nimare.meta.cbma.ale.BalancedALESubtraction

class BalancedALESubtraction(kernel_transformer=<class 'nimare.meta.kernel.ALEKernel'>, target_n=None, n_subsamples=2500, difference_iterations=1000, n_iters=1000, voxel_thresh=0.001, null_method='random-foci', mask_coverage='gm', alpha=0.05, memory=Memory(location=None), memory_level=0, n_cores=1, random_state=None, **kwargs)

Bases: PairwiseCBMAEstimator

Balanced ALE subtraction with matched-size subsampling.

“A balanced ALE subtraction using matched-size ” “subsampling within groups, averaged balanced ALE differences, and Monte Carlo null ” “extrema from balanced resamples. Frahm et al.[1]”

Parameters:

• null_method ({"random-foci", "label-permutation"}, optional) –

  Method used to generate the null distribution of balanced differences.

  "random-foci" (default) generates null MA maps by placing each study’s foci randomly within the mask_coverage while preserving per-study sample-size and focus-count metadata. Because balanced subsampling breaks label exchangeability, this is the statistically coherent null for balanced subtractions.

  "label-permutation" pools the prior-masked MA maps from both groups, randomly reassigns study labels (preserving group sizes), and computes the balanced difference. This directly tests group-label exchangeability at the cost of assuming the spatial structure of each study is fixed.

• mask_coverage ({"gm", "brain"}, optional) – Voxel set used both for restricting the balanced-difference computation and (when null_method="random-foci") for drawing random foci. "gm" uses mask-image intensity > 0.1 (the ICBM 10% GM probability map); "brain" uses all non-zero voxels. Default is "gm".

• alpha (float, optional) – Family-wise error rate for the per-group cluster threshold used inside _probabilistic_map and for the balanced-subtraction extrema percentiles in _fit. Default is 0.05.

Methods

correct_fwe_montecarlo(result[, ...])	Perform FWE correction using the max-value permutation method.
fit(dataset1, dataset2[, drop_invalid, ...])	Fit Estimator to two collections.
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.

correct_fwe_montecarlo(result, voxel_thresh=0.001, n_iters=5000, n_cores=1, vfwe_only=False)

Perform FWE correction using the max-value permutation method.

Only call this method from within a Corrector.

Changed in version 0.0.13: Change cluster neighborhood from faces+edges to faces, to match Nilearn.

Changed in version 0.0.12:

• Fix the vfwe_only option.

Changed in version 0.0.11:

• Rename *_level-cluster maps to *_desc-size_level-cluster.

• Add new *_desc-mass_level-cluster maps that use cluster mass-based inference.

Parameters:

• result (MetaResult) – Result object from a CBMA meta-analysis.

• voxel_thresh (float, default=0.001) – Cluster-defining p-value threshold. Default is 0.001.

• n_iters (int, default=5000) – Number of iterations to build the voxel-level, cluster-size, and cluster-mass FWE null distributions. Default is 5000.

• n_cores (int, default=1) – Number of cores to use for parallelization. If <=0, defaults to using all available cores. Default is 1.

• vfwe_only (bool, default=False) – If True, only calculate the voxel-level FWE-corrected maps. Voxel-level correction can be performed very quickly if the Estimator’s null_method was “montecarlo”. Default is False.

Returns:

• images (dict) – Dictionary of 1D arrays corresponding to masked images generated by the correction procedure. The following arrays are generated by this method:

  • logp_desc-size_level-cluster: Cluster-level FWE-corrected -log10(p) map based on cluster size. This was previously simply called “logp_level-cluster”. This array is not generated if vfwe_only is True.

  • logp_desc-mass_level-cluster: Cluster-level FWE-corrected -log10(p) map based on cluster mass. According to Bullmore et al.[2] and Zhang et al.[3], cluster mass-based inference is more powerful than cluster size. This array is not generated if vfwe_only is True.

  • logp_level-voxel: Voxel-level FWE-corrected -log10(p) map. Voxel-level correction is generally more conservative than cluster-level correction, so it is only recommended for very large meta-analyses (i.e., hundreds of studies), per Eickhoff et al.[4].

• description_ (str) – A text description of the correction procedure.

Notes

If vfwe_only is False, this method adds three new keys to the null_distributions_ attribute:

• values_level-voxel_corr-fwe_method-montecarlo: The maximum summary statistic value from each Monte Carlo iteration. An array of shape (n_iters,).

• values_desc-size_level-cluster_corr-fwe_method-montecarlo: The maximum cluster size from each Monte Carlo iteration. An array of shape (n_iters,).

• values_desc-mass_level-cluster_corr-fwe_method-montecarlo: The maximum cluster mass from each Monte Carlo iteration. An array of shape (n_iters,).

See also

nimare.correct.FWECorrector

The Corrector from which to call this method.

References

[1]

Lennart Frahm, Monimu, and Felix Hoffstaedter. Juaml/jale. URL: https://github.com/juaml/JALE.

[2]

Edward T Bullmore, John Suckling, Stephan Overmeyer, Sophia Rabe-Hesketh, Eric Taylor, and Michael J Brammer. Global, voxel, and cluster tests, by theory and permutation, for a difference between two groups of structural mr images of the brain. IEEE transactions on medical imaging, 18(1):32–42, 1999. URL: https://doi.org/10.1109/42.750253, doi:10.1109/42.750253.

[3]

Hui Zhang, Thomas E Nichols, and Timothy D Johnson. Cluster mass inference via random field theory. Neuroimage, 44(1):51–61, 2009. URL: https://doi.org/10.1016/j.neuroimage.2008.08.017, doi:10.1016/j.neuroimage.2008.08.017.

[4]

Simon B Eickhoff, Thomas E Nichols, Angela R Laird, Felix Hoffstaedter, Katrin Amunts, Peter T Fox, Danilo Bzdok, and Claudia R Eickhoff. Behavior, sensitivity, and power of activation likelihood estimation characterized by massive empirical simulation. Neuroimage, 137:70–85, 2016. URL: https://doi.org/10.1016/j.neuroimage.2016.04.072, doi:10.1016/j.neuroimage.2016.04.072.

Examples

>>> meta = MKDADensity()
>>> result = meta.fit(dset)
>>> corrector = FWECorrector(method='montecarlo', voxel_thresh=0.01,
                             n_iters=5, n_cores=1)
>>> cresult = corrector.transform(result)

fit(dataset1, dataset2, drop_invalid=True, ma_maps1=None, ma_maps2=None, inference_map1=None, inference_map2=None)

Fit Estimator to two collections.

Parameters:

• dataset1/dataset2 (Studyset or Dataset) – Collection objects to analyze.

• ma_maps1/ma_maps2 (scipy sparse matrix or sparse array, optional) – Precomputed study-wise MA maps aligned to dataset1 and dataset2, respectively. When provided, the estimator will reuse these maps instead of recomputing them from coordinates. These are typically 2D study-by-masked-voxel sparse matrices.

• inference_map1/inference_map2 (array_like or Niimg-like, optional) – Optional directional inference maps aligned to the common masked voxel space. Positive pairwise effects are only evaluated where inference_map1 > 0, and negative pairwise effects are only evaluated where inference_map2 > 0.

Returns:

• MetaResult – Results of Estimator fitting.

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

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

• preprocessing before fitting the actual model. Estimators’ individual

• ”fitting” methods are implemented as _fit, although users should

• call fit.

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

classmethod load(filename, compressed=True)

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)

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)

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