Source code for allensdk.model.biophysical.utils

# Allen Institute Software License - This software license is the 2-clause BSD
# license plus a third clause that prohibits redistribution for commercial
# purposes without further permission.
#
# Copyright 2015-2017. Allen Institute. All rights reserved.
#
# Redistribution and use in source and binary forms, with or without
# modification, are permitted provided that the following conditions are met:
#
# 1. Redistributions of source code must retain the above copyright notice,
# this list of conditions and the following disclaimer.
#
# 2. Redistributions in binary form must reproduce the above copyright notice,
# this list of conditions and the following disclaimer in the documentation
# and/or other materials provided with the distribution.
#
# 3. Redistributions for commercial purposes are not permitted without the
# Allen Institute's written permission.
# For purposes of this license, commercial purposes is the incorporation of the
# Allen Institute's software into anything for which you will charge fees or
# other compensation. Contact terms@alleninstitute.org for commercial licensing
# opportunities.
#
# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
# ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE
# LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
# CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
# SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
# INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
# CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
# ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
# POSSIBILITY OF SUCH DAMAGE.
#
import logging
import os
from ..biophys_sim.neuron.hoc_utils import HocUtils
from allensdk.core.nwb_data_set import NwbDataSet
from fractions import gcd
from skimage.measure import block_reduce
import scipy.interpolate
import numpy as np
from pkg_resources import resource_filename #@UnresolvedImport

PERISOMATIC_TYPE = "Biophysical - perisomatic"
ALL_ACTIVE_TYPE = "Biophysical - all active"


[docs]def create_utils(description, model_type=None): ''' Factory method to create a Utils subclass. Parameters ---------- description : Config instance used to initialize Utils subclass model_type : string Must be one of [PERISOMATIC_TYPE, ALL_ACTIVE_TYPE]. If none, defaults to PERISOMATIC_TYPE Returns ------- Utils instance ''' if model_type is None: try: model_type = description.data['biophys'][0]['model_type'] except KeyError as e: logging.error("Could not infer model type from description") if model_type == PERISOMATIC_TYPE: return Utils(description) elif model_type == ALL_ACTIVE_TYPE: return AllActiveUtils(description)
[docs]class Utils(HocUtils): '''A helper class for NEURON functionality needed for biophysical simulations. Attributes ---------- h : object The NEURON hoc object. nrn : object The NEURON python object. neuron : module The NEURON module. ''' _log = logging.getLogger(__name__) def __init__(self, description): self.update_default_cell_hoc(description) super(Utils, self).__init__(description) self.stim = None self.stim_curr = None self.simulation_sampling_rate = None self.stimulus_sampling_rate = None self.stim_vec_list = []
[docs] def update_default_cell_hoc(self, description, default_cell_hoc='cell.hoc'): ''' replace the default 'cell.hoc' path in the manifest with 'cell.hoc' packaged within AllenSDK if it does not exist ''' hoc_files = description.data['neuron'][0]['hoc'] try: hfi = hoc_files.index(default_cell_hoc) if not os.path.exists(default_cell_hoc): abspath_ch = resource_filename(__name__, default_cell_hoc) hoc_files[hfi] = abspath_ch if not os.path.exists(abspath_ch): raise IOError("cell.hoc does not exist!") self._log.warning("Using cell.hoc from the following location: %s", abspath_ch) except ValueError as e: pass
