spatial filter

A spatial filter combines the readings from many electrodes into a smaller set of cleaner "virtual" channels. The idea is that no single electrode sees the brain signal you care about by itself — each one picks up a bit of that signal plus a lot of shared noise. By adding and subtracting electrodes in just the right proportions, you can make the signal of interest stand out while the noise that they all share cancels away.

A homely analogy: if everyone in a noisy room hums the same background drone, you can subtract a neighbour's recording from yours to wipe out the drone and leave the voice you actually want. Spatial filters do the same trick across the scalp or cortex, mixing channels to amplify a target and suppress what is common to all of them.

There is a whole family of them, from simple fixed rules to learned ones. The common average reference subtracts the average of all electrodes; the Laplacian subtracts a ring of nearby neighbours to sharpen a local source; and CSP learns a custom mix tuned to tell two mental states apart. Cleaner channels mean an easier job for everything downstream.