Localist representation can improve efficiency for detection and counting

Almost all representations have both distributed and localist aspects, depe nding upon what properties of the data are being considered. With noisy dat a, features represented in a localist way can be detected very efficiently, and in binary representations they can be counted more efficiently than th ose represented in a distributed way. Brains operate in noisy environments, so the localist representation of behaviourally important events is advant ageous, and fits what has been found experimentally. Distributed representa tions require more neurons to perform as efficiently, but they do have grea ter versatility.