Spiking neuron


1.Transfer function

Select the "Integrate-and-fire" model. Apply a current step dI. What happens as you vary the size of the current step?

For a given series of values of dI, measure the mean spike frequency and draw the transfer function.

How does the transfer function change if you vary:

a) the membrane time constant Taum?

b) the threshold Thrsh?

2. Phase locking

Use a periodic current. Choose a signal frequency and time interval such that you can observe about ten cycles of the sinusoid in the display window.

a) Modify the neuron parameters in such a way as to obtain one spike per cycle with a constant phase delay relative to the signal.

b) Modify the threshold so that the ratio between the signal period and the spiking period is 3/4. That is, every 4 signal periods, there should be 3 spikes.

c) Modify one of the parameters so that the ratio is now 1/2.

d) Add noise to the neuron model. You can add noise to the threshold (variable threshold) by adding a nonzero value to the "gap" to the right of the threshold (Thrsh). You can add noise to the refractory period (Refr.) in the same way. Finally, noise can be added to the current (noise current). What happens to the phase-locking phenomenon you observed previously?

3. Spike response model

Select the "Spike response model (short memory)". Use a pulse train as input and choose the signal amplitude and threshold so that it takes several input pulses to produce a spike.

a) Change the time constant values. How do they affect the postsynaptic potential?

b) As above, for the resistance R.

c) Increase the amplitude of the input pulses so that each input pulse produces several output spikes. Compare the behavior of "short term memory" with "long term memory". What is the difference?