Atlas Annotation

Over the years the Allen Brain institute has constantly improved and updated their brain region annotation atlases. Unfortunately the old annotation atlases are not always aligned with the new ones. For example, the CCFv2 annotations and the Nissl volume are not compatible with the CCFv3 annotation and the corresponding average brain volume. This package proposes a number of methods for deforming the Nissl volume and the CCFv2 annotations in order to re-align them to CCFv3.

• Installation
  • Python Version and Environment
  • Install "Atlas Annotation"
• Data
  • Remote Storage Access
  • Get the Data
• Examples
• Notebooks, Widgets, and Experiments
• Funding & Acknowledgment

Installation

Python Version and Environment

The currently supported python versions are 3.7, 3.8, and 3.9.

If you are part of the Blue Brain Project and are working on the BB5 you can find the correct python version in the archive modules:

• python3.7: archive/2020-02 - archive/2020-12
• python3.8: archive/2021-01 - archive/2021-12
• python3.9: archive/2022-01 - unstable. Here's an example of a set of commands that will set up your environment on the BB5:

module purge
module load archive/2020-12
module load python
python -m venv venv
. ./venv/bin/activate
python --version

We also recommend that you make sure that pip is up-to-date and that the packages wheel and setuptools are installed:

pip install --upgrade pip wheel setuptools

Install "Atlas Annotation"

In order to access the data and the example scripts a local clone of this repository is required. Run these commands to get it:

git clone https://github.com/BlueBrain/atlas-annotation
cd atlas-annotation

The "Atlas Interpolation" package can now be installed directly from the clone we just created:

pip install '.[data, interactive]'

Data

The data for this project is managed by the DVC tool and all related files are located in the data directory. The DVC tool has already been installed together with the "Atlas Interpolation" package. Every time you need to run a DVC command (dvc ...) make sure to change to the data directory first (cd data).

Remote Storage Access

We have already prepared all the data, but it is located on a remote storage that is only accessible to people within the Blue Brain Project who have access permissions to project proj101. If you're unsure you can test your permissions with the following command:

ssh bbpv1.bbp.epfl.ch \
"ls /gpfs/bbp.cscs.ch/data/project/proj101/dvc_remotes"

Possible outcomes:

# Access OK
atlas_annotation
atlas_interpolation

# Access denied
ls: cannot open directory [...]: Permission denied

Depending on whether you have access to the remote storage in the following sections you will either pull the data from the remote (dvc pull) or download the input data manually and re-run the data processing pipelines to reproduce the output data (dvc repro).

If you work on the BB5 and have access to the remote storage then run the following command to short-circuit the remote access (because the remote is located on the BB5 itself):

cd data
dvc remote add --local gpfs_proj101 \
  /gpfs/bbp.cscs.ch/data/project/proj101/dvc_remotes/atlas_annotation
cd ..

Get the Data

The purpose of the "Atlas Annotation" package is to align brain volumes and the corresponding atlases. This section explains how to get these data.

If you have access to the remote storage (see above) then all data can be readily pulled from it:

cd data
dvc pull
cd ..

In the case where you don't have access to the remote storage, the data need to be downloaded from the original sources and the pre-processing needs to be run. Note that the pre-processing may take a long time (around an hour). Run the following commands to start this process:

cd data
dvc repro
cd ..

In some cases you might not need all data. Then it is possible to download unprepared data that you need by running specific DVC stages. Refer to the data/README.md file for the description of different data files.

Examples

Here are some examples of the functionalities that one can find in the atlannot package.

Registration

One can compute the registration between a fixed and a moving image. Those images can be of any type (for example Atlas Annotations or simply intensity images). The inputs can be 2D or 3D, the only constraint is that they have to be of the same shape.

