Discrepancy in Synapse-Replay Simulation Outputs between Neurodamus and BlueCelluLab #75
Comments
|
Thanks a lot @dhuruvapriyan for reporting this. May I ask how you are changing the release probability? Do you modify the simulation config for it? |
|
Hi @anilbey, "connection_overrides":[
{
"name": "Make_Use_1",
"source": "hex_O1",
"target": "hex_O1",
"synapse_configure": "%s.Use = 1"
}
] |
|
Hi @anilbey, I believe the issue might be related to random number streams. This seems to be connected to #72. Please let me know if you would like me to conduct validation experiments using both of these code bases. Thanks, |
|
Hi again @dhuruvapriyan. It's great that you are advancing on the issue.
Oh that would be great. I am currently stuck on this below issue with the Zenodo circuit. The simulation you are using - as well as the simulations provided on Zenodo - are setting the E.g. "conditions": {
"extracellular_calcium": 1.05,The mechanisms provided in Zenodo under O1_mods.xz however do not contain the How did you run BlueCellulab? Did you download the GluSynapse.mod elsewhere and compiled all the mechanisms together? |
|
Btw. last Friday the docker image of Neurodamus is created. It is easier to install Neurodamus now. Docker image: https://hub.docker.com/r/bluebrain/neurodamus |
|
Dear @anilbey, Thank you for your prompt response regarding this issue. We have recently rebuilt Neurodamus using the newly released Ecker's SSCx plasticity dataset, which can be found here: https://zenodo.org/record/8158472. Subsequently, we recompiled the "mods" present in the SSCx_plasticity dataset and incorporated them into BlueCelluLab. I successfully simulated VPM fibers with GluSynapse using BlueCelluLab. However, I encountered certain challenges when operating BlueCelluLab with the SSCx_plasticity model. I shall create a new ticket to provide a detailed description of these challenges encountered during the simulation of GluSynapse with BlueCelluLab. Thanks, |
|
Hey @anilbey, I'm reaching out to revisit the matter of the plasticity simulations, as they are critical to our current project. Previously, I successfully synchronized the CPre and CPost parameters in both BlueCelluLab and Neurodamus platforms. Yet, I've encountered a discrepancy: the synaptic voltage traces are misaligned when the USE parameter differs from 1. It appears that the 'use_scale_factor' might be improperly set to adjust the USE values. I'd appreciate your input on this observation. I am happy to assist you by running any necessary simulations as needed. Thanks in advance! Regards,
When Use = 1: BlueCelluLab Code: import pickle
from pathlib import Path
import numpy as np
from matplotlib import pyplot as plt
%matplotlib inline
import bluecellulab
from bluecellulab import SSim
bluecellulab.neuron.load_mechanisms("./circuit_sonata/")
cfg_json = ".../simulation_config.json"
sim = SSim(cfg_json,print_cellstate=True)
post_gid = ('S1nonbarrel_neurons', 170923)
pre_gid = ('S1nonbarrel_neurons', 170982)
bluecellulab.neuron.h.tau_effca_GB_GluSynapse = 278.3177658387
bluecellulab.neuron.h.gamma_d_GB_GluSynapse = 101.5387594661
bluecellulab.neuron.h.gamma_p_GB_GluSynapse = 216.1841700668
pre_spike_times = np.array([ 817, 829, 884, 995, 996, 1000, 1080, 1135, 1335, 1403])
sim.instantiate_gids([post_gid], add_synapses=True,
pre_spike_trains={pre_gid: pre_spike_times}, add_projections = False, add_minis = False, intersect_pre_gids=[pre_gid])
t_vec = bluecellulab.neuron.h.Vector()
t_vec.record(bluecellulab.neuron.h._ref_t)
v_synapses = {}
syns = {}
for post_gid in sim.cells:
if post_gid == post_gid:
n_of_syns = len(sim.cells[post_gid].synapses)
for syn_id in sim.cells[post_gid].synapses:
syns[syn_id] = sim.cells[post_gid].synapses[syn_id].hsynapse
if ("GluSynapse" in str(syns[syn_id])) and sim.cells[post_gid].synapses[syn_id].hsynapse.gmax0_AMPA:
v_synapses[syn_id] = neuron.h.Vector()
_ = v_synapses[syn_id].record(syns[syn_id]._ref_vsyn)
duration = 2500
sim.run(duration, cvode=False)Simulation Config: {
"manifest": {
"$CURRENT_DIR": "."
