Frequently asked questions

Table of Contents

  1. General
    1. Q: How can I find all the classes available in MOOSE?
    2. Q: How can I see the documentation on a class?
    3. Q: How can I see the documentation of a specific field of a class?
    4. Q: What unit system does MOOSE use?
    5. Q: How can I find all elements of a certain type in a model directly under a given element?
    6. Q: How can I find all elements of a certain type at all depths in a model?
    7. Q: How can I find an element with a particular name in a large model?
    8. Q: When trying to plot data from moose Table objects, I see this error “ValueError: x and y must have same first dimension, but have shapes ...”. What is the problem?
  2. Electrophysiology
    1. Q: Does setting length and diameter of a Compartment update its membrane resistance Rm and capacitance Cm fields?
    2. Q: What is the use of the field Ra in a Compartment?
    3. Q: What does HH in HHChannel, HHGate, etc. stand for?
    4. Q: How do I model Hodgkin-Huxley type ion channels?
    5. Q: How do I set up the channel gating equations in HHChannel?
    6. Q: How do I use HHGate.setupAlpha() or HHGate.setupTau() for setting up the lookup tables of an HHGate object?
    7. Q: How do I use explicit mathematical expressions for setting up the gating dynamics of HHGate?
    8. Q: How do I explicitly set the lookup tables for HHGates when using an HHChannel?
    9. Q: How do I model an ion channel so that the channel equations are evaluated exactly instead of using lookup/interpolation tables?
    10. Q: How do I model ion Hodgkin-Huxley-type channels that depend on both voltage and calcium concentration?

General

Q: How can I find all the classes available in MOOSE?

A: In Python, after importing moose, run the command moose.le('/classes')

Q: How can I see the documentation on a class?

A: You can run help(moose.{classname}) and moose.doc('{classname}') in Python.

Here {classname} should be replaced by the name of the class you want to know about.

For example, to learn more about the Compartment class you can run either of the following code snippets:

import moose
moose.doc('Compartment')

or

import moose
help(moose.Compartment)

Q: How can I see the documentation of a specific field of a class?

A: Run moose.doc('{classname}.{fieldname}')

Q: What unit system does MOOSE use?

A: MOOSE is unit agnostic. If all the parameters in your model are in a consistent unit system, results will be in the same system. We recommend that you use SI units throughout.

Experimentalists use different units for different quantities for convenience. These often get mixed up in computational models and can produce unexpected results due to inconsistent units. The physiological unit system uses mVmV for membrane potential, msms for time. Now, if you express membrane capacitance in pFpF, your membrane resistance must be in in GΩG\Omega to get consistent time course in your simulation. This is because only then your membrane time constant (τ=R(GΩ)×C(pF)\tau = R (G\Omega) \times C (pF)) will be in msms. But then your currents would be in pApA (I=V(mV)R(GΩ)I = \frac{V (mV) }{ R (G\Omega)}).

Another source of frustration is that dimensions of neurons are usually expressed in μm\mu m. But conveniently, the specific capacitance of cell membrane is close to 1uF/cm21 uF/cm^{2}. So if you are trying to set the absolute capacitance of a compartment based on this, you must convert its length and diameter to cmcm first and compute the surface area in cm2cm^{2}.

Q: How can I find all elements of a certain type in a model directly under a given element?

Example: You loaded the Kholodenko 2000 model as '/Kholodenko' , then the container for the chemicals and reactions is '/Kholodenko/kinetics/MAPK' . You want to search for elements of type Pool (which represents a pool of molecules or ions) directly under MAPK.

In [19]: moose.wildcardFind('/Kholodenko/kinetics[0]/MAPK[0]/#[TYPE==Pool]')
Out[19]: [<moose.Reac id=521 dataIndex=0 path=/Kholodenko[0]/kinetics[0]/MAPK[0]/Neg_feedback[0]>]

This returns a list of elements of type Pool that are children of /Kholodenko/kinetics/MAPK . But not its grand children or beyond.

Q: How can I find all elements of a certain type at all depths in a model?

