Hi,
Consider you have a network of two layers. the first layer contains N_e excitatory neurons and N_i inhibitory ones, fully connected to the second layer having N_o neurons. (ex. 6 exc and 4 inh neurons and 2 output neurons → 20 synapses )
I have the number of 5 runs and want to set random values (for weights) after each run for the number of randomely chosen synapses with a fixed probability , connected to second layer. For example,if I had 30 synapses,the weights of 0.2 of them( the number of 4) would be changed to a new random value after each run.
Thanks in advance.

This would set the random weight for each synapse with a probability of 20% – in the case of 20 synapses, this would change the weights for 4 synapses on average, but it will sometimes change the weight for 3 or 5 of them (Bernoulli trials → Binomial distribution).

If you want to change the weights of exactly 4 synapses, you can use: