replay_trajectory_classification.decoder.ClusterlessDecoder#

class ClusterlessDecoder(environment: Environment = Environment(environment_name='', place_bin_size=2.0, track_graph=None, edge_order=None, edge_spacing=None, is_track_interior=None, position_range=None, infer_track_interior=True, fill_holes=False, dilate=False, bin_count_threshold=0), transition_type: EmpiricalMovement | RandomWalk | RandomWalkDirection1 | RandomWalkDirection2 | Uniform = RandomWalk(environment_name='', movement_var=6.0, movement_mean=0.0, use_diffusion=False), initial_conditions_type: UniformInitialConditions = UniformInitialConditions(), infer_track_interior: bool = True, clusterless_algorithm: str = 'multiunit_likelihood', clusterless_algorithm_params: dict = {'mark_std': 24.0, 'position_std': 6.0})[source]#

Bases: _DecoderBase

Classifies neural population representation of position from multiunit spikes and waveforms.

Parameters:

  • environment (Environment, optional) – The spatial environment to fit

  • transition_type (EmpiricalMovement | RandomWalk | RandomWalkDirection1 | RandomWalkDirection2 | Uniform) – The continuous state transition matrix

  • initial_conditions_type (UniformInitialConditions, optional) – The initial conditions class instance

  • infer_track_interior (bool, optional) – Whether to infer the spatial geometry of track from position

  • clusterless_algorithm (str) – The type of clusterless algorithm. See _ClUSTERLESS_ALGORITHMS for keys

  • clusterless_algorithm_params (dict) – Parameters for the clusterless algorithms.

Methods

convert_results_to_xarray(results, time, ...)

Converts the results dict into a collection of labeled arrays.

copy()

Makes a copy of the classifier

fit(position, multiunits[, is_training])

Fit the spatial grid, initial conditions, place field model, and transition matrices.

fit_environment(position)

Discretize the spatial environment into bins.

fit_initial_conditions()

Set the initial probability of position.

fit_multiunits(position, multiunits[, ...])

get_metadata_routing()

Get metadata routing of this object.

get_params([deep])

Get parameters for this estimator.

load_model([filename])

Load the classifier from a file.

predict(multiunits[, time, ...])

Predict the probability of spatial position and category from the multiunit spikes and waveforms.

project_1D_position_to_2D(results[, ...])

Project the 1D most probable position into the 2D track graph space.

save_model([filename])

Save the classifier to a pickled file.

set_fit_request(*[, is_training, ...])

Request metadata passed to the fit method.

set_params(**params)

Set the parameters of this estimator.

set_predict_request(*[, is_compute_acausal, ...])

Request metadata passed to the predict method.

fit_state_transition

Methods

convert_results_to_xarray

Converts the results dict into a collection of labeled arrays.

copy

Makes a copy of the classifier

fit

Fit the spatial grid, initial conditions, place field model, and transition matrices.

fit_environment

Discretize the spatial environment into bins.

fit_initial_conditions

Set the initial probability of position.

fit_multiunits

fit_state_transition

get_metadata_routing

Get metadata routing of this object.

get_params

Get parameters for this estimator.

load_model

Load the classifier from a file.

predict

Predict the probability of spatial position and category from the multiunit spikes and waveforms.

project_1D_position_to_2D

Project the 1D most probable position into the 2D track graph space.

save_model

Save the classifier to a pickled file.

set_fit_request

Request metadata passed to the fit method.

set_params

Set the parameters of this estimator.

set_predict_request

Request metadata passed to the predict method.
convert_results_to_xarray(results: dict, time: ndarray, data_log_likelihood: float) → Dataset#

Converts the results dict into a collection of labeled arrays.

Parameters:

  • results (dict)

  • time (np.ndarray)

  • data_log_likelihood (float)

Returns:

results

Return type:

xr.Dataset

copy()#

Makes a copy of the classifier

fit(position: ndarray, multiunits: ndarray, is_training: ndarray | None = None)[source]#

Fit the spatial grid, initial conditions, place field model, and transition matrices.

Parameters:

  • position (np.ndarray, shape (n_time, n_position_dims)) – Position of the animal.

  • multiunits (np.ndarray, shape (n_time, n_marks, n_electrodes)) – Array where spikes are indicated by non-Nan values that correspond to the waveform features for each electrode.

  • is_training (None or np.ndarray, shape (n_time), optional) – Boolean array to indicate which data should be included in fitting of place fields, by default None

Return type:

self

fit_environment(position: ndarray) → None#

Discretize the spatial environment into bins. Determine valid track positions.

Parameters:

position (np.ndarray, shape (n_time, n_position_dims)) – Position of the animal in the environment.

fit_initial_conditions()#

Set the initial probability of position.

fit_multiunits(position: ndarray, multiunits: ndarray, is_training: ndarray | None = None)[source]#
Parameters:

  • position (np.ndarray, shape (n_time, n_position_dims))

  • multiunits (np.ndarray, shape (n_time, n_marks, n_electrodes))

  • is_training (None or array_like, shape (n_time,))

get_metadata_routing()#

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:

routing – A MetadataRequest encapsulating routing information.

Return type:

MetadataRequest

get_params(deep=True)#

Get parameters for this estimator.

Parameters:

deep (bool, default=True) – If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:

params – Parameter names mapped to their values.

Return type:

dict

static load_model(filename='model.pkl')#

Load the classifier from a file.

Parameters:

filename (str, optional)

Return type:

classifier instance

predict(multiunits: ndarray, time: ndarray | None = None, is_compute_acausal: bool = True, use_gpu: bool = False) → Dataset[source]#

Predict the probability of spatial position and category from the multiunit spikes and waveforms.

Parameters:

  • multiunits (np.ndarray, shape (n_time, n_marks, n_electrodes)) – Array where spikes are indicated by non-Nan values that correspond to the waveform features for each electrode.

  • time (np.ndarray or None, shape (n_time,), optional) – Label the time axis with these values.

  • is_compute_acausal (bool, optional) – If True, compute the acausal posterior.

  • use_gpu (bool, optional) – Use GPU for the state space part of the model, not the likelihood.

Returns:

results

Return type:

xarray.Dataset

project_1D_position_to_2D(results: Dataset, posterior_type='acausal_posterior') → ndarray#

Project the 1D most probable position into the 2D track graph space.

Only works for single environment.

Parameters:

  • results (xr.Dataset)

  • posterior_type (causal_posterior | acausal_posterior | likelihood)

Returns:

map_position2D

Return type:

np.ndarray

save_model(filename='model.pkl')#

Save the classifier to a pickled file.

Parameters:

filename (str, optional)

set_fit_request(*, is_training: bool | None | str = '$UNCHANGED$', multiunits: bool | None | str = '$UNCHANGED$', position: bool | None | str = '$UNCHANGED$') → ClusterlessDecoder#

Request metadata passed to the fit method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to fit if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to fit.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

  • is_training (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for is_training parameter in fit.

  • multiunits (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for multiunits parameter in fit.

  • position (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for position parameter in fit.

Returns:

self – The updated object.

Return type:

object

set_params(**params)#

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:

**params (dict) – Estimator parameters.

Returns:

self – Estimator instance.

Return type:

estimator instance

set_predict_request(*, is_compute_acausal: bool | None | str = '$UNCHANGED$', multiunits: bool | None | str = '$UNCHANGED$', time: bool | None | str = '$UNCHANGED$', use_gpu: bool | None | str = '$UNCHANGED$') → ClusterlessDecoder#

Request metadata passed to the predict method.

Note that this method is only relevant if enable_metadata_routing=True (see sklearn.set_config()). Please see User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to predict if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to predict.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Note

This method is only relevant if this estimator is used as a sub-estimator of a meta-estimator, e.g. used inside a Pipeline. Otherwise it has no effect.

Parameters:

  • is_compute_acausal (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for is_compute_acausal parameter in predict.

  • multiunits (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for multiunits parameter in predict.

  • time (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for time parameter in predict.

  • use_gpu (str, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED) – Metadata routing for use_gpu parameter in predict.

Returns:

self – The updated object.

Return type:

object