[docs] def generate_morphology(self, morph_filename): '''Load a swc-format cell morphology file. Parameters ---------- morph_filename : string Path to swc. ''' h = self.h swc = self.h.Import3d_SWC_read() swc.input(morph_filename) imprt = self.h.Import3d_GUI(swc, 0) h("objref this") imprt.instantiate(h.this) h("soma[0] area(0.5)") for sec in h.allsec(): sec.nseg = 1 + 2 * int(sec.L / 40.0) if sec.name()[:4] == "axon": h.delete_section(sec=sec) h('create axon[2]') for sec in h.axon: sec.L = 30 sec.diam = 1 sec.nseg = 1 + 2 * int(sec.L / 40.0) h.axon[0].connect(h.soma[0], 0.5, 0.0) h.axon[1].connect(h.axon[0], 1.0, 0.0) h.define_shape()
[docs] def load_cell_parameters(self): '''Configure a neuron after the cell morphology has been loaded.''' passive = self.description.data['passive'][0] genome = self.description.data['genome'] conditions = self.description.data['conditions'][0] h = self.h h("access soma") # Set fixed passive properties for sec in h.allsec(): sec.Ra = passive['ra'] sec.insert('pas') for seg in sec: seg.pas.e = passive["e_pas"] for c in passive["cm"]: h('forsec "' + c["section"] + '" { cm = %g }' % c["cm"]) # Insert channels and set parameters for p in genome: if p["section"] == "glob": # global parameter h(p["name"] + " = %g " % p["value"]) else: if p["mechanism"] != "": h('forsec "' + p["section"] + '" { insert ' + p["mechanism"] + ' }') h('forsec "' + p["section"] + '" { ' + p["name"] + ' = %g }' % p["value"]) # Set reversal potentials for erev in conditions['erev']: h('forsec "' + erev["section"] + '" { ek = %g }' % erev["ek"]) h('forsec "' + erev["section"] + '" { ena = %g }' % erev["ena"])
[docs] def setup_iclamp(self, stimulus_path, sweep=0): '''Assign a current waveform as input stimulus. Parameters ---------- stimulus_path : string NWB file name ''' self.stim = self.h.IClamp(self.h.soma[0](0.5)) self.stim.amp = 0 self.stim.delay = 0 # just set to be really big; doesn't need to match the waveform self.stim.dur = 1e12 self.read_stimulus(stimulus_path, sweep=sweep) # NEURON's dt is in milliseconds simulation_dt = 1.0e3 / self.simulation_sampling_rate stimulus_dt = 1.0e3 / self.stimulus_sampling_rate self._log.debug("Using simulation dt %f, stimulus dt %f", simulation_dt, stimulus_dt) self.h.dt = simulation_dt stim_vec = self.h.Vector(self.stim_curr) stim_vec.play(self.stim._ref_amp, stimulus_dt) stimulus_stop_index = len(self.stim_curr) - 1 self.h.tstop = stimulus_stop_index * stimulus_dt self.stim_vec_list.append(stim_vec)
[docs] def read_stimulus(self, stimulus_path, sweep=0): '''Load current values for a specific experiment sweep and setup simulation and stimulus sampling rates. NOTE: NEURON only allows simulation timestamps of multiples of 40KHz. To avoid aliasing, we set the simulation sampling rate to the least common multiple of the stimulus sampling rate and 40KHz. Parameters ---------- stimulus path : string NWB file name sweep : integer, optional sweep index ''' Utils._log.info( "reading stimulus path: %s, sweep %s", stimulus_path, sweep) stimulus_data = NwbDataSet(stimulus_path) sweep_data = stimulus_data.get_sweep(sweep) # convert to nA for NEURON self.stim_curr = sweep_data['stimulus'] * 1.0e9 # convert from Hz hz = int(sweep_data['sampling_rate']) neuron_hz = Utils.nearest_neuron_sampling_rate(hz) self.simulation_sampling_rate = neuron_hz self.stimulus_sampling_rate = hz if hz != neuron_hz: Utils._log.debug("changing sampling rate from %d to %d to avoid NEURON aliasing", hz, neuron_hz)
[docs] def record_values(self): '''Set up output voltage recording.''' vec = {"v": self.h.Vector(), "t": self.h.Vector()} vec["v"].record(self.h.soma[0](0.5)._ref_v) vec["t"].record(self.h._ref_t) return vec
