Abstract
Significance: The ability to store and retrieve information is a crucial function of memory in the brain. One well-established paradigm in which this is tested is free recall, in which a number of robust features and biases characterize the statistics of recall. Although these laws are consistently reproduced, the neuronal basis of how they emerge is not clear. In this manuscript, we propose a neural network model with a plausible learning rule that expresses these laws that may provide a mechanistic insight into memory function. In particular, our model reproduces power laws governing the recall capacity, as well as the biases that human subjects use to guide recall, such as the use of temporal and semantic cues.
All Related Versions
Funding Information
  • RCUK | Biotechnology and Biological Sciences Research Council (BB/N013956/1, BB/N019008/1)
  • Wellcome (200790/Z/16/Z)
  • Simons Foundation (564408)
  • RCUK | Engineering and Physical Sciences Research Council (EP/R035806/1)

This publication has 78 references indexed in Scilit: