Brain is most expensive organ of human body. It takes up more metabolic energy (approx. 20%) than in accord with ratio of it's weight with rest of the body weight. For the same reason it has to be efficient and robust to utilize it's energy budget. The human brain is complex in its working and very little is understood about its working and language. It's working is intractable in the sense that simulating all of its basic processing element (neuron) is not possible with today computing technology. The limitation of creating such an infrastructure which can simulate whatever is known, thanks to neuroscience, is one of the barrier stopping research community to empirically model working of the brain. Moreover, it is not only neurons and synapses, which are emulated in artificial neural networks in the over-generalized sense, plays major role in working of the brain but many supporting structures eg. Glia cells also have critical role to play. But, the problem is glia cells don't communicate through electrical signalling so recording their activity in-vitro is very hard. Another, hurdle is time-space trade off being faced by recording instruments. For example, MRI is good in spatial resolution, fMRI is tuned between spatial and time resolution but not good in recording very complex processing, and EEG/MEG are very poor in spatial resolution. These all challenges have to be resolved one by one or in parallel so that true understanding about working of the brain can be internalized.
Comentários