Psychology & Neuroscience Stack Exchange is a question and answer site for practitioners, researchers, and students in cognitive science, psychology, neuroscience, and psychiatry. It only takes a minute to sign up.
Sign up to join this community
I read a NY Times article about the European effort to simulate a human brain and the criticism regarding the (non-)feasibility of such an endeavor (not to mention the astronomical costs).
In this article, there's mention of the current inability to simulate the 302 neurons in a nematode brain, so my (layman's) question is simple:
why is it so hard to simulate a neuron?
Complexity. That is, the fundamental laws governing the behavior isn't terribly difficult to comprehend or model. However, putting it all together and hoping it reproduces the correct behavior is a great challenge.
Once you develop a model that has the appropriate underlying mechanisms in it's components, you still have to put them all together and choose the right parameters and hope you didn't miss anything in terms of modulators that would actively adjust those parameters. So even if the functional mathematical form of your model is correct, you still have 5-10 parameters for each neuron in the network that you have to get to the right value, including the parameters describing the strength and dynamics of the coupling between neurons. Just being off by values as small as .001 can sometimes mean the difference between a working model and a failing model. Now you have 5-10 parameters for 300 neurons... that's 1500-3000 parameters that all need to be "correct" at the same time.
That being said, it's not like we are exceedingly accurate in our models of even single neurons. We make simplifications and take averages. There's lots of second messenger and genetic processes involved in neuron function that many models neglect. This isn't meant to disparage modeling. It implies that no complex mathematical model will probably ever be able to generalize the neuron for all cases. Instead, we make models focused on asking specific questions. We investigate two or three mechanisms at a time and hold all the other considerations constant in order to isolate those mechanisms we're interested in. So a model I develop to study the effects of the diversity of calcium channels on neuron excitability will have many differences (in parameters selection) from the same model used by somebody studying diversity of potassium channels. And in some cases, we may even simplify less relevant aspects of the model, e.g., I may only have one potassium channel in my neuron to generalize and average all potassium channels so that I can focus only on calcium diversity and not fuss with more parameters.
I don't think it is hard to simulate a neuron. See NEURON.
Simulating a brain however is a much more difficult (and unaccomplished) task, even if the brain only consists of 302 neurons.
