This study evaluated the recent proposition that LFP recordings generally sample over an extremely confined spatial extent of several hundred microns surrounding the electrode contact. We find that through volume conduction, the LFP typically spreads well beyond this micro-domain extent, and indeed is observable many millimeters distant to the active neuronal tissue in which it is generated. It is worth noting that the conclusion the LFP in general spreads only over a ~250 μm domain is fundamentally inconsistent with the evidence indicating that stimulus-evoked and event-related potentials recorded on the scalp in humans reflect a summation of LFP generated in the brain (Luck, 2005;Mitzdorf, 1985;Nunez et al., 1991;Nunez and Srinivasan, 2006;Schroeder et al., 1991). We have discussed a number of ways in which LFP recordings can be managed to improve their spatial resolution and the precision of their physiological interpretation. We conclude that both physiological factors (e.g., strength, spatial extent and symmetry of activation in the neuronal substrate), and technical factors (e.g., electrode reference site) are critical to understanding the source and sampling area of an LFP, and that any general model of the LFP must account for these factors.
