Brain performance depends on the efficient and rapid exchange of information across neural ensembles that are dispersed spatially. It is overwhelmingly well established that information is communicated proximally, chiefly by virtue of synaptic connections. That is, one neural structure -the dendrite- receives the information that a complementary other -the axon- emits. In this picture, the influence of Shannonís theory of communication cannot be overstated. Synaptic networks have been supported by a deluge of empirical evidence that in turn has helped to shape the entire field of Neuroscience. The scheme of information exchange however suffers delays of several tens of milliseconds for the communication of information between distant neural groups. Here we examine a theoretical proposal that brains also operate in the currency of spatial patterns of extracellular fields. On the extracellular side, dendrites are exposed to a local ionic environment that changes over time under the influence of proximal synaptic release and global activity-dependent fluctuations of extracellular fields. Fluctuations arising from local synapses have the largest magnitude and as such have received most experimental attention. Nonetheless low amplitude fluctuations of extracellular fields driven by the activity of distant neural population have the potentiality to convey information. We hypothesize that neurons could employ their tridimensional -spatially extended- dendritic arborescence to sample extracellular field gradients thereby facilitating quasi-instantaneous ìawarenessî of the global patterning of brain activity. We identify recent empirical evidence that renders this hypothesis plausible and draw putative implications for empirical and theoretical neuroscience. In particular, it is argued that this hypothesis, if supported by empirical evidence, has profound consequences for artificial neural networks and brain in silico: the spatial arrangement of neural populations in which distance and orientation are key variables is missing from most computational models of the brain. Our hypothesis would also bring fresh light to spatial morphosis occurring in the evolving and developing brain: putatively, directionally favorable arrangements of neural ensembles may explain idiosyncrasies in performance. With the growing importance of machine intelligence in todayís society and only partial successes garnered by artificial systems modeled upon synaptic principles, we conclude that the extracellular field hypothesis deserves further theoretical and empirical exploration.
* This work has been presented at the Dynamic Brain Forum held in Carmona, Spain, September 3rd-5th, 2012.