AxSI 0.0.17

AxSI for MRI data

Usage Guide

Overview

This documentation provides details on how to use the AxSI parser for analyzing MRI data using various input parameters.

Run from Command Line

To execute the program via the command line, use the following syntax:

axsi-main.py \
  --subj-folder /path/to/subject_folder \
  --run-name "run_001" \
  --data /path/to/data.nii.gz \
  --bval /path/to/bval.bval \
  --bvec /path/to/bvec.bvec \
  --mask /path/to/mask.nii.gz \
  --small-delta 20 \
  --big-delta 50 \
  --gmax 8.0 \
  --gamma-val 4258 \
  --num-processes-pred 35 \
  --num-threads-pred 1 \
  --num-processes-axsi 1 \
  --num-threads-axsi 10 \
  --linear-lsq-method R-quadprog \
  --nonlinear-lsq-method gurobi \
  --debug-mode

Required Arguments

• --subj-folder: Path to the subject folder (must exist)
• --run-name: Specify the name for the run
• --data: Path to the data file
• --bval: Path to the bval file
• --bvec: Path to the bvec file
• --mask: Path to the mask file

Optional Arguments

• --small-delta (default: 15): Gradient duration in milliseconds.
• --big-delta (default: 45): Time to scan (time interval) in milliseconds.
• --gmax (default: 7.9): Gradient maximum amplitude in G/cm.
• --gamma-val (default: 4257): Gyromagnetic ratio.
• --num-processes-pred (default: 1): Number of processes to run in parallel in prediction step.
• --num-threads-pred (default: 1): Number of threads to run in parallel in prediction step.
• --num-processes-axsi (default: 1): Number of processes to run in parallel in AxSI step.
• --num-threads-axsi (default: 1): Number of threads to run in parallel in AxSI step.
• --linear-lsq-method (default: R-quadprog): Method for linear least squares. **Choices **: R-quadprog, gurobi, scipy, cvxpy
• --nonlinear-lsq-method (default: R-minpack): Method for nonlinear least squares. **Choices **: R-minpack, scipy, lsq-axsi
• --debug-mode: Enable debug mode. If not provided, debug mode is disabled by default.

NIfTI Viewer

Overview

This is a Dash-based web application that allows users to interactively visualize slices of 3D or 4D NIfTI files ( commonly used in neuroimaging). Users can select slices along different axes, apply various color maps, and visualize data dynamically using sliders for timepoints (in 4D data) and slice indices.

Features

Input File Handling

• The NIfTI file path is provided as a command-line argument using argparse.
• The script loads and processes the provided file using the nibabel library.

Visualization

• Supports visualization along three axes: axial, sagittal, and coronal.
• Provides several color maps (e.g., gray, viridis, plasma) for the visualization.
• Interactive user interface for exploring slices.

4D Data Support

• If the input NIfTI file contains 4D data (e.g., time-series or multi-volume data), a slider lets users navigate through different timepoints.

User Controls

• Dropdown menus to select the viewing axis and color map.
• Sliders for selecting specific slices and timepoints.

Output

• The visualization is rendered as an interactive plot using Plotly.

Getting Started

Run the script from the command line, providing the NIfTI file path as an argument:

nifti-viewer --nifti-file /path/to/your/file.nii.gz

Example

nifti-viewer.py --nifti-file example_data/pasi.nii.gz

The app runs at http://127.0.0.1:8050 by default, displaying the interactive visualization.

Installation Instructions

Step 1: Install Miniconda

1. Download the Miniconda installer: Open your terminal and run the following command to download the Miniconda installer:

   wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh
   

2. Make the installer executable: Change the permissions of the downloaded script to make it executable:

   chmod +x Miniconda3-latest-Linux-x86_64.sh
   

3. Run the installer: Execute the installer script to install Miniconda:

   ./Miniconda3-latest-Linux-x86_64.sh
   

   Follow the on-screen instructions to complete the installation. You may need to agree to the license terms and specify the installation location.

Step 2: Configure Conda

4. Add the conda-forge channel: After installing Miniconda, configure it to use the conda-forge channel, which provides additional packages:

   conda config --add channels conda-forge
   

Step 3: Create a New Conda Environment

5. Create a new conda environment: Create a new environment named axsi with R version 4.4.2 and Python version 3.12:

   conda create --name axsi -c conda-forge r-base=4.4.2 python=3.12
   

6. Activate the new environment: Activate the newly created environment:

   conda activate axsi
   

Step 4: Verify Installation

7. Check the R installation: Verify that R is installed correctly by running:

   which R
   

8. Check the Python installation: Verify that Python is installed correctly by running:

   which python
   

9. Check Python version: Confirm the installed version of Python:

   python --version
   

10. Check R version: Confirm the installed version of R:

    R --version
    

Step 5: Install AxSI Packages

11. Install the axsi package: install the axsi package using pip:

    pip install axsi
    

12. Install AxSI-specific R components: The axsi package provide a command to install necessary R packages: 'quadprog' and 'minpack.lm'.

**Note**: R version 4.4.2 already installed on your system in the creation of the environment:

```bash
axsi-install-r
```

13. Optional: Run AxSI tests: AxSI comes with a test suite, you can run it using: bash axsi-run-tests

Execution

The main script shipped with this project is axsi-main, see its options by running:

axsi-main -h

This program implements the methodology from the paper:

Gast, H., Horowitz, A., Krupnik, R., Barazany, D., Lifshits, S., Ben-Amitay, S., & Assaf, Y. A Method for In-Vivo Mapping of Axonal Diameter Distributions in the Human Brain Using Diffusion-Based Axonal Spectrum Imaging (AxSI).

The Python code was originally developed by Hila Gast and improved by Refael Kohen [e-mail], Yaniv Assaf Lab, Tel Aviv University.