A: Use wildcardFind in the form moose.wildcardFind('{root}/##[TYPE={typename}]')

Example: In the Kholodenko 2000 model you want to search for elements of type Pool (which represents a pool of molecules or ions) you loaded the model as "/Kholodenko\", then

In [19]: moose.wildcardFind('/Kholodenko/##[TYPE==Pool]')
Out[19]: [<moose.Reac id=521 dataIndex=0 path=/Kholodenko[0]/kinetics[0]/MAPK[0]/Neg_feedback[0]>]

The ## here tells wildcardFind to do a recursive seach, i.e., starting at {root} , look through its children, grand children, all the way down to the leaves of the element tree.

Q: How can I find an element with a particular name in a large model?

A: Use wildcardFind in the form moose.wildcardFind('{root}/##[FIELD(name)={name}]')

This always returns a list of elements. If the name is unique under root, it will be a list with a single element. If there are multiple elements with the same name, this will return a list containing all those elements. Example: In the Kholodenko 2000 model you want to search for the element "Neg_feedback", and you loaded the model as "/Kholodenko\", then

In [19]: moose.wildcardFind('/Kholodenko/##[FIELD(name)==Neg_feedback]')
Out[19]: [<moose.Reac id=521 dataIndex=0 path=/Kholodenko[0]/kinetics[0]/MAPK[0]/Neg_feedback[0]>]

If you search for "MAPK" :

In [21]: moose.wildcardFind('/Kholodenko/##/MAPK')
Out[21]:
[<moose.Neutral id=489 dataIndex=0 path=/Kholodenko[0]/kinetics[0]/MAPK[0]>,
<moose.Pool id=491 dataIndex=0 path=/Kholodenko[0]/kinetics[0]/MAPK[0]/MAPK[0]>,
<moose.Table2 id=552 dataIndex=0 path=/Kholodenko[0]/data[0]/MAPK[0]>]

Note that the recursive search wildcard (##) must be separated by / (slash). moose.wildcardFind('/Kholodenko/##MAPK') returns an empty list.

Q: When trying to plot data from moose Table objects, I see this error “ValueError: x and y must have same first dimension, but have shapes ...”. What is the problem?

  File "C:\moose-core\tests\core\debug_hhchanf2d.py", line 86, in test_vclamp
	axes[1].plot(t, gktab.vector, label=f'({vstep * 1e3} mV)')
  File "C:\miniforge3\envs\track\Lib\site-packages\matplotlib\axes\_axes.py", line 1721, in plot
	lines = [*self._get_lines(self, *args, data=data, **kwargs)]
        	^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\miniforge3\envs\track\Lib\site-packages\matplotlib\axes\_base.py", line 303, in __call__
	yield from self._plot_args(
           	^^^^^^^^^^^^^^^^
  File "C:\miniforge3\envs\track\Lib\site-packages\matplotlib\axes\_base.py", line 499, in _plot_args
	raise ValueError(f"x and y must have same first dimension, but "
ValueError: x and y must have same first dimension, but have shapes (1050001,) and (2100002,)

A: This may happen due to multiple connections.

In this instance, you are creating the time vector based on the length of one Table, and plotting another. One of the tables (here the second one) was doubly connected, i.e., its requestOut message was passed in two connect calls, like this:

moose.connect(gktab, 'requestOut', channel1, 'Gk')
moose.connect(gktab, 'requestOut', channel2, 'Gk')

You should check the length of the vector attribute of a table against the number of time points expected from total simulation time (simtime in moose.start(simtime) ) and the dt of the table. If the length of vector is a multiple of the expected number, then you have multiple connection issue with that table.

Electrophysiology

Q: Does setting length and diameter of a Compartment update its membrane resistance Rm and capacitance Cm fields?

A: No.

length and diameter are utility fields that can be used for visualization and format conversion. They are not used by MOOSE for computing Rm and Cm. In some other tools the user specifies the specific resistance and specific capacitance of the membrane along with length and diameter of a compartment. And the software computes the total mebrane resistance and capacitance based on the surface area. However, in MOOSE Rm and Cm represent total membrane resistance and capacitance respectively, and are set directly. If you have a cylindrical compartment with given length and diameter, and the specific resistance and capacitance are RM and CM respectively, then you should set:

surface_area = pi * compartment.length * compartment.diameter
compartment.Rm = RM / surface_area
compartment.Cm = CM * surface_area

where pi is imported from any of the modules like math or numpy, or set directly to 3.14159… You must ensure unit consistency between length and specific resistance/capacitance here.

Q: What is the use of the field Ra in a Compartment?

