This is a simple Izhikevich model of a neuron. The Izhikevich model is a spiking neuron that is able to replicate the behavior of many different types of neurons. The model is described by the following equations: v' = 0.04v^2 + 5v + 140 - u + I u' = a(bv - u) if v >= 30 then v = c, u = u + d
The model has four parameters: a, b, c, and d. These parameters can be adjusted to replicate the behavior of different types of neurons. The model also has an input current I that can be used to stimulate the neuron. The model is implemented in the izhikevich_neuron module.
The izhikevich_neuron module has the following ports:
clk: The clock signalreset: The reset signaluo_out: The output voltage of the neuronui_in: Configuration input for the neuronuio_in: Configuration input for the neuronThe followiing parameters and IO are exposed through these module pins:
| Name | Bits | Direction | Pins | Description |
|---|---|---|---|---|
Input Current |
5 | Input | ui[0-4] | Input current (mA) |
Neuron Mode |
3 | Input | uio[0-2] | See Table Below |
A Param |
4 | Input | ui[5-7], uio[3] | 4-bit custom A-parameter |
B Param |
4 | Input | uio[4:7] | 4-bit custom B-parameter |
Membrane Potential |
8 | Output | uo[0:7] | Signed 8-bit voltage (mV) |
| Neuron Mode | Behavior | A | B | C | D |
|---|---|---|---|---|---|
| 0 | RS (Regular Spiking) | .02 | .02 | -65 | 8 |
| 1 | IB (Intrinsically Bursting) | .02 | .02 | -55 | 4 |
| 2 | CH (Chattering) | .02 | .02 | -50 | 2 |
| 3 | FS (Fast Spiking) | 0.1 | 0.2 | -65 | 2 |
| 4 | TC (Thalamo-Cortical) | .02 | 0.25 | -65 | .05 |
| 5 | RZ (Resonator) | 0.1 | 0.25 | -65 | 2 |
| 6 | LTS (Low Threshold Spiking) | .02 | 0.25 | -65 | 2 |
| 7 | Custom | A | B | MRU | MRU |
MRU: Most Recently Used Value. For example in custom mode, if the user was previously in RS mode, the C and D values will be set to -65 and 8 respectively, but A-B will be set to the custom values.
Default state: RS mode
The cocotb test bench provides a sweep across all neuron modes and a sweep across all A and B parameters. The test bench also provides a sweep across all input currents. The test bench checks that the output voltage of the neuron is within the expected range for each configuration. This can be used to plot the output voltage of the neuron for different configurations.
This module requires a driver to interface with the neuron. The driver should be able to set the input current and the neuron mode, and read the output voltage. The driver should also be able to reset the neuron and provide a clock signal.
| # | Input | Output | Bidirectional |
|---|---|---|---|
| 0 | Input Current [0] | Membrane Potential [0] | Neuron Select [0] |
| 1 | Input Current [1] | Membrane Potential [1] | Neuron Select [1] |
| 2 | Input Current [2] | Membrane Potential [2] | Neuron Select [2] |
| 3 | Input Current [3] | Membrane Potential [3] | A Parameter [3] |
| 4 | Input Current [4] | Membrane Potential [4] | B Parameter [0] |
| 5 | A Parameter [0] | Membrane Potential [5] | B Parameter [1] |
| 6 | A Parameter [1] | Membrane Potential [6] | B Parameter [3] |
| 7 | A Parameter [2] | Membrane Potential [7] | B Parameter [4] |