![Execution time: related to a Pentium 4, 1.6-GHz, 512 GB, Windows XP computer using MATLAB version 7.0.
Nicolle Correa, Tu ̈lay Adali, and Vince D Calhoun. Performance of blind source separation algorithms for fMRI analysis using a group ICA method.
Magnetic resonance imaging, 25(5):684–94, jun 2007.
Algorithms
Spatial
Maps
Accuracy
Time
Courses
Accuracy
Parameters
Settings
Execution
Time [s]
Infomax 70
FastICA 38
JADE 62
EVD 3
ICA Algorithms
3](https://image.slidesharecdn.com/soaproject-180628211100/85/SoA_project-3-320.jpg)
![Related Works
6
User Friendliness
Execution Time
FSL[1]
Execution Time: 20 min
[1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT](https://image.slidesharecdn.com/soaproject-180628211100/85/SoA_project-6-320.jpg)
![Related Works
User Friendliness
Execution Time
Gift:
Execution Time: 8 min
7
FSL[1]
Execution Time: 20 min
[1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT](https://image.slidesharecdn.com/soaproject-180628211100/85/SoA_project-7-320.jpg)
![Related Works
User Friendliness
Execution Time
Gift:
Execution Time: 8 min
8
FSL[1]
Execution Time: 20 min
[1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT
FastBRaIn](https://image.slidesharecdn.com/soaproject-180628211100/85/SoA_project-8-320.jpg)
SoA_project
- 1. Valentina Corbetta: valentina.corbetta@mail.polimi.it Giada Casagrande: giada.casagrande@mail.polimi.it Filippo Carloni: filippo.carloni@mail.polimi.it 16th May 2018 XOHW18 Meeting @NECSTLab State of Art FastBRaIn
- 2. ICA The Independent Component Analysis (ICA) is a computational method for separating a multivariate signal into additive subcomponents. There are different approaches to implement ICA: Infomax maximum likelihood FastICA maximizes nongaussianity JADE joint diagonalization EVD second order correlation 2
- 3. Execution time: related to a Pentium 4, 1.6-GHz, 512 GB, Windows XP computer using MATLAB version 7.0. Nicolle Correa, Tu ̈lay Adali, and Vince D Calhoun. Performance of blind source separation algorithms for fMRI analysis using a group ICA method. Magnetic resonance imaging, 25(5):684–94, jun 2007. Algorithms Spatial Maps Accuracy Time Courses Accuracy Parameters Settings Execution Time [s] Infomax 70 FastICA 38 JADE 62 EVD 3 ICA Algorithms 3
- 4. 4
- 5. 5 Related Works Execution Time User Friendliness
- 6. Related Works 6 User Friendliness Execution Time FSL[1] Execution Time: 20 min [1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT
- 7. Related Works User Friendliness Execution Time Gift: Execution Time: 8 min 7 FSL[1] Execution Time: 20 min [1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT
- 8. Related Works User Friendliness Execution Time Gift: Execution Time: 8 min 8 FSL[1] Execution Time: 20 min [1]https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FEAT FastBRaIn
- 9. FastBRaIn Thank you for your attention FastBrainatNECST FastBRaIn_NECST Valentina Corbetta: valentina.corbetta@mail.polimi.it Giada Casagrande: giada.casagrande@mail.polimi.it Filippo Carloni: filippo.carloni@mail.polimi.it
Editor's Notes
- #4: FastICA the best overall performance Iterative algorithm Maximize non-Gaussianity of estimated sources Easy to use: requires initialization and setting of a limited number of parameters Independent components estimated individually
- #6: FSL is a comprehensive library of analysis tools for fMRI, MRI and DTI. It runs on Apple and PCs Gift is a MATLAB toolbox. It implements different algorithms for ICA applied to the analysis of fMRI