Recent talk

I recently gave a talk which I'm placing a link to in the sidebar.  The talk is  a little more than 1 hour and is a whopping 40mb, just so you know before you DL.   It basically sums up what I've been working on for the past few years.

Here is the title and summary:

Modeling network correlations in cortical tissue

from juvenile human epileptics

Summary

Models of neural tissue can make predictions about a real neural network, but these

predictions rely on the data to determine parameters. Hence, the model is only as good as

the data. I collected in vitro data removed from juvenile humans with refractory epilepsy,

and found human-specific spatial and temporal dynamics that are not found in rats. I will

first describe the general characteristics of the human data in comparison with rat data,

and my attempts to model these differences with three popular models of neural

networks: branching, pair-wise maximum entropy, and a forest fire model. I will

describe three key discoveries from this exploration: first, spatial dynamics are more

easily satisfied than temporal in both the rat and human tissue, second temporal

correlations are not captured by the branching or the maximum entropy model, and

thirdly, strong temporal correlations can be accounted for with the addition of a

parameter in the forest fire model. Finally I will suggest new questions that this research

has revealed about human tissue, and models of neural networks.