Image thresholding

The goal of this example is to illustrate the use of the function threshold_img together with threshold and two_sided parameters to view certain values in image data.

The threshold parameter can take both positive and negative values. two_sided parameter is complementary to threshold effecting its behavior.

Image without threshold

We will first load the dataset and display the image without manipulation.

from nilearn import datasets
from nilearn.plotting import plot_stat_map

image = datasets.load_sample_motor_activation_image()

plot_param = {
    "display_mode": "ortho",
    "cut_coords": [5, -26, 21],
    "draw_cross": False,
}

plot_stat_map(image, title="image without threshold", **plot_param)

<nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f6ed33a4700>

Image thresholded at 2 when two_sided=True

Now we will use threshold=2 together with two_sided=True to threshold the image. When two_sided=True, we can only use positive values for threshold.

This will set all image values between -2 and 2 to 0.

Note

You can get a similar result by passing threshold directly when plotting. In this case the colorbar will adapt to show which part of the image was masked.

import matplotlib.pyplot as plt

from nilearn.image import threshold_img

thresholded_img = threshold_img(
    image,
    threshold=2,
    cluster_threshold=0,
    two_sided=True,
    copy=True,
    copy_header=True,
)

figure_width = 8

fig, axes = plt.subplots(
    3,
    1,
    figsize=(figure_width, 13),
)

plot_stat_map(
    image, title="image without threshold", axes=axes[0], **plot_param
)

plot_stat_map(
    thresholded_img,
    title="image thresholded at 2 before plotting with two_sided=True",
    axes=axes[1],
    **plot_param,
)

plot_stat_map(
    image,
    title="image thresholded during plotting",
    threshold=2,
    axes=axes[2],
    **plot_param,
)

<nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f6ed4ff2fa0>

Image thresholded at 2 when two_sided=False

Now we will use threshold=2 together with two_sided=False to see the effect.

This will set all image values below 2 to 0.

thresholded_img = threshold_img(
    image,
    threshold=2,
    cluster_threshold=0,
    two_sided=False,
    copy=True,
    copy_header=True,
)


fig, axes = plt.subplots(
    2,
    1,
    figsize=(figure_width, 8),
)


plot_stat_map(
    image, title="image without threshold", axes=axes[0], **plot_param
)

plot_stat_map(
    thresholded_img,
    cmap="Reds",
    title="image thresholded at 2 with two_sided=False",
    axes=axes[1],
    **plot_param,
)

<nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f6ed5163ca0>

Image thresholded at -2 when two_sided=False

Now we will use threshold=-2 together with two_sided=False to see the effect.

This will set all image values above -2 to 0.

thresholded_img = threshold_img(
    image,
    threshold=-2,
    cluster_threshold=0,
    two_sided=False,
    copy=True,
    copy_header=True,
)


fig, axes = plt.subplots(
    2,
    1,
    figsize=(figure_width, 8),
)

plot_stat_map(
    image, title="image without threshold", axes=axes[0], **plot_param
)

plot_stat_map(
    thresholded_img,
    cmap="Blues_r",
    title="image thresholded at -2 with two_sided=False",
    axes=axes[1],
    **plot_param,
)

<nilearn.plotting.displays._slicers.OrthoSlicer object at 0x7f6ed319d310>

Estimated memory usage: 107 MB