AMICO

Implementation of the linear framework for Accelerated Microstructure Imaging via Convex Optimization (AMICO) described here:

Accelerated Microstructure Imaging via Convex Optimization (AMICO) from diffusion MRI data
Alessandro Daducci, Erick Canales-Rodriguez, Hui Zhang, Tim Dyrby, Daniel Alexander, Jean-Philippe Thiran
NeuroImage 105, pp. 32-44 (2015)

Installation

Download and install external software

• NODDI MATLAB toolbox. Download the software and follow the instructions provided here to install it.

• CAMINO toolkit. Download the software and follow the instructions provided here to install it.

• SPArse Modeling Software (SPAMS). Download the software and follow the instructions provided here to install it.

Setup paths/variables in MATLAB

Add the folder containing the source code of AMICO to your MATLAB PATH.

Copy the file AMICO_Setup.txt and rename it to AMICO_Setup.m. Modify its content to set the paths to your specific needs, as follows:

• AMICO_code_path : path to the folder containing the source code of AMICO (this repository). E.g. /home/user/AMICO/code.

• NODDI_path : path to the folder containing the source code of the NODDI toolbox (in case you want to use NODDI, not needed for ActiveAx). E.g. /home/user/NODDI_toolbox_v0.9.

• CAMINO_path : path to the bin folder containing the executables of the Camino toolkit (in case you want to use ActiveAx, not needed for NODDI). E.g. /home/user/camino/bin.

• SPAMS_path : path to the folder containing the source code of the SPAMS Library. E.g. /home/user/spams.

• AMICO_data_path : path to the folder where you store all your datasets. E.g. /home/user/AMICO/data. Then, the software assumes the folder structure is the following:

  ├── data
      ├── Study_01                 --> all subjects acquired with protocol "Study_01"
          ├── Subject_01
          ├── Subject_02
          ├── ...
      ├── Study_02                 --> all subjects acquired with protocol "Study_02"
          ├── Subject_01
          ├── Subject_02
          ├── ...
      ├── ...
  

  This way, the kernels need to be computed only once per each study, i.e. same protocol (number of shells, b-values etc), and subsequently adapted to each subject (specific gradient directions) very efficiently.

Getting started

Tutorials/demos are provided in the folder doc/demos to help you get started with the AMICO framework.