Do neurons generate monopolar current sources?

According to the “standard model,” electric potentials such as the local field potential (LFP) or the electroencephalogram (EEG) are generated by current dipoles made by cerebral cortex neurons arranged in parallel. In this issue of Journal of Neurophysiology, Riera et al. (2012) present experimental evidence that this standard model may be insufficient to account for LFP and EEG signals in the rat brain. The authors have designed a set of technically impressive experiments that question the validity of the dipole model. We briefly summarize these findings and then speculate on possible physical mechanisms to explain these surprising results. According to the standard model, a given current source (for example due to the opening of a postsynaptic conductance as in Fig. 1A) will be instantaneously balanced by an extracellular current and a “return current”, which will enter the neuron at another location (for example the soma; see Fig. 1B). This configuration implies that a dipole will instantaneously appear in the neuron, as illustrated in Fig. 1. In these conditions, the system is described by Kirchhoff’s laws, similar to an electronic circuit (see Fig. 1, bottom, for an example of equivalent circuit). According to this model, the instantaneous dipole that appears in asymmetric neurons (such as pyramidal cells) will be responsible for the production of an electric field outside of the neurons. If these cellular dipoles are oriented in parallel, a situation which is called “open field” configuration (Lorente de No 1947; such as typically in cerebral cortex), the field generated by the different dipoles will summate and create a signal strong enough to be recorded with extracellular microelectrodes, the LFP, or even give rise to potentials recordable at the surface of the scalp, such as the EEG. Motivated by this standard model, a number of methods have appeared to estimate the dipolar sources from LFP or EEG recordings (Jones et al. 2007; Pascual-Marqui et al. 1994; Ramirez 2008). These methods are quite popular in the EEG literature, and there exists several commercial or open-source programs to perform this estimate of underlying dipolar sources from the EEG (Pascual-Marqui et al. 2002; http://www.uzh.ch/keyinst/loreta. htm) and/or from magnetoencephalographic (MEG) recordings. These estimates are of course entirely dependent on the standard dipole model. In their study, Riera et al. (2012), investigated the validity of this model in several ways. They first recorded local neuronal activity in three dimensions, using a set of multicontact electrodes inserted in the rat barrel cortex (spanning several barrels). The system records both units and LFPs at all locations, thereby providing a dense three-dimensional coverage of this area of cerebral cortex. By applying variants of the currentsource density (CSD) analysis (Nicholson and Freeman 1975), which estimates current sources (without making dipole assumptions), they demonstrate that, following whisker activation, there appears current sources and sinks, which are not necessarily balanced. Most interestingly, they designed a procedure to estimate the different multipolar components of the current profiles and quantified their respective importance. Surprisingly, while there were significant dipolar and multipolar components as expected, they also found an unexpected strong monopolar component. This component was necessary to explain the data. In a second series of experiments, Riera et al. (2012) directly tested the dipolar assumption in the rat brain. They conceived a high-resolution EEG cap for the rat brain, specifically designed for this purpose (a quite impressive achievement in itself), and used standard dipolar source estimation techniques to estimate the underlying dipoles in cerebral cortex. This estimate was aided by a three-dimensional reconstruction of the rat cerebral cortex and electrode position using magnetic resonance imaging (MRI). The MRI images were used to constrain the location of the dipoles in cerebral cortex, as is routinely done in human EEG (Dehghani et al. 2010b). In addition, the laminar LFPs were simultaneously recorded using microelectrodes. By using an approach that takes into account 1 According to Kirchhoff’s current law, the sum of the currents at any node of a circuit is zero, which implies that there cannot be any charge accumulation at any node of the circuit. Address for reprint requests and other correspondence: A. Destexhe, Unité de Neurosciences, Information and Complexité, Centre National de la Recherche Scientifique, 1 Ave. de la Terrasse (BAT 33) 91190, Gif sur Yvette, France. 2 Note that this estimate of the different multipolar contributions is made based on a very simplified model of the LFP, and this estimate could be different with a more realistic model. Fig. 1. Illustration of the flow of ions following the activation of a synaptic conductance. A: activation of a synaptic conductance at a given position in the dendrite of a neuron The conductance is assumed in this example to be associated to a net entry of positive ions. B: according to the “standard model,” the synaptic current (downward arrow) is instantaneously balanced by a return current in another region of the neuron, resulting in a dipole. The equivalent electrical circuit corresponding to this situation is shown at bottom. J Neurophysiol 108: 953–955, 2012. doi:10.1152/jn.00357.2012.