作者: Stanislas Dehaene
DOI:
关键词:
摘要: In this chapter, I put together the first elements of a mathematical theory relating neurobiological observations to psychological laws in domain numerical cognition. The starting point is postulate neuronal code whereby numerosity—the cardinal set objects—is represented approximately by firing population numerosity detectors. Each these neurons fires certain preferred numerosity, with tuning curve which Gaussian function logarithm numerosity. From logGaussian code, decisions are taken using Bayesian mechanisms log-likelihood computation and accumulation. resulting equations for response times errors classical tasks number comparison same–different judgments shown tightly fit behavioral neural data. Two more speculative issues discussed. First, new chronometric evidence presented supporting hypothesis that acquisition symbols changes mental line, both increasing its precision changing coding scheme from logarithmic linear. Second, examine how symbolic nonsymbolic representations numbers affect performance arithmetic computations such as addition subtraction.