LogFAQs > #923062278

AsucaHayashi posted...

so like, you're like, a really smart person, huh?

Naw, most academics are just normal people who spend a lot of time trying to understand very specific things. Lots of domain specific knowledge.

PatrickMahomes posted...

What does this all mean in gamefaqs human words

Well, so, the neuron is the basic functional unit in the brain, right? But action potentials, the 'firing' of a neuron, happen on the scale of milliseconds. But we know behavior and memory does not. If it did, we wouldn't be able to hold a conversation, because by the time I ended the speech output of a single word, I would have forgotten that I started speaking it. So, for long term memory this is something that's easy to reconcile, structural changes occur that modify the strength of certain connections between neurons, creating a 'memory' composed of the synchronous activation of a set of neurons. However, for short term memory/working memory, it's unclear how this happens, because it's not operating on a timescale that would allow for structural modifications.

So, basically, a large class of working memory models connect together a bunch of neurons together with balanced inhibition that allows for a reverberating signal, allowing for the maintenance of information across longer timescales. My dissertation is an exploration of these models, examining what assumptions they hold, what predictions they make, and how to generate data to constrain parameter selection for these models.
---

=E[(x-E[x])(y-E[y])]