Seminar starting now over zoom! Featuring @NicoleCRust #neuroscience
@NicoleCRust What a lovely talk. Rather amused, but not surprised, to learn that we find memorable the same things that monkeys do. And that damage to the hippocampus can result in overgeneralization.
@NicoleCRust I am looking forward to a whole brain #connectome for an #amniote (like, say, a juvenile small lizard, that could be done reasonably within a grant period) to understand the position of the hippocampus in whole brain circuits. Because I don't understand it.
@albertcardona TIL https://en.wikipedia.org/wiki/Amniote
Sounds intriguing! Might also inform about why smell evokes more powerful and older memories? There's some human work on that but ...
My favorite result of that line is: Both scents and photos can trigger memories, but scents tend to trigger older memories, from childhood (<10 years).
https://pubmed.ncbi.nlm.nih.gov/19686154/
I'm doing some conceptual grappling about PFC-hippocampal generalization, interference, disambiguation, etc., at the moment, but I'm a newbie. Can you recommend a particular paper for this Felleman/Van Essen model? Thank you.
Nice, thanks a million.
We have as of yet unpublished mouse data that shows an interesting disconnect between disinhibited/impulsive overgeneralization with hippocampal lesion (dorsal CA1) versus excessively deliberative overgeneralization in aged and transgenic mice (both theorized to have a deficit in PFC->hippocampal inhibitory control). Potentially interesting point to consider.
@WorldImagining @albertcardona Interesting! Please let me know when there's something to read/see.
@albertcardona Thank you! Former: crazy, right? Latter: You picked up on an important issue. By default, we think about higher stages of hierarchies as the ones responsible for generalization (e.g. abstractions). Felleman/Van Essen put the hippocampus at the very top (even above PFC). But if you destroy it, it leads to overgeneralization. That does not mean that it's not necessary for generalization (eg pattern completion) but it's an intriguing data point that constrains things.