I am working on spiking neural networks and i simulated the networks under random inputs (from gaussian distribution). The connections between them are sparse and random. It is obvious that the neurons which are connected can show linear relations with respect to inputs, but how about any neurons that are unconnected.
For example group of neurons in the cluster A (a1,a2,a3) and another cluster B(b1,b2,b3). Can we assume that that unconnected neurons (cluster A (a1), cluster B(b1)) cannot show correlations, if they are driven by random stimuli? Can i get references regarding neuronal correlations under random stimuli?
If two neurons are connected by polisynaptic connections (e.g., neurons A & B connected through neuron C: A --> C --> B), correlations between neurons A and B will be attenuated but might still be significant.
Statistically, you may find [even strong] correlation between any two random variables though it does not mean a causation exists between them.
Practically, computer simulators rely on random generators to derive random numbers which:
The above issues may cause a fake correlation appear in simulation result.
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity