…and this is a 
Research into adult neural cell integration
Category: Academics
Posted on: October 30, 2007 7:42 PM, by Lua Yar
I found an article about new brain cells that I thought was really interesting. Researchers at the Yale School of Medicine discovered the mechanism behind how new neural cells are integrated into the adult brain. It turns out that new neural cells take a while to mature and fully integrate themselves into existing neural networks in the brain. While they are maturing, they rely on signals from other brain regions so that they do not disturb ongoing functions of the brain. They can receive input from these other regions for up to 10 days before they are ready to make any of their own outputs. So how long does it take to fully develop their synaptic connections so that they can talk to one another? Up to 3 weeks.
So why do we care; what is significant about this discovery? This mechanism sheds light on how neural cells integrate themselves into existing networks, which will impact how stem cells are used to replace neurons lost to injury or disease. The main concern is about neurons firing inappropriately, which could cause seizures or cognitive dysfunction.
The full article can be found in the Journal of Neuroscience, Vol. 27
Comments
Posted by: Duff | October 30, 2007 8:27 PM
I've always believed that if I want to remember something I had to repeat it to myself over a period of a couple of weeks. Maybe I was on to something. Certainly not from any visceral understanding of neuronal networks, but just dumb observation.
Posted by: Brian English | October 30, 2007 9:06 PM
This is very cool. I wonder why we need so many neurons when the soul is immaterial and communicates with the brain via an unmeasurable energy which violates the 1st law of thermodynamics? Seems like a waste. I thought nature was normally very frugal and wouldn't waste billions of energy requiring cells. Odd no?
/end sarcasm
[Ducks and runs before being slayed by monists.]
Posted by: Mark (Monty) Montague | October 30, 2007 9:49 PM
hey, that's cool. I wonder if anyone has tried to see if this works well to add "cell" units to simplified neural network models, like Discreet Hopfield Nets... it'd be really cool if this turns out to have a simple method for new cells to "learn the ropes" to join in a network doing the same computation before coming online. It would also be interesting if it didn't work, since qualifying the ways that things like simple feedback networks are a naive model of actual brain tissues could help our understanding as well.
Posted by: Andy J | October 31, 2007 12:31 AM
So you're saying there's still hope for me... ;-)
Posted by: caynazzo | October 31, 2007 7:15 AM
"The main concern is about neurons firing inappropriately..."
This may be true, if only because the journalist chose to end with that quote. What I find interesting is how these cells get to their final destination in the first place, neural migration, hypothesized to play a major role in learning disorders such as ADHD and dyslexia.
What I found confusing in the article was a clear distinction between neural integration and maturation.
Posted by: The flying Trilobite | October 31, 2007 8:09 AM
Fascinating. Read this last night and found myself thinking about it on my coffee-accompanied walk to work. Does the caffeine speed anything up? I certainly feel more neuronally-dense and connected.
Thanks for that post, Lua Yar.
(By the way, do you get a lot of comments about the classic videogame Yar's Revenge? 'Cuz if you are a giant space insect destroying installations, I, for one, would like to extend my servitude. Ok. Now you can make all the "Mellow Yellow" jokes you want about my name.)
Posted by: Bunjo | October 31, 2007 8:27 AM
Curious. I've read that a new habit takes 21 days to be formed and 63 days to become 'embedded'. I imagine that forming habits and integrating new neural cells must correlate... I wonder if the 63 days are needed to de-tune any conflicting old habits/old neural connections?
Posted by: Eamon Knight | October 31, 2007 9:54 AM
I'd heard something along these lines before, as being part of the reason it takes anti-depressants three or four weeks to take effect.
Posted by: goodtime johnny | October 31, 2007 3:58 PM
Interesting in another way. T-cells and b-cells have to "mature" before being let loose in the body. However, "learn" in this case actually is being tested to see if they respond appropriately. If they respond to too many possible surface proteins they are destroyed because they will mis-flag cells as being invaders and cause an auto-immune type response. On the other hand, if they never identify anything as non-self, they have no use in the immune system and are destroyed.
I wonder whether neural cells that fail to function correct are destroyed?
BTW, I'm not supposing there is anything similar in the mechanisms involved, but there may be a fundamental principle of cellular "learning" involved here.
Posted by: LeeLeeOne | October 31, 2007 4:32 PM
Can this be "reduced" to simplified mathematics, i.e., simple statistical analysis (blech!) [as so many of my personal colleagues seem to revel in]?
Sorry... am having a bad day with co-workers and their arguments for/against behaviors, ideology, social linguistics, analysis, etc.
This is the wrong box for discussion? Probably.
Long day with patients and personal research. Isn't college fun?!
Posted by: Ichthyic | October 31, 2007 10:30 PM
The main concern is about neurons firing inappropriately, which could cause seizures or cognitive dysfunction.
PZ, you should send a copy of this article to Michael Egnor ASAP.
Posted by: Evan Garber | November 1, 2007 3:00 PM
Is there a recognize mechanism that prevents premature integration possessed by mature cells - or neural networks? If so, how does it operate and what happens when it malfunctions - are there diseases which are the result of such failure - or cause failure? some questions - may be irrelevant or wrongly focused, if the mechanism is solely contained within the new cells.