 |
Theoretical Neurobiology and Computational Neuroscience |  |
Many forms of learning depend on the ability of an organism to sense and react to the adaptive value of its behavior. Such value, if reflected in the activity of specific neural structures (neural value systems), can selectively increase the probability of adaptive behaviors by modulating synaptic changes in the circuits relevant to those behaviors. Neuromodulatory systems in the brain are well suited to carry out this process since they respond to evolutionarily important cues (innate value), broadcast their response to widely distributed areas of the brain through diffuse projections, and release substances that can modulate changes in synaptic strength. Initial modeling studies utilized innate value, for example, in linking certain sets of sensory features to adaptive motor responses in perceptual categorization. A more general approach to the problem of adaptive value includes a way to modify value itself on the basis of experience. If value-dependent modulation is extended to the inputs of neural value systems themselves, initially neutral cues could acquire value. This process has important implications for the acquisition of behavioral sequences.
A synthetic neural model was used to illustrate value-dependent acquisition of a simple foveation response to a visual stimulus. When the connections to the value system were themselves plastic and thus became able to mediate acquired value, a significant improvement in the response ensued. Using a second-order conditioning paradigm, auditory discrimination was demonstrated to occur in the model in the absence of direct positive reinforcement and even in the presence of slight negative reinforcement. The discriminative responses were accompanied by value-dependent plasticity of receptive fields, as reflected in the selective augmentation of unit response to valuable sensory cues. The time-course during learning of the responses of the value system and the transfer of these responses from one sensory modality to another were found to be in close agreement with experimental results. Finally, the relation of value-dependent learning to models of reinforcement learning was addressed. The results obtained from these simulations can be directly related to various experimental findings and provide additional support for the usefulness of selectional principles in the analysis of brain and behavior. (See Publications 28, 42, and 54.)
 |
The Neurosciences Institute |
| May 18, 1998, mercurio |