A problem about the strength of the synaptic weight

[spike bush]

Hello everyone,

         As we all know , the parameters we use in CARLsim are dimensionless quantity.When I want to use the real parameter ,how should I convert it.For example ,The value of my real device is 10 Kilohm ,When I want to use it in CARLsim , How much I should set the connection of the synaptic weight ，like 0.1f ,or 1.0 f ..

[kashyaph]

Hi, that's a matter of scaling. You can try different scaling factors of the parameters as long as they do not saturate or under-excite the neurons, i.e. the linear slope part of the tuning curves. Another approach would be to first make a rate model of your system and then convert it to a spike model. This is easy for simple models, such as LIF, which is now supported by CARLsim4. 

For further on how to do that, you can read Chapter 7 of Dayan and Abbott Theoretical Neuroscience book or this paper:

https://arxiv.org/pdf/1510.08829.pdf