Fluctuating spike threshold independently for each neuron

[Guillaume Dumas]

Dear PyNN magicians,

Is there any simple way to make the spike thresholds varying at each time-step of the simulation?

Thanks for your time,

Guillaume

[Andrew Davison]

Dear Guillaume,

On Tuesday, 29 September 2015 18:03:59 UTC+2, Guillaume Dumas wrote:

 

Is there any simple way to make the spike thresholds varying at each time-step of the simulation?

You can probably do this using callbacks, but it would be very inefficient.

Example (I haven't tested this; having a callback equal to the timestep may not work):

dt = 0.1

sim.setup(timestep=dt)

p = sim.Population(N, sim.IF_cond_exp())

def sinusoidal_threshold(t):

    p.set(v_thresh=-50 + sin(t))

    return t + dt

sim.run(t_stop, callbacks=[sinusoidal_threshold])

It would be much more efficient to implement a model with time-varying threshold in the backend simulator(s) you want to use - Brian, NEST or NEURON, and then wrap it for PyNN.

Cheers,

Andrew