IGITUGraz / SpikeDetectorFuse

The nest module that implements a spike detector that terminates the simulation on large spiking

Spike Detector Fuse for NEST

This module monitors the spiking activity of the network, and throws a python exception if the spiking activity goes beyond the given threshold for the given length of time. See the comments in the code for more details.

It is a module for the NEST simulator.

Installation

With the appropriate nest-config in your PATH run ./install.sh in the root directory.

Once the installation is done, add /path/to/nest/installation/lib/nest/ to your LD_LIBRARY_PATH environment variable.

Usage

import nest
nest.Install('spikedetfusemodule')
...  # Create your neurons and connect up your network
spike_det = nest.Create('spike_detector_fuse')
nest.Connect(neurons, spike_det)
# Setting termination criterion
nest.SetStatus(spike_det, {'frequency_thresh':200.0, 'length_thresh':50.0, 'n_connected_neurons':len(exc_neurons)})
...  # Run your simulation
try:
    nest.Simulate(200)
except nest.NESTError as E:
    E_msg = E.args[0]
    if E_msg.startswith('UnstableSpiking'):
        print(E_msg)
    else:
        raise

In this example, if your average network activity stays beyond 200Hz for longer than 50ms, a nest.NESTError will be thrown with message 'UnstableSpiking in Simulate_d: The Network seems to be in a regime of unstable spiking, terminating simulation'

Important notes

• nest.Simulate() cannot be run again without resetting the kernel by running nest.ResetKernel() once an unstable spiking exception is thrown
• Data upto the simulation slice where the exception is thrown can be safely retrieved and parsed even if the above exception is thrown.