Connectivity Matrices, ROI Intersections and Density MapsΒΆ

This example is meant to be an introduction to some of the streamline tools available in DIPY. Some of the functions covered in this example are target, connectivity_matrix and density_map. target allows one to filter streamlines that either pass through or do not pass through some region of the brain, connectivity_matrix groups and counts streamlines based on where in the brain they begin and end, and finally, density map counts the number of streamlines that pass though every voxel of some image.

To get started we’ll need to have a set of streamlines to work with. We’ll use EuDX along with the CsaOdfModel to make some streamlines. Let’s import the modules and download the data we’ll be using.

from dipy.tracking.eudx import EuDX
from dipy.reconst import peaks, shm
from dipy.tracking import utils
from dipy.tracking.streamline import Streamlines

from dipy.data import read_stanford_labels, fetch_stanford_t1, read_stanford_t1

hardi_img, gtab, labels_img = read_stanford_labels()
data = hardi_img.get_data()
labels = labels_img.get_data()

fetch_stanford_t1()
t1 = read_stanford_t1()
t1_data = t1.get_data()

We’ve loaded an image called labels_img which is a map of tissue types such that every integer value in the array labels represents an anatomical structure or tissue type [1]. For this example, the image was created so that white matter voxels have values of either 1 or 2. We’ll use peaks_from_model to apply the CsaOdfModel to each white matter voxel and estimate fiber orientations which we can use for tracking.

white_matter = (labels == 1) | (labels == 2)
csamodel = shm.CsaOdfModel(gtab, 6)
csapeaks = peaks.peaks_from_model(model=csamodel,
                                  data=data,
                                  sphere=peaks.default_sphere,
                                  relative_peak_threshold=.8,
                                  min_separation_angle=45,
                                  mask=white_matter)

Now we can use EuDX to track all of the white matter. To keep things reasonably fast we use density=2 which will result in 8 seeds per voxel. We’ll set a_low (the parameter which determines the threshold of FA/QA under which tracking stops) to be very low because we’ve already applied a white matter mask.

seeds = utils.seeds_from_mask(white_matter, density=2)
streamline_generator = EuDX(csapeaks.peak_values, csapeaks.peak_indices,
                            odf_vertices=peaks.default_sphere.vertices,
                            a_low=.05, step_sz=.5, seeds=seeds)
affine = streamline_generator.affine

streamlines = Streamlines(streamline_generator, buffer_size=512)

The first of the tracking utilities we’ll cover here is target. This function takes a set of streamlines and a region of interest (ROI) and returns only those streamlines that pass though the ROI. The ROI should be an array such that the voxels that belong to the ROI are True and all other voxels are False (this type of binary array is sometimes called a mask). This function can also exclude all the streamlines that pass though an ROI by setting the include flag to False. In this example we’ll target the streamlines of the corpus callosum. Our labels array has a sagittal slice of the corpus callosum identified by the label value 2. We’ll create an ROI mask from that label and create two sets of streamlines, those that intersect with the ROI and those that don’t.

cc_slice = labels == 2
cc_streamlines = utils.target(streamlines, cc_slice, affine=affine)
cc_streamlines = Streamlines(cc_streamlines)

other_streamlines = utils.target(streamlines, cc_slice, affine=affine,
                                 include=False)
other_streamlines = Streamlines(other_streamlines)
assert len(other_streamlines) + len(cc_streamlines) == len(streamlines)

We can use some of DIPY‘s visualization tools to display the ROI we targeted above and all the streamlines that pass though that ROI. The ROI is the yellow region near the center of the axial image.

from dipy.viz import window, actor
from dipy.viz.colormap import line_colors

# Enables/disables interactive visualization
interactive = False

# Make display objects
color = line_colors(cc_streamlines)
cc_streamlines_actor = actor.line(cc_streamlines, line_colors(cc_streamlines))
cc_ROI_actor = actor.contour_from_roi(cc_slice, color=(1., 1., 0.),
                                      opacity=0.5)

vol_actor = actor.slicer(t1_data)

vol_actor.display(x=40)
vol_actor2 = vol_actor.copy()
vol_actor2.display(z=35)

# Add display objects to canvas
r = window.Renderer()
r.add(vol_actor)
r.add(vol_actor2)
r.add(cc_streamlines_actor)
r.add(cc_ROI_actor)

# Save figures
window.record(r, n_frames=1, out_path='corpuscallosum_axial.png',
              size=(800, 800))
if interactive:
    window.show(r)
r.set_camera(position=[-1, 0, 0], focal_point=[0, 0, 0], view_up=[0, 0, 1])
window.record(r, n_frames=1, out_path='corpuscallosum_sagittal.png',
              size=(800, 800))
if interactive:
    window.show(r)
../_images/corpuscallosum_axial.png

Corpus Callosum Axial

../_images/corpuscallosum_sagittal.png

Corpus Callosum Sagittal

Once we’ve targeted on the corpus callosum ROI, we might want to find out which regions of the brain are connected by these streamlines. To do this we can use the connectivity_matrix function. This function takes a set of streamlines and an array of labels as arguments. It returns the number of streamlines that start and end at each pair of labels and it can return the streamlines grouped by their endpoints. Notice that this function only considers the endpoints of each streamline.

