Unsupervised Spatiotemporal Analysis of FMRI Data Using Graph-Based Visualizations of Self-Organizing Maps

Abstract

We present novel graph-based visualizations of self-organizing maps for unsupervised functional magnetic resonance imaging (fMRI) analysis. A self-organizing map is an artificial neural network model that transforms high-dimensional data into a low-dimensional (often a 2-D) map using unsupervised learning. However, a postprocessing scheme is necessary to correctly interpret similarity between neighboring node prototypes (feature vectors) on the output map and delineate clusters and features of interest in the data. In this paper, we used graph-based visualizations to capture fMRI data features based upon 1) the distribution of data across the receptive fields of the prototypes (density-based connectivity); and 2) temporal similarities (correlations) between the prototypes (correlation-based connectivity). We applied this approach to identify task-related brain areas in an fMRI reaction time experiment involving a visuo-manual response task, and we correlated the time-to-peak of the fMRI responses in these areas with reaction time. Visualization of self-organizing maps outperformed independent component analysis and voxelwise univariate linear regression analysis in identifying and classifying relevant brain regions. We conclude that the graph-based visualizations of self-organizing maps help in advanced visualization of cluster boundaries in fMRI data enabling the separation of regions with small differences in the timings of their brain responses.

Variance decomposition for single-subject task-based fMRI activity estimates across many sessions

“…It may be that substantial reductions in σ 2 err at the single-subject level may only be obtained via alternative multivariate data-driven analytical methods, such as Independent Component Analysis (Calhoun et al, 2001) or Self Organizing Maps (Katwal et al, 2013) that rely on a less stringent set of underlying assumptions about the data. Any such efforts may be vital to the success of longitudinal single-subject examinations with fMRI.…”

Evolving, dynamic clustering of spatio/spectro-temporal data in 3D spiking neural network models and a case study on EEG data

“…EFuNN (Kasabov 2001) was introduced for evolving supervised clustering related to classification or regression. In (Katwal et al 2013), a graph-based visualisations of SOM has been used for clustering fMRI data. In another research (Liao et al 2008), an integrated SOM and hierarchical clustering architecture was designed to detect activation patterns of fMRI data.…”

Extending parallelization of the self-organizing map by combining data and network partitioned methods

“…Kohonen's Self-Organizing Map (SOM) [10] allows for low-dimensional visualization of a high-dimensional space while the data's topology is preserved. It has been used extensively to avoid many of the issues noted with modern data visualization methods [11][12][13][14]. The Contextual SOM (CSOM) supplements the standard SOM with the addition of a contextual label on the individual nodes of the resulting SOM.…”