recordings shows now a depolarized and fluctuating membrane potential, a low input resistance and high levels of spontaneous firing. To characterize cortical neurons during EEG-activated states, it is necessary to perform intracellular measurements in parallel with EEG recordings. Unfortunately there are but few of these measurements.Destexhe et al speaked of EEG recordings by awake animals, those under ketamine-xylazine anaesthesia and animals under barbiturate anaesthesia. The recorded activity in those three cases contrasts with activity is usually seen in intracellularly-recorded neurons in cortical slices.

To evaluate the integrative properties of pyramidal neurons during high-conductance states, in vivo approaches are insufficient. They have the advantage of an intact network, but do not allow us to control synaptic inputs with sufficient precision. In vitro approaches leave out the background activity and the high-conductance state of neurons in an intact network. Destexhe et al (2003) used computational modelling to integrate the information from in vivo and in vitro studies. Computational modelling showed that this high-conductance state of neurons in vivo was accompanied by enhanced responsiveness and gain modulation, equalization of synaptic efficacies, probabilistic and irregular behaviour and increased temporal resolution. It also shows that synaptic activity has an important impacton the operating mode (coincidence detection versus firing-rate integration) of cortical neurons, and that enhanced voltage attenuation during high-conductance states should favour the electrical isolation of dendritic segments with respect to each other.

In dynamic-clamp experiments, in vitro electrophysiology is combined with computational modelling to create high-conductance states in cortical slices. This way, the dynamic clamp offers a way of studying neurons in slices while simulating in vivo synaptic input.The term dynamic clamp refers to a variety of hardware and software used to create artificial conductances in neurons.To simulate a particular conductance, the dynamic clamp computes the difference between the measured membrane potential and the reversal potential for that conductance, multiplies this 'driving force' by the desired amount of conductance, and injects the resulting current into the neuron. Such dynamic clamp experiments confirm that synaptic noise enhances neuronal responsiveness for small-amplitude inputs and decreases the responsiveness forr large-amplitude inputs (gain of modulation). They could also be used to test the predictions that it equalizes synaptic efficacies, increases temporal resolution, and induces probabilistic behaviour. Consistent with the increased temporal resolution, a relationship was found between the high-conductance state and the irregular firing of neocortical neurons. The most striking consequence is that neurons can resolve higher frequency inputs in high-conductance states than when quiescent. Models therefore predict that cortical neurons in high-conductance states can efficiently and rapidly track temporal variations in synaptic inputs.

After the publication of Destexthe et al (2003), a lot of researchers looked at the implications for dendritic function of the high conductance state in vivo when doing research. With help of this paper Fellous, Rudolph, Destexhe & Sejnowski (2003) state that although the study of stochastic