Characterize single cell dynamics in NetPyNE.

In batch_simulation.py, simulations can be run one at a time or in batch. Parameter sets are handled by the ParameterGrid module from scikit-learn. A dictionary consisting of lists of parameter values is passed to ParamaterGrid and a list of all possible paramater value combinations is returned. The default parameter set can be found in config/default-config.py. All parameters defined with ParameterGrid will overwrite the default values.

Current supported models:

• Hay et al. 2011 Layer 5b Pyramidal Cell from NeuroML-DB
• Hay and Segev 2015 excitatory (AMPA, NMDA) and inhibitory (GABAA, GABAB) synpases from ModelDB

Output Location

All output from the simulation will be saved in the following path:

models/NEURON/<nmldb-id>-NEURON/output/<sim_name>/<sim_label>

Example Parameter Sets

Injected Current

Single Simulation

Run a simulation for 1000 ms with injected current to the soma with an amplitude of 0.44 nA for 600 ms with a 200 ms delay.

paramGrids = {'sim_name': ['INJ'],
              'sim_flag': ['iclamp'],
              'nmldb_id': ['NMLCL000073'],
              'enable_syns': [False],
              'input_amp': [0.44],
              'input_sec': ['soma_0'],
              'sim_dur': [1000],
              'stim_dur': [600],
              'stim_delay': [200]}

Batch Simulations

Run simulations for 1000 ms each with injected current to the soma with an amplitudes of 0.44 nA, 0.63 nA, and 0.8 nA for 600 ms with a 200 ms delay. ParameterGrid will return 3 parameter sets with unique values for the parameter input_amp.

paramGrids = {'sim_name': ['INJ'],
              'sim_flag': ['iclamp'],
              'nmldb_id': ['NMLCL000073'],
              'enable_syns': [False],
              'input_amp': [0.44, 0.63, 0.8],
              'input_sec': ['soma_0'],
              'sim_dur': [1000],
              'stim_dur': [600],
              'stim_delay': [200]}

Synaptic Current

Single Simulation

Run a simulation for 1000 ms with 50 excitatory synapses on basal dendrites that are stimulated by Poisson spike trains with frequency 40 Hz for 600 ms with a 200 ms delay.

paramGrids = {'sim_name': ['EXC'],
             'sim_flag': ['syns'],
             'nmldb_id': ['NMLCL000073'],
             'syns_type': ['basal'],  
             'num_syns_E': [50],
             'sim_dur': [1000],
             'stim_dur': [600],
             'stim_delay': 200]}

Batch Simulations

Run simulations for 1000 ms each with 50 and 100 excitatory synapses on basal and apical distal dendrites that are stimulated by Poisson spike trains with frequency 40 Hz for 600 ms with a 200 ms delay. ParameterGrid will return 4 parameter sets with unique combinations of the parameters syns_type and num_syns_E.

paramGrids = {'sim_name': ['EXC'],
             'sim_flag': ['syns'],
             'nmldb_id': ['NMLCL000073'],
             'syns_type': ['basal', 'apical_distal'],  
             'num_syns_E': [50, 100],
             'enable_syns': [True],
             'record_LFP': [False],
             'sim_dur': [1000],
             'stim_dur': [600],
             'stim_delay': [200]}

Combined

Run simulations with different input types.

# unique parameters for each parameter set
param_sets = {'syns': {'enable_syns': [True],
                       'syns_type': ['basal', 'apical_distal'],
                       'num_syns_E': [50, 100]},
              'iclamp': {'enable_syns': [False],
                         'input_amp': [0.44, 0.63, 0.8],
                         'input_sec': ['soma_0']}}

paramGrids = []

for sim_flag, param_set in param_sets.items():
  # common parameters between parameter sets
  paramGrid = {'sim_name': ['IN'],
               'sim_flag': [sim_flag],
               'nmldb_id': ['NMLCL000073'],
               'record_LFP': [False],
               'sim_dur': [1000],
               'stim_dur': [600],
               'stim_delay': [200]}
	            
  for param_name, param in param_set.items():
    paramGrid[param_name] = [param]
		
  paramGrids.append(paramGrid)

Different number of synpases for each location

Run simulations with different numbers of synapses for each group of synapse locations.

# unique parameters for each parameter set
param_sets = {'syns': {'enable_syns': [True]}
group_num_syns = {'basal': [50, 100],
                  'apical_distal': [200, 300]}

paramGrids = []

for sim_flag, param_set in param_sets.items():
  for syn_group, num_syns in group_num_syns.items():
    # common parameters between parameter sets
    paramGrid = {'sim_name': ['IN'],
                 'sim_flag': [sim_flag],
                 'nmldb_id': ['NMLCL000073'],
                 'syns_type': [syn_group],
                 'num_syns_E': num_syns,
                 'record_LFP': [False],
                 'sim_dur': [1000],
                 'stim_dur': [600],
                 'stim_delay': [200]}
  	            
    for param_name, param in param_set.items():
      paramGrid[param_name] = [param]
  		
    paramGrids.append(paramGrid)