The main use-case of atlannot is the registration of brain volumes from one coordinate framework to another. It is then needed to allow some flexibility in terms of inputs type to accept any data such as regions annotations, intensity images.

import numpy as np

from atlannot.ants import register, transform

fixed = np.random.rand(20, 20)   # replace by a real image
moving = np.random.rand(20, 20)  # replace by a real image
# Computation of the displacement field from moving image to fixed image.
nii_data = register(fixed.astype(np.float32), moving.astype(np.float32)) 
# Apply the displacement to moving image.
warped = transform(moving.astype(np.float32), nii_data)

Image Manipulation

atlannot has also a lot of utility functions to manipulate images in order to make some pre-processing/post-processing on images.

A concrete example could be to combine a region annotation and an intensity image together and use the final result as an input to the registration. To merge information from both images, one could superpose regions borders of the annotation on top of the intensity image.

import numpy as np
from atlannot.utils import edge_laplacian_thin, merge

intensity_img = np.random.rand(20, 20) # Load intensity image here

# Create fake annotation image
annotation_img = np.zeros((20, 20))    # Load annotation image here
annotation_img[5:15, 5:15] = 1         # Load annotation image here

# Compute the borders of the annotation image
borders = edge_laplacian_thin(annotation_img)

# Merge intensity image and annotation image
merge_img = merge(intensity_img, borders)

See here other manipulation one can do on any kind of images:

import numpy as np
from atlannot.utils import (
  add_middle_line, 
  edge_laplacian_thick, 
  edge_laplacian_thin, 
  edge_sobel, 
  image_convolution,
  split_halfs,
)

# Instantiate an image 
img = np.random.rand(20, 20)  # Please replace by a real image

# Apply some convolution to the image
kernel = [[-1, -1, -1], [-1, 8, -1], [-1, -1, -1]]
img1 = image_convolution(img, kernel=kernel)
img2 = edge_laplacian_thick(img)
img3 = edge_sobel(img)
img4 = edge_laplacian_thick(img)

# Add a middle line, can choose the axis, the tickness, ...
img5 = add_middle_line(img, axis=0, thickness=2)

# Split the image into two 
half_imgs = split_halfs(img2, axis=0)[0]

Utilities

The atlannot contains other utilities:

• Atlas utilities:
  • Merge atlases to harmonize the scripts
  • Unfurl regions if the regions are structured in tree
  • Compute misalignments
  • Remapping the labels
• Notebook utilities:
  • Volume Viewer to see volume in every directions
  • Add colored legend to atlas images

Concrete examples

You can find numerous examples of the usage of atlannot package in the scripts located in the experiments directory.

git clone https://github.com/BlueBrain/atlas-annotation#egg=atlannot
cd atlas-annotation/experiments

To execute the scripts in this experiments folder, please first follow the data preparation instructions found in the data section.

Next, one needs also to install additional packages for interactive use.

pip install git+https://github.com/BlueBrain/atlas-annotation#egg=atlannot[interactive]

Once the cloning, the installation and the download of data is done, you can use any script, for example:

python ants2d_atlas_fine.py

Notebooks, Widgets, and Experiments

The additional functionality related to notebooks, widgets, and experiment scripts is not activated by default. In order to use it you need to specify an additional interactive option upon installing this package. This can be done as follows:

pip install git+https://github.com/BlueBrain/atlas-annotation#egg=atlannot[interactive]

Furthermore, you will need JupyterLab or Jupyter Notebook installed in your virtual environment, as well as the corresponding ipywidgets plugin. Follow the following online instructions in order to do so:

• How to install JupyterLab/Jupyter Notebook: https://jupyter.org
• How to install the ipywidgets plugin: https://ipywidgets.readthedocs.io/en/latest/user_install.html

Funding & Acknowledgment

The development of this software was supported by funding to the Blue Brain Project, a research center of the École polytechnique fédérale de Lausanne (EPFL), from the Swiss government’s ETH Board of the Swiss Federal Institutes of Technology.

Copyright (c) 2021-2022 Blue Brain Project/EPFL