Shamir M. (2006). The scaling of winner-takes-all accuracy with population size. Neural computation. 18 [PubMed]