Resting state fMRI analysis using unsupervised learning algorithms

ABSTRACT

RS-fMRI data analysis for functional connectivity explorations is a challenging topic in computational neuroimaging. Several approaches have been investigated to discover whole-brain data features. Among these, clustering techniques based on Competitive Learning (CL) and Spectral Methods (SM) have been shown effective in providing useful information in various contexts. We selected three clustering algorithms and two spectral methods, i.e the clustering algorithm are Self-organising Maps (SOM), Neural Gas (NG) and Growing Neural Gas (GNG), whereas the spectral methods are the classic Principal Component Analysis (PCA) and the Nonlinear Robust Fuzzy Principal Component Analysis (NRFPCA). We validated clustering with Davies–Bouldin Index (DBI) and we selected informative principal components using Random Matrix Theory (RMT). tools. We adopted these techniques to study the intrinsic functional properties of images coming from a shared repository of resting state fMRI experiments (1000 Functional Connectome Project).

KEYWORDS:

• RS-fMRI
• functional connectivity
• competitive clustering
• self organizing map
• neural gas
• growing neural gas
• Davies-Bouldin index
• spectral methods
• principal component analysis
• Nonlinear Robust Fuzzy Principal Component Analysis
• random matrix theory

Acknowledgments

This work is the extended version of the conference paper titled “Resting State fMRI Functional Connectivity Analysis Using Soft Competitive Learning Algorithms” that won the Taylor & Francis prize for the section “Imaging & Visualization” at the 15^th International Symposium on Computer Methods in Biomechanics and Biomedical Engineering and 3^rd Conference on Imaging and Visualisation (CMBBE2018, Lisbon, 26-28 March 2018). The authors thank the conference chairs Paulo R. Fernandes and João Manuel R. Tavares for this important achievement. The authors are also thankful to Dr.ssa Sabina Strocchi for the scientific support she gave about the experimental design and the data analysis of the conference paper.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Notes on contributors

Alberto Arturo Vergani

Alberto Arturo Vergani is an early stage researcher in computer science and mathematics. His works are focused on computational neuroimaging, neuroinformatics and clinical computer science. He is engaged in several international educative programmes organized by International Neuroinformatics Coordinating Facility (INCF), Partnership for Advance Computing in Europe (PRACE) and Human Brain Project (HBP). He is student member of IEEE Computational Intelligence Society and Association for Computing Machinery.