Friday, October 24, 2008

Scale-free learning: the synapse

Synaptic Learning Rules and Sparse Coding in a Model Sensory System (Finelli et al)

The major impetus for the scale-free/fractal, modular, stochastic model for cognition presented in my last post is to investigate the link between the structure and function of the brain with (at least some) of the structure and function of the mind. Although it is not yet specified in the graphical models, my thought is that as we explore the deeper nestings of cognitive resources, we'll get closer and closer to clear links with descrete brain structures like neural networks or even individual neurons. I think it's important to show that learning is not just a mental event, but also a brain event; ie: when learning occurs it is the result of physical and chemical changes in the brain. Although I'm not necessarily arguing that the mental changes that occur in learning are themselves physical and/or chemical, the modeling of learning at the cognitive scale involves changes in the relationships and patterns of activity among cognitive resources.

The above-linked article from Finelli et al uses the locust olfactory system to demonstrate that learning occurs in the synapse level (a change in behavior at the synapse). Although certainly the human brain is much more complex than the locust brain, an interesting ability in both is to be able to process sensory data in a way that involves fewer and fewer neuron activations as the data is more and more processed. The authors found that the activity of a very small number of cells in higher-level processing of smells were sufficient for the locust odor recognition. Interestingly, the authors also suggest that exposure of the olfactory system to a wide variety of odors helps the system to process new odors more effectively.

While I'm not sure that the demonstration of learning at the synaptic level has any practical application to human learning experiences in the classroom or in other traditional learning settings other than to help educators appreciate more deeply that so many different people *can* actually learn the same skills and concepts in the same way, I do think that the implications of this research can improve educational experiences by helping us to focus on how the brain processes information. Particularly, I think it's interesting and important to note that early exposure to a wide variety of information helps the brain to process new information more effectively at later points in time. This suggests that our educational systems, particularly those targeted at young students, should strive to present very broad curricula - perhaps even at the expense of depth - so that later in life those learners can process the new and more complex information that they are exposed to more deeply.