Spatial reach of extracellular recording
The spatial reach (or listening sphere) of an extracellular electrode is the effective radius within which single-neuron activity contributes detectably to the recorded signal. It is not a hard boundary but a signal-to-noise threshold: because each neuron's contribution decays with distance while uncorrelated background activity from the surrounding neuropil sets a noise floor, only cells close enough to rise above that floor are resolvable as sorted units. For spikes (high frequency) this radius is typically a few tens of microns; for the low-frequency LFP it is far larger and much debated, ranging from a few hundred microns to millimeters depending on the spatial coherence of the underlying synaptic currents.
Reach depends on frequency band, electrode size (a larger contact averages over more tissue and integrates a bigger volume but with lower single-unit selectivity), and the correlation structure of the sources: spatially correlated synaptic inputs summate constructively and remain visible far away, whereas incoherent spikes cancel. This is why spike yield and LFP interpretation are governed by different geometry, and why claims about how local the LFP is remain contested.
The LFP is often described as sampling a fixed volume; in reality its reach is set by the correlation of the sources, not by a physical constant, so the same electrode can have a small effective reach for one process and a large one for another.