A: Ra represents total axial resistance (also called cytoplasmic resistance) of the compartment.

It is relevant only for multicompartmental models, and represents the resistance faced by current flowing from one compartment to the next.

Q: What does HH in HHChannel, HHGate, etc. stand for?

A: HH stands for Hodgkin and Huxley.

They figured out the dynamics of ion channels that generate action potentials in nerve fibres, and these models use their formulation to implement channel dynamics. See here, here and here to learn more about this.

Q: How do I model Hodgkin-Huxley type ion channels?

A: Create instances of the HHChannel class for fast computations, or HHChannelF class for more flexible computations.

The HHChannel class is designed for faster simulations, and avoids explicitly evaluating Hodgkin-Huxley-type expressions at each simulation step. Instead, it finds the alpha(V) and beta(V) for membrane voltage V using lookup/interpolation tables.

The HHChannelF class is for convenience and accuracy, and allows you to explicitly set the gate expressions as strings. However, it evaluates the formula at each time step, which is slower than table lookup in HHChannel.

Q: How do I set up the channel gating equations in HHChannel?

A: The gating parameters in HHChannel are computed via HHGate objects. They use tables mapping voltage values to gating parameter values to find the gating parameters for a given voltage.

  1. Set HHChannel.Xpower, HHChannel.Ypower corresponding to the number of gating particles in the channel model.
    • For example, for the Na+ channel conductance in the squid giant axon is calculated at voltage VV as GNa(V)=GˉNa×m(V)3×h(V)G_{Na}(V) = \bar{G}_{Na} \times m(V)^{3} \times h(V). So set channel.Xpower= 3 and ~channel.Ypower=1.
    • When these powers are set to positive values, corresponding HHGate objects are created.
  2. Next setup the lookup tables for gating parameters based on their equations using one of the following methods:
    1. Use HHGate.setupAlpha() (for alpha-beta form of the HH equations) or HHGate.setupTau() (for the tau-inf form of the HH equations) function.
    2. (Introduced in 2025-03 in development branch) Assign mathematical expression strings to the fields HHGate.alphaExpr and HHGate.betaExpr for alpha-beta form, or HHGate.tauExpr and HHGate.infExpr for tau-inf form
    3. Explicitly set the HHGate.tableA and HHGate.tableB fields with arrays computed in Python.

Q: How do I use HHGate.setupAlpha() or HHGate.setupTau() for setting up the lookup tables of an HHGate object?

A: Use HHGate.setupAlpha(...) for alpha-beta form of the HH equations or HHGate.setupTau(...) for the tau-inf form of the HH equations.

These functions compute the HH-gate expressions in the generic form:

y(x)=A+BxC+exp((x+D)/F)y(x) = \frac{A + Bx}{C + exp((x+D)/F)}

Thus for each of alpha and beta (or tau and inf depeding on the formulation) there are 5 constant coefficients: A…F, some of which may be 0. The setupAlpha() function (or setupTau()) takes a list of 13 parameters: 5 each for the constants in the expressions for alpha and beta (or tau and inf), number of divisions in the interpolation table (divs), the lower bound of the interpolation table (min), and the upper bound of the interpolation table (max).

Below is an example of setting up Hodgkin-Huxley’s Na channel using setupAlpha. In their formulation

αm=0.1(25−V)exp(25−V10)−1\alpha_{m} = \frac{0.1 (25 - V)}{exp(\frac{25 - V}{10}) - 1}

Thus A=0.1∗25A = 0.1 * 25 , B=−0.1B = -0.1 , C=−1C = -1 , D=−25D = -25 and F=−10F = -10 . Similarly,

βm=4exp(−V18)\beta_{m} = 4 exp(\frac{-V}{18})

so, A=4A = 4, B=0B = 0, C=0C = 0, D=0D = 0 and F=18F = 18 .

