Decode regions of interest and subsets of Datasets

We can use the methods in nimare.decode.discrete to apply functional characterization analysis to regions of interest or subsets of the Dataset.

import os

import nibabel as nib
import numpy as np
import pandas as pd
from nilearn.plotting import plot_roi, plot_stat_map

import nimare
from nimare.decode import discrete
from nimare.tests.utils import get_test_data_path

Load dataset with abstracts

We’ll load a small dataset composed only of studies in Neurosynth with Angela Laird as a coauthor, for the sake of speed.

dset = nimare.dataset.Dataset.load(
    os.path.join(get_test_data_path(), "neurosynth_laird_studies.pkl.gz")
)
dset.annotations.head(5)

	id	study_id	contrast_id	Neurosynth_TFIDF__001	Neurosynth_TFIDF__01	Neurosynth_TFIDF__05	Neurosynth_TFIDF__10	Neurosynth_TFIDF__10 healthy	Neurosynth_TFIDF__100	Neurosynth_TFIDF__11	Neurosynth_TFIDF__12	Neurosynth_TFIDF__12 healthy	Neurosynth_TFIDF__13	Neurosynth_TFIDF__14	Neurosynth_TFIDF__14 healthy	Neurosynth_TFIDF__15	Neurosynth_TFIDF__15 healthy	Neurosynth_TFIDF__16	Neurosynth_TFIDF__16 healthy	Neurosynth_TFIDF__17	Neurosynth_TFIDF__18	Neurosynth_TFIDF__18 healthy	Neurosynth_TFIDF__19	Neurosynth_TFIDF__20	Neurosynth_TFIDF__20 healthy	Neurosynth_TFIDF__200	Neurosynth_TFIDF__2007	Neurosynth_TFIDF__2014	Neurosynth_TFIDF__2014 2014	Neurosynth_TFIDF__2014 wiley	Neurosynth_TFIDF__21	Neurosynth_TFIDF__22	Neurosynth_TFIDF__23	Neurosynth_TFIDF__24	Neurosynth_TFIDF__25	Neurosynth_TFIDF__26	Neurosynth_TFIDF__27	Neurosynth_TFIDF__28	Neurosynth_TFIDF__29	Neurosynth_TFIDF__30	...	Neurosynth_TFIDF__weak	Neurosynth_TFIDF__weaker	Neurosynth_TFIDF__week	Neurosynth_TFIDF__weeks	Neurosynth_TFIDF__weight	Neurosynth_TFIDF__weighted	Neurosynth_TFIDF__whilst	Neurosynth_TFIDF__white	Neurosynth_TFIDF__white matter	Neurosynth_TFIDF__wide	Neurosynth_TFIDF__widely	Neurosynth_TFIDF__widespread	Neurosynth_TFIDF__wiley	Neurosynth_TFIDF__wiley periodicals	Neurosynth_TFIDF__wise	Neurosynth_TFIDF__wm	Neurosynth_TFIDF__wm task	Neurosynth_TFIDF__women	Neurosynth_TFIDF__word	Neurosynth_TFIDF__word form	Neurosynth_TFIDF__word pairs	Neurosynth_TFIDF__word recognition	Neurosynth_TFIDF__words	Neurosynth_TFIDF__work	Neurosynth_TFIDF__working	Neurosynth_TFIDF__working memory	Neurosynth_TFIDF__world	Neurosynth_TFIDF__worse	Neurosynth_TFIDF__written	Neurosynth_TFIDF__year	Neurosynth_TFIDF__year old	Neurosynth_TFIDF__years	Neurosynth_TFIDF__yield	Neurosynth_TFIDF__yielded	Neurosynth_TFIDF__young	Neurosynth_TFIDF__young adults	Neurosynth_TFIDF__young healthy	Neurosynth_TFIDF__younger	Neurosynth_TFIDF__younger adults	Neurosynth_TFIDF__zone
2225	17029760-1	17029760	1	0.0	0.0	0.0	0.0	0.0	0.0	0.000000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3565	18760263-1	18760263	1	0.0	0.0	0.0	0.0	0.0	0.0	0.000000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3908	19162389-1	19162389	1	0.0	0.0	0.0	0.0	0.0	0.0	0.179402	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4397	19603407-1	19603407	1	0.0	0.0	0.0	0.0	0.0	0.0	0.000000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
5031	20197097-1	20197097	1	0.0	0.0	0.0	0.0	0.0	0.0	0.000000	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	...	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

5 rows × 3172 columns

Create a region of interest

# First we'll make an ROI
arr = np.zeros(dset.masker.mask_img.shape, int)
arr[65:75, 50:60, 50:60] = 1
mask_img = nib.Nifti1Image(arr, dset.masker.mask_img.affine)
plot_roi(mask_img, draw_cross=False)

# Get studies with voxels in the mask
ids = dset.get_studies_by_mask(mask_img)

Decode an ROI image using the Neurosynth method

# Run the decoder
decoder = discrete.NeurosynthDecoder(correction=None)
decoder.fit(dset)
decoded_df = decoder.transform(ids=ids)
decoded_df.sort_values(by="probReverse", ascending=False).head()

	pForward	zForward	probForward	pReverse	zReverse	probReverse
Term
Neurosynth_TFIDF__motor	2.779546e-08	5.554784	0.375000	0.003054	2.962263	1.000000
Neurosynth_TFIDF__presence	2.553341e-04	3.656850	0.545455	0.005002	2.806918	0.916667
Neurosynth_TFIDF__behavioral	3.583888e-04	3.568969	0.383333	0.041038	2.043146	0.869565
Neurosynth_TFIDF__primary	3.583888e-04	3.568969	0.383333	0.041038	2.043146	0.869565
Neurosynth_TFIDF__finally	9.176489e-02	1.686161	0.583333	0.057395	1.900292	0.857143

Decode an ROI image using the BrainMap method

# Run the decoder
decoder = discrete.BrainMapDecoder(correction=None)
decoder.fit(dset)
decoded_df = decoder.transform(ids=ids)
decoded_df.sort_values(by="probReverse", ascending=False).head()

	pForward	zForward	likelihoodForward	pReverse	zReverse	probReverse
Term
Neurosynth_TFIDF__critical	1.0	0.0	40.958333	1.0	0.0	0.025862
Neurosynth_TFIDF__communication	1.0	0.0	40.958333	1.0	0.0	0.025862
Neurosynth_TFIDF__correction	1.0	0.0	40.958333	1.0	0.0	0.025862
Neurosynth_TFIDF__pitch	1.0	0.0	40.958333	1.0	0.0	0.025862
Neurosynth_TFIDF__tasks	1.0	0.0	40.958333	1.0	0.0	0.025862