cleo.opto module

Contains opsin models, light sources, and some parameters

class cleo.opto.BansalFourStateOpsin(extra_namespace: dict = NOTHING, spectrum: list[tuple[float, float]] = NOTHING, spectrum_interpolator: Callable = <function log_pchip_interpolator>, extrapolate: bool = False, Gd1: Quantity = 66. * hertz, Gd2: Quantity = 10. * hertz, Gr0: Quantity = 0.333 * hertz, g0: Quantity = 3.2 * nsiemens, phim: Quantity = 1.e+22 * metre ** -2 * second ** -1, k1: Quantity = 0.4 * khertz, k2: Quantity = 120. * hertz, Gf0: Quantity = 18. * hertz, Gb0: Quantity = 8. * hertz, kf: Quantity = 10. * hertz, kb: Quantity = 8. * hertz, gamma: Quantity = 0.05, p: Quantity = 1, q: Quantity = 1, E: Quantity = 0. * volt, *, name: str = NOTHING, save_history: bool = True)

Bases: MarkovOpsin

4-state model from Bansal et al. 2020.

The difference from the PyRhO model is that there is no voltage dependence.

rho_rel is channel density relative to standard model fit; modifying it post-injection allows for heterogeneous opsin expression.

IOPTO_VAR_NAME and V_VAR_NAME are substituted on injection.

Method generated by attrs for class BansalFourStateOpsin.

E: Quantity

Gb0: Quantity

Gd1: Quantity

Gd2: Quantity

Gf0: Quantity

Gr0: Quantity

g0: Quantity

gamma: Quantity

init_syn_vars(opto_syn: Synapses) → None

Initializes appropriate variables in Synapses implementing the model

Also called on reset().

Parameters:

syn (Synapses) – The synapses object implementing this model

k1: Quantity

k2: Quantity

kb: Quantity

kf: Quantity

model: str

Basic Brian model equations string.

Should contain a rho_rel term reflecting relative expression levels. Will likely also contain special NeuronGroup-dependent symbols such as V_VAR_NAME to be replaced on injection in modify_model_and_params_for_ng().

p: Quantity

phim: Quantity

q: Quantity

class cleo.opto.BansalThreeStatePump(extra_namespace: dict = NOTHING, spectrum: list[tuple[float, float]] = NOTHING, spectrum_interpolator: Callable = <function log_pchip_interpolator>, extrapolate: bool = False, Gd: Quantity = 250. * hertz, Gr: Quantity = 50. * hertz, ka: Quantity = msecond ** -1, p: Quantity = 0.7, q: Quantity = 0.1, phim: Quantity = 1.2e+24 * metre ** -2 * second ** -1, E: Quantity = -0.4 * volt, g0: Quantity = 22.34 * nsiemens, a: Quantity = 2.e+08 * metre ** -3 * second ** -1 * amp ** -1 * mole, b: float = 12, vartheta_max: Quantity = 5. * katal / (metre ** 3), kd: Quantity = 16. * mmolar, g_Cl: Quantity = 2.3 * nsiemens, Cl_out: Quantity = 124. * mmolar, Psi0: Quantity = 4.4286 * katal / (metre ** 3), E_Cl0: Quantity = -70. * mvolt, vmin: Quantity = -0.4 * volt, vmax: Quantity = 50. * mvolt, *, name: str = NOTHING, save_history: bool = True)

Bases: MarkovOpsin

3-state model from Bansal et al. 2020. Defaults are for eNpHR3.0.

rho_rel is channel density relative to standard model fit; modifying it post-injection allows for heterogeneous opsin expression.

IOPTO_VAR_NAME and V_VAR_NAME are substituted on injection.

Method generated by attrs for class BansalThreeStatePump.

Cl_out: Quantity

E: Quantity

E_Cl0: Quantity

Gd: Quantity

Gr: Quantity

Psi0: Quantity

a: Quantity

b: float

g0: Quantity

g_Cl: Quantity

init_syn_vars(opto_syn: Synapses) → None

Initializes appropriate variables in Synapses implementing the model

Also called on reset().

Parameters:

syn (Synapses) – The synapses object implementing this model

ka: Quantity

kd: Quantity

model: str

Basic Brian model equations string.

Should contain a rho_rel term reflecting relative expression levels. Will likely also contain special NeuronGroup-dependent symbols such as V_VAR_NAME to be replaced on injection in modify_model_and_params_for_ng().

p: Quantity

phim: Quantity

q: Quantity

vartheta_max: Quantity

vmax: Quantity

Needed to avoid jumps in [Cl_in] for EIF neurons

vmin: Quantity

Needed to avoid jumps in [Cl_in] for EIF neurons

class cleo.opto.FourStateOpsin(extra_namespace: dict = NOTHING, spectrum: list[tuple[float, float]] = NOTHING, spectrum_interpolator: Callable = <function log_pchip_interpolator>, extrapolate: bool = False, g0: Quantity = 114. * nsiemens, gamma: Quantity = 0.00742, phim: Quantity = 2.33e+23 * metre ** -2 * second ** -1, k1: Quantity = 4.15 * khertz, k2: Quantity = 0.868 * khertz, p: Quantity = 0.833, Gf0: Quantity = 37.3 * hertz, kf: Quantity = 58.1 * hertz, Gb0: Quantity = 16.1 * hertz, kb: Quantity = 63. * hertz, q: Quantity = 1.94, Gd1: Quantity = 105. * hertz, Gd2: Quantity = 13.8 * hertz, Gr0: Quantity = 0.33 * hertz, E: Quantity = 0. * volt, v0: Quantity = 43. * mvolt, v1: Quantity = 17.1 * mvolt, *, name: str = NOTHING, save_history: bool = True)

Bases: MarkovOpsin

4-state model from PyRhO (Evans et al. 2016).

rho_rel is channel density relative to standard model fit; modifying it post-injection allows for heterogeneous opsin expression.