We expect the membrane voltage to stay within -110 mV to 50 mV under physiological conditions, and divide this range into in 3000 points for lookup. We also want the lookup to use linear interpolation for voltage values falling between two entries in the table.

import moose

chan = moose.HHChannel('channel')
chan.Xpower = 3  # this will also initialize the HHGate element channel/gateX
chan.Ypower = 1  # this will also initialize the HHGate element channel/gateY
m_gate = moose.element(f'{chan.path}/gateX')
h_gate = moose.element(f'{chan.path}/gateY')
vmin = -110
vmax = 50
vdivs = 3000
m_gate.setupAlpha([
    0.1 * 25.0,                # A_A
    -0.1,                      # A_B
    -1.0,                      # A_C
    -25.0,                     # A_D
    -10.0,                     # A_F
    4.0,                       # B_A
    0.0,                       # B_B
    0.0,                       # B_C
    0.0,                       # B_D
    18.0,                      # B_F
    vdivs,
    vmin,
    vmax])
m_gate.useInterpolation = True   # use linear interpolation instead of direct lookup

# Similarly for h_gate ...

Q: How do I use explicit mathematical expressions for setting up the gating dynamics of HHGate?

(Introduced in 2025-03 in development branch) Assign mathematical expression strings to the fields HHGate.alphaExpr and HHGate.betaExpr for alpha-beta form, or HHGate.tauExpr and HHGate.infExpr for tau-inf form.

In this case you must explicitly set the min, max, and divs fields of each gate. The expressions should use exprtk syntax. Do not forget to convert the expressions to reflect the unit system of your entire model. We recommend adhering to SI units, but the code sample below shows original Hodgkin-Huxley formulation in physiological units. Note that the min and max voltages are in mV.

import moose

chan = moose.HHChannel('channel')
chan.Xpower = 3  # this will also initialize the HHGate element channel/gateX
chan.Ypower = 1  # this will also initialize the HHGate element channel/gateY
m_gate = moose.element(f'{chan.path}/gateX')
h_gate = moose.element(f'{chan.path}/gateY')
m_gate.alphaExpr = '0.1 * (25 - v)/(exp((25 - v)/10) - 1)'
m_gate.betaExpr = '4 * exp(-v/18)'
h_gate.alphaExpr = '0.07 * exp(-v/20)'
h_gate.betaExpr = '1/(exp((30-v)/10) + 1)'
for gate in (m_gate, h_gate):
   gate.useInterpolation = True
   gate.divs = 1000
   gate.min = -30.0
   gate.max = 120.0

Q: How do I explicitly set the lookup tables for HHGates when using an HHChannel?

A: Compute the alpha(V) and beta(V) values for the desired range of voltages.

Then for the target gate, assign alpha(V) series to the tableA field, and alpha(V) + beta(V) to the tableB field. If you have few voltage values, or if you want higher accuracy, set useInterpolation to True. You must explicitly set the min, max, and divs fields of each gate. Note that we are using SI units here and the HH equations have been modified accordingly.

import numpy as np
import moose

chan = moose.HHChannel('channel')
chan.Xpower = 3
chan.Ypower = 1
m_gate = moose.element(f'{chan.path}/gateX')
h_gate = moose.element(f'{chan.path}/gateY')
vmin = -110e-3
vmax = 50e-3
vdivs = 1000
v = np.linspace(vmin, vmax, vdivs + 1) - (-70e-3)
m_alpha = (
    1e3 * (25 - v * 1e3) / (10 * (np.exp((25 - v * 1e3) / 10) - 1))
)
m_beta = 1e3 * 4 * np.exp(-v * 1e3 / 18)
m_gate.min = vmin
m_gate.max = vmax
m_gate.divs = vdivs
m_gate.tableA = m_alpha
m_gate.tableB = m_alpha + m_beta
h_alpha = 1e3 * 0.07 * np.exp(-v / 20e-3)
h_beta = 1e3 * 1 / (np.exp((30e-3 - v) / 10e-3) + 1)
h_gate.min = vmin
h_gate.max = vmax
h_gate.divs = vdivs
h_gate.tableA = h_alpha
h_gate.tableB = h_alpha + h_beta

Q: How do I model an ion channel so that the channel equations are evaluated exactly instead of using lookup/interpolation tables?

A: Use the HHChannelF class for formula based evaluation of gating variables.

HHChannelF is similar to HHChannel but allows you to specify expression strings for gate dynamics. It allows both alpha-beta form and the tau-inf forms. To use alpha-beta form assign the formula for alpha and beta to the fields alpha and beta of the gate in the channel.

Q: How do I model ion Hodgkin-Huxley-type channels that depend on both voltage and calcium concentration?

A: If your channel depends on two independent variables, for example voltage and calcium concentration, use HHChannel2D or HHChannelF2D.