Adapting without reinforcement

Our data rule out a broad class of behavioral models in which behavioral change is guided by differential reinforcement. To demonstrate this, we showed that the number of reinforcers missed before the subject shifted its behavior was not sufficient to drive behavioral change. What’s more, many subjects shifted their behavior to a more optimal strategy even when they had not yet missed a single reinforcer. Naturally, differential reinforcement cannot be said to drive a process that shifts to accommodate to new conditions so adeptly that it doesn’t miss a single reinforcer: it would have no input on which to base this shift.

[1]  C. Gallistel,et al.  The learning curve: implications of a quantitative analysis. , 2004, Proceedings of the National Academy of Sciences of the United States of America.

[2]  David J. C. MacKay,et al.  Information Theory, Inference, and Learning Algorithms , 2004, IEEE Transactions on Information Theory.

[3]  Richard S. Sutton,et al.  Reinforcement Learning: An Introduction , 1998, IEEE Trans. Neural Networks.

[4]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[5]  P. Dayan,et al.  Decision theory, reinforcement learning, and the brain , 2008, Cognitive, affective & behavioral neuroscience.

[6]  C. Gallistel,et al.  Risk assessment in man and mouse , 2009, Proceedings of the National Academy of Sciences.

[7]  C. Gallistel,et al.  Mice take calculated risks , 2012, Proceedings of the National Academy of Sciences.