At each time point, Common Average Reference treats the across-channel mean as a shared baseline and removes that value from every electrode’s voltage. The resulting channel values represent deviations from the simultaneous multichannel average, so activity that differs spatially across the array becomes more prominent for comparison and interpretation.
Broad electrode coverage matters because the average should represent distributed recording activity rather than a narrow portion of the measured field. If coverage is limited, or one channel contributes disproportionately, the calculated reference may be less representative. These conditions can alter the spatial differences that CAR is intended to clarify.
Unlike a reference tied to one electrode, Common Average Reference reduces dependence on the behavior of that single channel by using information from the recording array. This makes channel-to-channel comparisons less dependent on one selected location. The advantage is greatest when the electrodes provide broad coverage and no individual channel dominates the mean.
CAR is especially useful when the analytical goal is to examine spatial differences in neural signals rather than only the voltage at one site. Subtracting the common average changes each channel relative to the same multichannel baseline, which can support interpretation of distributed patterns and differences among electrodes in biological recordings.
A CAR workflow begins with a multichannel recording, calculates the mean voltage across available channels separately at each time point, and subtracts that mean from every channel. The transformed signals can then enter preprocessing or later analysis. Consistent channel inclusion matters because changing the set of channels changes the reference itself.
Researchers may apply Common Average Reference during EEG preprocessing when they want a common reference for subsequent event-related analysis. The method produces signals expressed relative to the array-wide mean, allowing responses to be examined across channels without relying on a single reference electrode. It therefore supports comparisons while preserving spatial differences in the recording.