M, grouping = utils.connectivity_matrix(cc_streamlines, labels, affine=affine,
                                        return_mapping=True,
                                        mapping_as_streamlines=True)
M[:3, :] = 0
M[:, :3] = 0

We’ve set return_mapping and mapping_as_streamlines to True so that connectivity_matrix returns all the streamlines in cc_streamlines grouped by their endpoint.

Because we’re typically only interested in connections between gray matter regions, and because the label 0 represents background and the labels 1 and 2 represent white matter, we discard the first three rows and columns of the connectivity matrix.

We can now display this matrix using matplotlib, we display it using a log scale to make small values in the matrix easier to see.

import numpy as np
import matplotlib.pyplot as plt
plt.imshow(np.log1p(M), interpolation='nearest')
plt.savefig("connectivity.png")
../_images/connectivity.png

Connectivity of Corpus Callosum

In our example track there are more streamlines connecting regions 11 and 54 than any other pair of regions. These labels represent the left and right superior frontal gyrus respectively. These two regions are large, close together, have lots of corpus callosum fibers and are easy to track so this result should not be a surprise to anyone.

However, the interpretation of streamline counts can be tricky. The relationship between the underlying biology and the streamline counts will depend on several factors, including how the tracking was done, and the correct way to interpret these kinds of connectivity matrices is still an open question in the diffusion imaging literature.

The next function we’ll demonstrate is density_map. This function allows one to represent the spatial distribution of a track by counting the density of streamlines in each voxel. For example, let’s take the track connecting the left and right superior frontal gyrus.

lr_superiorfrontal_track = grouping[11, 54]
shape = labels.shape
dm = utils.density_map(lr_superiorfrontal_track, shape, affine=affine)

Let’s save this density map and the streamlines so that they can be visualized together. In order to save the streamlines in a ”.trk” file we’ll need to move them to “trackvis space”, or the representation of streamlines specified by the trackvis Track File format.

To do that, we will use tools available in nibabel)

import nibabel as nib
from dipy.io.streamline import save_trk

# Save density map
dm_img = nib.Nifti1Image(dm.astype("int16"), hardi_img.affine)
dm_img.to_filename("lr-superiorfrontal-dm.nii.gz")

# Move streamlines to "trackvis space"
voxel_size = labels_img.header.get_zooms()
trackvis_point_space = utils.affine_for_trackvis(voxel_size)
# lr_sf_trk = utils.move_streamlines(lr_superiorfrontal_track,
#                                   trackvis_point_space, input_space=affine)

lr_sf_trk = Streamlines(lr_superiorfrontal_track)

# Save streamlines
save_trk("lr-superiorfrontal.trk", lr_sf_trk, shape=shape, vox_size=voxel_size, affine=affine)

Let’s take a moment here to consider the representation of streamlines used in DIPY. Streamlines are a path though the 3D space of an image represented by a set of points. For these points to have a meaningful interpretation, these points must be given in a known coordinate system. The affine attribute of the streamline_generator object specifies the coordinate system of the points with respect to the voxel indices of the input data. trackvis_point_space specifies the trackvis coordinate system with respect to the same indices. The move_streamlines function returns a new set of streamlines from an existing set of streamlines in the target space. The target space and the input space must be specified as affine transformations with respect to the same reference [2]. If no input space is given, the input space will be the same as the current representation of the streamlines, in other words the input space is assumed to be np.eye(4), the 4-by-4 identity matrix.

All of the functions above that allow streamlines to interact with volumes take an affine argument. This argument allows these functions to work with streamlines regardless of their coordinate system. For example even though we moved our streamlines to “trackvis space”, we can still compute the density map as long as we specify the right coordinate system.

dm_trackvis = utils.density_map(lr_sf_trk, shape, affine=np.eye(4))
assert np.all(dm == dm_trackvis)

This means that streamlines can interact with any image volume, for example a high resolution structural image, as long as one can register that image to the diffusion images and calculate the coordinate system with respect to that image.

Footnotes

[1]The image aparc-reduced.nii.gz, which we load as labels_img, is a modified version of label map aparc+aseg.mgz created by FreeSurfer. The corpus callosum region is a combination of the FreeSurfer labels 251-255. The remaining FreeSurfer labels were re-mapped and reduced so that they lie between 0 and 88. To see the FreeSurfer region, label and name, represented by each value see label_info.txt in ~/.dipy/stanford_hardi.
[2]An affine transformation is a mapping between two coordinate systems that can represent scaling, rotation, sheer, translation and reflection. Affine transformations are often represented using a 4x4 matrix where the last row of the matrix is [0, 0, 0, 1].

Example source code

You can download the full source code of this example. This same script is also included in the dipy source distribution under the doc/examples/ directory.