There are three main arguments that axons make negligible contribution to EEG-MEG. First, the quadrupolar field falls of rapidly with distance, and is likely to be dominated at the scalp by the dipolar fields of dendrites. Second, axons are not arranged systematically in the cortex as are the apical dendrites of pyramidal cells, thus the geometric superposition of axonal fields can not occur to the same degree as it does for dendrites. Third, action potentials have ~ 1 ms duration, and therefore have a dominant frequency near 1000 Hz. The EEG-MEG signal has most of its power below 100 Hz, which is more like the time scale over which dendritic potentials vary. The extracellular field due to a single neuron is not detectable at the scalp, but the superposition of many synchronously active neurons is detectable. In order for action potential fields to superimpose to measurable levels at the scalp, it would be necessary for the action potentials of multiple neurons to occur with high temporal synchrony. While it does appear that spike synchrony plays a fundamental role in neural processing, the requirements on synchrony are much more demanding for axons than dendrites due to the shorter duration of their potentials. For these reasons, it is expected that EEG-MEG is dominated by dendritic potentials. Up to the size of the integration volume of an EEG or MEG sensor, dipolar sheets created by synchronously active patches of cortex make contributions to the scalp potential in proportion to their size.
For each argument against axon contributions, there is a reasonable counter-argument, First, dipole fields likely dominate the scalp potential, but that does not mean that quadrupole fields are totally negligible. Second, axons run in fiber bundles, and synchronous input to their neurons generates a compound action potential. Third, sensory input typically generates a neural response with abrupt onset and high degree of neural synchrony, at least at the dendritic level. This increases the firing probability in time and can increase spike synchrony. Thus spike synchrony in fiber bundles could potentially superimpose to be measurable at the scalp. Thus, although cortical pyramidal dendrites likely dominate axonal fields in resting EEG, action potentials could conceivably contribute to the scalp potential, particularly in early sensory response. Still, the short duration of spikes puts their power at high frequencies, which are filtered out in many EEG recordings.
Was this article helpful?