The mammalian cortex is the outer mantle of cells surrounding the central structures, e.g., brainstem and thalamus. It is unique to mammals, and is believed to be necessary for most higher-level brain functions. Topologically the cortex is comprised of two spherical shells, corresponding to the two hemispheres. The hemispheres are connected by the corpus callosum. Cortical thickness varies mostly between 2-3 mm in the human, and is folded around the subcortical structures so as to appear wrinkled. Its average surface area is about 2200 cm2 (Zilles 1990).
It is estimated that there are roughly 1011 neurons in the human brain, and 1010 of these in the cortex. Of these, approximately 85% are pyramidal cells (Braitenberg and Schuz 1991), whose dendritic trees have a distinctive, elongated geometry that makes possible the generation of extracellular fields at large distances. The remaining 15% may be broadly classified as stellate cells, whose dendritic trees are approximately spherical, and make little or no contribution to distant fields. Of course, both cells types are interconnected to form a single dynamical network, but it is believed that the fields at large distances are dominated by pyramidal cells.
Synaptic connections in the cortex are dense. Each cortical neuron receives 104-105 synaptic connections, with most inputs coming from distinct neurons. Pyramidal cells make excitatory connections to both cell types. They make intracortical connections over lengths ranging 0.5-3 mm, and cortico-cortical connections over lengths ranging 1-20 cm. Stellate cells make inhibitory connections to both cell types. They make intracortical connections over lengths
ranging only 0.02-0.03 mm. Thus connections in the cortex are said to exhibit long-range excitation and short-range inhibition. Because of the diversity of scales of these synaptic connections, and the nonlocal nature of the cortico-cortical connections, we expect the cortex to exhibit rich spatio-temporal dynamics spanning a wide range of length and time scales.
Figure 1 shows a depiction of several layers of the human head and the positioning of EEG electrodes relative to the cortex. The folds of the cortex are such that even nearby patches of cortex can have different orientations and distances from the detectors. Section 6 shows how individual EEG and MEG detectors spatially integrate neural activity over as much as 100 cm2. Combining the data from many scalp probes, however, yields an improvement to the order of several cm2. Using the latter estimate, we must still conclude that EEG and MEG detectors integrate brain activity over a volume including as many as 107-109 cortical neurons.
Was this article helpful?