},
"run": {
"dt": 0.025,
"tstop": 2500.0,
"random_seed": 12345,
"using_cvode": 0
},
"conditions": {
"extracellular_calcium": 2.0,
"v_init": -80.0,
"spike_location": "AIS",
"mechanisms": {
"ProbAMPANMDA_EMS": {
"init_depleted": true,
"minis_single_vesicle": true
},
"ProbGABAAB_EMS": {
"init_depleted": true,
"minis_single_vesicle": true
},
"GluSynapse": {
"init_depleted": false,
"minis_single_vesicle": false,
"cao_CR": 2.0,
"tau_effca_GB": 278.3177658387,
"gamma_d_GB" : 101.5387594661,
"gamma_p_GB" : 216.1841700668
}
}
},
"target_simulator": "CORENEURON",
"network": "/project/ctb-emuller/datasets/SSCx_plasticity/O1_2023a_Ecker/circuit_config.json",
"node_sets_file": "./node_sets.json",
"node_set": "test_circuit",
"output": {
"output_dir": "./reporting",
"spikes_file": "out.h5"
},
"reports": {
"soma": {
"cells": "hex_O1ExcitatoryPlastic",
"type": "compartment",
"variable_name": "v",
"unit": "mV",
"dt": 1.0,
"start_time": 0.0,
"end_time": 122500.0
},
"rho": {
"cells": "hex_O1ExcitatoryPlastic",
"type": "synapse",
"sections": "all",
"variable_name": "GluSynapse.rho_GB",
"unit": "nd",
"dt": 0.025,
"start_time": 0.0,
"end_time": 122500.0
},
"v_syn": {
"cells": "hex_O1ExcitatoryPlastic",
"type": "synapse",
"sections": "all",
"variable_name": "GluSynapse.vsyn",
"unit": "mV",
"dt": 0.025,
"start_time": 0.0,
"end_time": 122500.0
},
"effcai_GB": {
"cells": "hex_O1ExcitatoryPlastic",
"type": "synapse",
"sections": "all",
"variable_name": "GluSynapse.effcai_GB",
"unit": "mM",
"dt": 0.025,
"start_time": 0.0,
"end_time": 122500.0
}
},
"inputs": {
"s1_syns_spikes": {
"node_set": "hex_O1ExcitatoryPlastic",
"input_type": "spikes",
"delay": 0.0,
"duration": 2000.0,
"module": "synapse_replay",
"spike_file": "./input_spikes/s1_syn_spikes.dat",
"source": "hex_O1"
}
},
"connection_overrides": [
{
"name": "plasticity",
"source": "hex_O1Excitatory",
"target": "hex_O1Excitatory",
"modoverride": "GluSynapse",
"weight": 1.0
},
{
"name": "init",
"source": "hex_O1",
"target": "hex_O1",
"spont_minis": 0.0,
"weight": 1.0
},
{
"name": "VPM_init",
"source": "proj_Thalamocortical_VPM_Source",
"target": "hex_O1",
"spont_minis": 0.0,
"weight": 1.0
},
{
"name": "POm_init",
"source": "proj_Thalamocortical_POM_Source",
"target": "hex_O1",
"spont_minis": 0.0,
"weight": 1.0
}
]
}s1_syn_spikes.dat |
|
Thanks @dhuruvapriyan for the detailed report. It is good that you narrowed the issue down to the use of |
|
Hi @dhuruvapriyan sorry for the late reply I returned to the office this week. In the meantime my colleague was looking into this issue. I believe the issue is raised from these lines. self.hsynapse.Use_d = self.syn_description["Use_d_TM"] * \
self.syn_description["u_scale_factor"]
self.hsynapse.Use_p = self.syn_description["Use_p_TM"] * \
self.syn_description["u_scale_factor"]In the code above |
|
Ignore my previous message. That part of the code looks correct. The issue may be due to the use of these parameters in the simulation_config. "tau_effca_GB": 278.3177658387,
"gamma_d_GB" : 101.5387594661,
"gamma_p_GB" : 216.1841700668What are these parameters? Is Neurodamus reading them? |
|
Hi @anilbey, I'm sorry for not getting back to you sooner. Those are the parameters fitted for plasticity experiments. All three parameters are computed along with others to determine depression and potentiation thresholds. |
Hi @anilbey,
I would like to report a discrepancy between the two codebases when using the synapse replay feature. I attempted to simulate synapses of a cell using SScx dataset, specifically focusing on the edge population labeled "S1_nonbarrel_neurons__S1_nonbarrel_neurons." The somatic voltage traces obtained from BlueCelluLab do not align with those from Neurodamus. Interestingly, when the release probability is set to 1, the traces do match, suggesting that the issue may lie in my random seed settings. I have attached my simulation_config, spikes files, and results for your review. Please let me know if you need additional details to help resolve this issue. Thank you in advance!
Matching Results for Use = 1:
My simulation_config.json:
{ "run": { "dt": 0.025, "tstop": 2000, "random_seed": 12345 }, "conditions": { "extracellular_calcium": 1.05, "v_init": -80.0, "spike_location": "AIS", "mechanisms": { "ProbAMPANMDA_EMS": { "init_depleted": false, "minis_single_vesicle": false }, "ProbGABAAB_EMS": { "init_depleted": false, "minis_single_vesicle": false } } }, "target_simulator": "CORENEURON", "network": "./circuit_config.json", "node_sets_file": "./node_sets.json", "node_set": "my_circuit", "output": { "output_dir": "./reporting", "spikes_file": "spikes.h5" }, "inputs":{ "s1_syns_spikes": { "node_set": "hex_O1", "input_type": "spikes", "delay": 0.0, "duration": 2000.0, "module": "synapse_replay", "spike_file": "./input_spikes/s1_syn_spikes.dat", "source": "hex_O1" } }, "connection_overrides": [ { "name": "init", "source": "hex_O1", "target": "hex_O1", "spont_minis":0.0, "weight": 1.0 }, { "name": "VPM_init", "source": "proj_Thalamocortical_VPM_Source", "target": "hex_O1", "spont_minis": 0.0, "weight": 1.0 }, { "name": "VPM-L5E_meanE", "source": "proj_Thalamocortical_VPM_Source", "target": "Layer5Excitatory", "spont_minis": 1.0, "weight": 0.733 }, { "name": "POm_init", "source": "proj_Thalamocortical_POM_Source", "target": "hex_O1", "spont_minis": 0.0, "weight": 1.0 } ], "reports": { "soma_report": { "type": "compartment", "cells": "my_circuit", "variable_name": "v", "dt": 0.25, "start_time": 0.0, "end_time": 11000.0 } } }My spikes data:
Thanks,
Dhuruva Priyan G M
The text was updated successfully, but these errors were encountered: