nimare.meta.cbmr.CBMREstimator

class CBMREstimator(group_categories=None, moderators=None, mask=None, spline_spacing=10, model=<class 'nimare.meta.models.PoissonEstimator'>, penalty=False, n_iter=2000, lr=1, lr_decay=0.999, tol=1e-09, device='cpu', random_state=None, **kwargs)[source]

Bases: Estimator

Coordinate-based meta-regression with a spatial model.

Warning

Support for Dataset inputs is deprecated and will be removed in a future release. Prefer Studyset.

Added in version 0.1.0.

Parameters:

  • group_categories (str or obj:list or obj:None, optional) – CBMR allows a collection to be categorized into multiple groups according to one or more group categories. Default is one-group CBMR.

  • moderators (str or obj:list or obj:None, optional) – CBMR can accommodate experiment-level moderators (e.g. sample size, year of publication). Default is CBMR without experiment-level moderators.

  • model (subclass of GeneralLinearModelEstimator, optional) –

    Stochastic model class used by CBMR. Available options are:

    • PoissonEstimator (default): the most efficient and widely used option, but slightly less accurate because it approximates low-rate binomial data, cannot account for over-dispersion in foci counts, and may underestimate standard errors.

    • NegativeBinomialEstimator: slower and sometimes less stable, but slightly more accurate. This model allows anticipated excess variance relative to Poisson via a group-wise overdispersion parameter shared by all experiments and voxels.

    • ClusteredNegativeBinomialEstimator: a random-effects Poisson variant that models experiment-level latent characteristics shared across the brain for a given experiment.

  • penalty (bool, optional) – Currently, the only available option is Firth-type penalty, which penalizes the likelihood function by Jeffreys’ invariant prior and encourages convergence.

  • spline_spacing (int, optional) – Spatial structure of foci counts is parameterized by the coefficients of cubic B-spline bases in CBMR. Spatial smoothness in CBMR is determined by spline spacing, which is shared across the x, y, and z dimensions. Default is 10.

  • n_iter (int, optional) – Number of iterations allowed in the log-likelihood optimization. Default is 2000.

  • lr (float, optional) – Learning rate in optimization of log-likelihood function. Default is 1.

  • lr_decay (float, optional) – Multiplicative factor of learning rate decay. Default is 0.999.

  • tol (float, optional) – Stopping criterion based on the change in log-likelihood between two consecutive iterations. Default is 1e-9.

  • device (string, optional) – Device type (‘cpu’ or ‘cuda’) representing where operations will be allocated. Default is ‘cpu’.

  • random_state (int, optional) – Random seed used for torch-based weight initialization. Default is None.

  • **kwargs – Keyword arguments. Arguments for the Estimator can be assigned here, Another optional argument is mask.

Variables:

  • masker (NiftiMasker or similar) – Masker object.

  • inputs (dict) – Inputs to the Estimator. For CBMR estimators, this includes the following keys: coordinates, mask_img (brain mask image), id (experiment ids), ids_by_group (experiment ids categorized by groups), moderators_by_group (experiment-level moderators categorized by groups, if present), coef_spline_bases (spatial matrix of cubic B-spline coefficients in x, y, and z), foci_per_voxel (voxelwise sum of foci counts across experiments, categorized by groups), foci_per_experiment (experiment-wise sum of foci counts across space, categorized by groups).

Notes

Follow-up inference is exposed through CBMRResult and CBMRInference.

Methods

fit(dataset[, drop_invalid])

Fit Estimator to a collection.

get_params([deep])

Get parameters for this estimator.

load(filename[, compressed])

Load a pickled class instance from file.

save(filename[, compress])

Pickle the class instance to the provided file.

set_params(**params)

Set the parameters of this estimator.
fit(dataset, drop_invalid=True)[source]

Fit Estimator to a collection.

Parameters:

  • dataset (Studyset or Dataset) – Collection object to analyze.

  • drop_invalid (bool, optional) – Whether to automatically ignore any studies without the required data or not. Default is True.

Returns:

Result of Estimator fitting. Subclasses may return a MetaResult subclass.

Return type:

MetaResult

Variables:

  • inputs (dict) – Inputs used in _fit.

  • warning:: (..) – Support for Dataset inputs is deprecated and will be removed in a future release. Prefer Studyset.

  • and (The fit method is a light wrapper that runs input validation)

  • individual (preprocessing before fitting the actual model. Estimators')

  • should ("fitting" methods are implemented as _fit, although users)

  • fit. (call)

get_params(deep=True)[source]

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

classmethod load(filename, compressed=True)[source]

Load a pickled class instance from file.

Parameters:

  • filename (str) – Name of file containing object.

  • compressed (bool, default=True) – If True, the file is assumed to be compressed and gzip will be used to load it. Otherwise, it will assume that the file is not compressed. Default = True.

Returns:

obj – Loaded class object.

Return type:

class object

save(filename, compress=True)[source]

Pickle the class instance to the provided file.

Parameters:

  • filename (str) – File to which object will be saved.

  • compress (bool, optional) – If True, the file will be compressed with gzip. Otherwise, the uncompressed version will be saved. Default = True.

set_params(**params)[source]

Set the parameters of this estimator.

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

Return type:

self