IOPTO_VAR_NAME and V_VAR_NAME are substituted on injection.

Defaults are for ChR2.

Method generated by attrs for class FourStateOpsin.

E: Quantity

Gb0: Quantity

Gd1: Quantity

Gd2: Quantity

Gf0: Quantity

Gr0: Quantity

g0: Quantity

gamma: Quantity

init_syn_vars(opto_syn: Synapses) → None

Initializes appropriate variables in Synapses implementing the model

Also called on reset().

Parameters:

syn (Synapses) – The synapses object implementing this model

k1: Quantity

k2: Quantity

kb: Quantity

kf: Quantity

model: str

fv as described in Evans et al., 2016 can give us problems when we divide by zero. Specifically, when v - E = 0, we need to use l’Hopital’s rule to find the limit of fv. This is v1/v0. However…to make things simpler, we see that the (v-E) term cancels out with the same term in the I_post term.

p: Quantity

phim: Quantity

q: Quantity

v0: Quantity

v1: Quantity

class cleo.opto.Opsin(extra_namespace: dict = NOTHING, spectrum: list[tuple[float, float]] = NOTHING, spectrum_interpolator: Callable = <function log_pchip_interpolator>, extrapolate: bool = False, *, name: str = NOTHING, save_history: bool = True)

Bases: LightDependent, SynapseDevice

Base class for opsin model.

Requires that the neuron model has a current term (by default Iopto) which is assumed to be positive (unlike the convention in many opsin modeling papers, where the current is described as negative).

We approximate dynamics under multiple wavelengths using a weighted sum of photon fluxes, where the ε factor indicates the activation relative to the peak-sensitivy wavelength for an equivalent number of photons (see Mager et al, 2018). This weighted sum is an approximation of a nonlinear peak-non-peak wavelength relation; see notebooks/multi_wavelength_model.ipynb for details.

Method generated by attrs for class Opsin.

property action_spectrum

Alias for light_receptor.spectrum

class cleo.opto.ProportionalCurrentOpsin(extra_namespace: dict = NOTHING, spectrum: list[tuple[float, float]] = NOTHING, spectrum_interpolator: Callable = <function log_pchip_interpolator>, extrapolate: bool = False, *, name: str = NOTHING, save_history: bool = True, I_per_Irr: Quantity)

Bases: Opsin

A simple model delivering current proportional to light intensity

Method generated by attrs for class ProportionalCurrentOpsin.

I_per_Irr: Quantity

How much current (in amps or unitless, depending on neuron model) to deliver (must include per unit irradiance).

model: str

Basic Brian model equations string.

Should contain a rho_rel term reflecting relative expression levels. Will likely also contain special NeuronGroup-dependent symbols such as V_VAR_NAME to be replaced on injection in modify_model_and_params_for_ng().

required_vars: list[Tuple[str, Unit]]

Default names of state variables required in the neuron group, along with units, e.g., [(‘Iopto’, amp)].

It is assumed that non-default values can be passed in on injection as a keyword argument [default_name]_var_name=[non_default_name] and that these are found in the model string as [DEFAULT_NAME]_VAR_NAME before replacement.

cleo.opto.chr2_4s() → FourStateOpsin

Returns a 4-state ChR2 model.

Params taken from try.projectpyrho.org’s default 4-state configuration. Action spectrum from Nagel et al., 2003, Fig. 4a, extracted using Plot Digitizer.

Parameters can be changed after initialization but before injection.

cleo.opto.chr2_b4s() → BansalFourStateOpsin

Returns a 4-state ChR2 model.

Params given in Bansal et al., 2020. Action spectrum from Nagel et al., 2003, Fig. 4a, extracted using Plot Digitizer.

Parameters can be changed after initialization but before injection.

cleo.opto.chr2_h134r_4s() → BansalFourStateOpsin

Returns a 4-state ChR2(H134R) model.

Params given in Bansal et al., 2020. Action spectrum is same as for chr2_4s(), but blue-shifted 20 nm (I cannot find it directly in the literature).

Parameters can be changed after initialization but before injection.

cleo.opto.chrimson_4s() → BansalFourStateOpsin

Returns a 4-state Chrimson model.

Params given in Bansal et al., 2020. Action spectrum from Mager et al., 2018, Supp. Fig. 1a, extracted using Plot Digitizer.

Parameters can be changed after initialization but before injection.

cleo.opto.enphr3_3s()

Returns a 3-state model of eNpHR3, a chloride pump.

Params given in Bansal et al., 2020. Action spectrum from Gradinaru et al., 2010, Figure 3F, extracted using Plot Digitizer.

cleo.opto.gtacr2_4s() → BansalFourStateOpsin

Returns a 4-state model of GtACR2, an anion channel.

Params given in Bansal et al., 2020. Action spectra from Govorunova et al., 2015, Fig. 1f, extracted using Plot Digitizer.

Parameters can be changed after initialization but before injection.

cleo.opto.vfchrimson_4s() → BansalFourStateOpsin

Returns a 4-state vf-Chrimson model.

Params given in Bansal et al., 2020. Action spectrum from Mager et al., 2018, Supp. Fig. 1a, extracted using Plot Digitizer.

Parameters can be changed after initialization but before injection.