I'm trying to find a paper that states otherwise -- that we can't have all neurons fire simultaneously, but all I can find is Quora questions (like this) which mentions that if all neurons will fire together we will have a seizure.
5
-
1$\begingroup$ What does "simultaneously" mean? What's the purpose of the question/what will it help you accomplish or understand? $\endgroup$– Bryan Krause ♦Dec 19, 2022 at 5:22
-
$\begingroup$ "simultaneously" means at the same time. I assume that all neurons are being used in general at some point, but my question is, can there be a given time when they are all active at the same time. The link I shared from Quora states that there isn't such time, as this will result in a seizure. I'm trying to find a publishable paper that states the same (or the opposite). $\endgroup$– PenguinDec 19, 2022 at 5:35
-
$\begingroup$ But why does it matter if they can or not? In principle, nothing prevents it, but also it would just never happen, so I don't see any purpose in asking the question. $\endgroup$– Bryan Krause ♦Dec 19, 2022 at 5:54
-
$\begingroup$ If they don't it raises some interesting ideas from a computational side. I'm writing a paper and need to ensure that the ideas are based on a real phenomena (and not an assumption). But I do understand your reasoning and appreciate the help on this! $\endgroup$– PenguinDec 19, 2022 at 6:00
-
$\begingroup$ Why does it raise interesting ideas from a computational side? Imagine, if neurons are firing at 100Hz (very fast, unusually fast for neurons, but it'll help us get to simultaneity), and you consider firing within a 1 ms window to be simultaneous (which isn't actually simultaneous, which is why I asked "what does simultaneously mean?"), the chance of all N neurons firing at the same time is equal to 1/0.1^N - you'll see that once you have a few neurons, this chance is very small, let alone 10s of billions. Nothing interesting there, just simple probability. $\endgroup$– Bryan Krause ♦Dec 19, 2022 at 14:48
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
4
Note: this answer is imprecise; read its comments below.
"...based on the Hodgkin–Huxley model calculates by the method of ion counting and power integration that an action potential consumes 2.468 × 10−7 J of biological energy produced by the hydrolysis of Adenosine triphosphate (ATP)..." (ref. https://link.springer.com/article/10.1007/s11571-018-9503-3)
Supposing we have 100 billion neurons in the brain, the total energy consumed would be:
1E11 * 2.468E-7 J = 2.468E4 J
So 24KJ to generate a single action potential for each neuron of the brain.
To lift an apple 1 meter requires roughly 1 J. So the energy needed to make each single neuron fire once would be enough to lift 24 thousand apples 1 meter, all at the same time.
So that is why it is impossible: the brain would catch fire. An we would starve.
-
4$\begingroup$ Few things wrong with this answer... First of all, action potentials are almost entirely passive. They are powered by pre-existing electrical gradients; yes, it takes energy to maintain those gradients over the long run, but this maintenance occurs over seconds to minutes; the precise timing of a spike doesn't matter. Further, all neurons of the brain are spiking all the time, some might be as slow as 1 Hz but others as fast as 100 Hz. There's no additional energy expenditure for every neuron of the brain to fire at once within, say, a second, versus firing at once within, say, 1 ms. $\endgroup$– Bryan Krause ♦Dec 21, 2022 at 16:06
-
5$\begingroup$ Lastly, the HH model is based on the giant squid; humans don't have giant squid neurons, we have human neurons, and mammalian neurons have some very important modifications like voltage-gated potassium channels that greatly minimize the metabolic costs of single action potentials by limiting the counter-currents during the rising phase of the action potential. $\endgroup$– Bryan Krause ♦Dec 21, 2022 at 16:07
-
$\begingroup$ @BryanKrause - Thank you for your constructive comments, and I would like to read a correct answer from you. From what you wrote, each neuron still needs/dissipates some amount of energy when firing (either during the spike process of before that). How much is that? $\endgroup$– PietroDec 21, 2022 at 16:46
-
2$\begingroup$ This paper ncbi.nlm.nih.gov/pmc/articles/PMC4189373 references another that estimates 0.66 J released per minute per gram of brain tissue. Most metabolic activity in the brain is synaptic currents rather than action potentials - action potentials are inexpensive energetically: what's metabolically expensive is when lots of ions move across membranes. Very few ions need to move for an AP; lots of ions move when you have competing excitatory and inhibitory conductances. I don't know what the ratio is or what you'd estimate a single spike to cost, though. $\endgroup$– Bryan Krause ♦Dec 21, 2022 at 16:58