[docs] def get_recorded_data(self, vec): '''Extract recorded voltages and timestamps given the recorded Vector instance. If self.stimulus_sampling_rate is smaller than self.simulation_sampling_rate, resample to self.stimulus_sampling_rate. Parameters ---------- vec : neuron.Vector constructed by self.record_values Returns ------- dict with two keys: 'v' = numpy.ndarray with voltages, 't' = numpy.ndarray with timestamps ''' junction_potential = self.description.data['fitting'][0]['junction_potential'] v = np.array(vec["v"]) t = np.array(vec["t"]) if self.stimulus_sampling_rate < self.simulation_sampling_rate: factor = self.simulation_sampling_rate / self.stimulus_sampling_rate Utils._log.debug("subsampling recorded traces by %dX", factor) v = block_reduce(v, (factor,), np.mean)[:len(self.stim_curr)] t = block_reduce(t, (factor,), np.min)[:len(self.stim_curr)] mV = 1.0e-3 v = (v - junction_potential) * mV return { "v": v, "t": t }
[docs] @staticmethod def nearest_neuron_sampling_rate(hz, target_hz=40000): div = gcd(hz, target_hz) new_hz = hz * target_hz / div return new_hz
[docs]class AllActiveUtils(Utils):
[docs] def generate_morphology(self, morph_filename): '''Load a neurolucida or swc-format cell morphology file. Parameters ---------- morph_filename : string Path to morphology. ''' morph_basename = os.path.basename(morph_filename) morph_extension = morph_basename.split('.')[-1] if morph_extension.lower() == 'swc': morph = self.h.Import3d_SWC_read() elif morph_extension.lower() == 'asc': morph = self.h.Import3d_Neurolucida3() else: raise Exception("Unknown filetype: %s" % morph_extension) morph.input(morph_filename) imprt = self.h.Import3d_GUI(morph, 0) self.h("objref this") imprt.instantiate(self.h.this) for sec in self.h.allsec(): sec.nseg = 1 + 2 * int(sec.L / 40.0) self.h("soma[0] area(0.5)") axon_diams = [self.h.axon[0].diam, self.h.axon[0].diam] self.h.distance(sec=self.h.soma[0]) for sec in self.h.allsec(): if sec.name()[:4] == "axon": if self.h.distance(0.5, sec=sec) > 60: axon_diams[1] = sec.diam break for sec in self.h.allsec(): if sec.name()[:4] == "axon": self.h.delete_section(sec=sec) self.h('create axon[2]') for index, sec in enumerate(self.h.axon): sec.L = 30 sec.diam = axon_diams[index] for sec in self.h.allsec(): sec.nseg = 1 + 2 * int(sec.L / 40.0) self.h.axon[0].connect(self.h.soma[0], 1.0, 0.0) self.h.axon[1].connect(self.h.axon[0], 1.0, 0.0) # make sure diam reflects 3d points self.h.area(.5, sec=self.h.soma[0])
[docs] def load_cell_parameters(self): '''Configure a neuron after the cell morphology has been loaded.''' passive = self.description.data['passive'][0] genome = self.description.data['genome'] conditions = self.description.data['conditions'][0] h = self.h h("access soma") # Set fixed passive properties for sec in h.allsec(): sec.Ra = passive['ra'] sec.insert('pas') # for seg in sec: # seg.pas.e = passive["e_pas"] # for c in passive["cm"]: # h('forsec "' + c["section"] + '" { cm = %g }' % c["cm"]) # Insert channels and set parameters for p in genome: section_array = p["section"] mechanism = p["mechanism"] param_name = p["name"] param_value = float(p["value"]) if section_array == "glob": # global parameter h(p["name"] + " = %g " % p["value"]) else: if hasattr(h, section_array): if mechanism != "": print('Adding mechanism %s to %s' % (mechanism, section_array)) for section in getattr(h, section_array): if self.h.ismembrane(str(mechanism), sec=section) != 1: section.insert(mechanism) print('Setting %s to %.6g in %s' % (param_name, param_value, section_array)) for section in getattr(h, section_array): setattr(section, param_name, param_value) # Set reversal potentials for erev in conditions['erev']: erev_section_array = erev["section"] ek = float(erev["ek"]) ena = float(erev["ena"]) print('Setting ek to %.6g and ena to %.6g in %s' % (ek, ena, erev_section_array)) if hasattr(h, erev_section_array): for section in getattr(h, erev_section_array): if self.h.ismembrane("k_ion", sec=section) == 1: setattr(section, 'ek', ek) if self.h.ismembrane("na_ion", sec=section) == 1: setattr(section, 'ena', ena) else: print("Warning: can't set erev for %s, " "section array doesn't exist" % erev_section_array) self.h.v_init = conditions['v_init'] self.h.celsius = conditions['celsius']