[CSTP] IM refractory period only works with RUNGE_KUTTA4

[nmsutton]

Currently, when using the feat/ca3net (CSTP) branch, users should know that the Izhikevich model (IM) refractory period (RP) only works with the RUNGE_KUTTA4 (RK4) integration method. Adding a forward Euler (FE) version of IMRP would resolve this issue. As a short-term solution, if there are issues with CSTP running using FE, an assert statement could prevent simulations with the branch from running unless RK4 is used. An update relevant to RP has been made in a forked version of the ca3net branch link. It corrects kernel_findFiring() code for allowing any RP time to be used correctly.