Real-time neural-network models provide a conceptual framework for formulat
ing questions about the nature of cognition, an architectural framework far
mapping cognitive Functions to brain regions, a semantic framework far def
ining terms, and a computational framework for testing hypotheses. This art
icle considers key questions about how a physical system might simultaneous
ly support one-trial teaming and lifetime memories, in the context of neura
l models that test possible solutions to the problems posed. Model properti
es point to partial answers, and model limitations lead to new questions. P
lacing individual system components in the context of a unified real-time n
etwork allows analysis to move from the revel of neural processes, includin
g teaming laws and rules of synaptic transmission, to cognitive processes,
including attention and consciousness.