Electrophysiology of Natural and Artificial Vision - Artificial Sight: Basic Research, Biomedical Engineering, and Clinical Advances - page 359

the magnitude of an electric field falls of with the square of the distance from
the source; this means that the contribution of an individual cortical neuron
to the potential recorded at the scalp (a few centimeters away) is extremely
small.
Field potential recordings made from the body surface are often referred to
as surface potential recordings. These signals are potential differences measured
between two points on the body surface. If you picture the dipole source of
the potential difference within the conductive body, the associated electric field
can be visualized as a series of increasing shells surrounding the source (like
layers of an onion), where each shell represents all of the locations in space
that are at the same potential at a given snapshot in time. These isopotential
surfaces eventually intersect the boundary of the body, where the electrodes are
located. The potential difference recorded on the body surface results from the
two electrode locations intercepting two different isopotential surfaces. Attempts
to mathematically infer the location of the dipole source from knowledge of
the surface potentials is solving the inverse problem (as opposed to solving
for the field potential a known distance from the dipole, the forward problem)
and is called functional mapping; this is a highly developed area of biomedical
modeling. Much of the challenge in functional mapping arises from the complex
shapes of the isopotential surfaces due to anatomy and differences in conductivity
of the various tissues between the dipole source and the electrodes.
Recent trends in electrophysiology are driven by attempts to overcome the
limitations of each of the methods just described. The relative attributes of each
method are summarized in Table 20.1.
New technologies and analysis strategies are aimed at the ultimate goal
of recording from many individual neurons without sacrificing the ability to
resolve individual action potentials. The development of multielectrode arrays
attempts to scale up the extracellular recording approach, but is ultimately limited
by the displacement of tissue required to introduce the many electrodes into
the target structure. The spatial resolution of functional mapping from surface

Table 20.1. Summary of four basic strategies for recording neural activity. Existing
methods are a compromise between resolution (temporal and spatial) and the number of
individual units that can be resolved in the recorded signal.

Method Temporal resolution Spatial resolution Number of units

Intracellular
Recording

Excellent (direct
measure of V m

Excellent Very Limited (1 unit)

Extracellular
Recording

Excellent (spikes
resolved; signal is a
derivative of V m )

Very Good Limited ( ∼ 1-3 units)

Moderate ∼ 1mm 3 Moderate (100's-1000's
of units)

Local Field Potential
Recording

Moderate (spikes not
resolved)

Field (Surface)
Potential
Recording

Moderate (spikes not
resolved)

Poor (resolve voxels
of a few mm 3 )

Many (e.g. entire cortex
accessible)