Construction of embedded fMRI resting state functional connectivity networks using manifold learning
This work addresses improving diagnostic tools for schizophrenia patients, but it is incremental as it compares existing manifold learning methods on a known dataset.
The paper tackled the problem of constructing functional connectivity networks from fMRI data for schizophrenia classification, showing that using Diffusion Maps with a lagged cross-correlation metric outperformed other combinations in classification performance.
We construct embedded functional connectivity networks (FCN) from benchmark resting-state functional magnetic resonance imaging (rsfMRI) data acquired from patients with schizophrenia and healthy controls based on linear and nonlinear manifold learning algorithms, namely, Multidimensional Scaling (MDS), Isometric Feature Mapping (ISOMAP) and Diffusion Maps. Furthermore, based on key global graph-theoretical properties of the embedded FCN, we compare their classification potential using machine learning techniques. We also assess the performance of two metrics that are widely used for the construction of FCN from fMRI, namely the Euclidean distance and the lagged cross-correlation metric. We show that the FCN constructed with Diffusion Maps and the lagged cross-correlation metric outperform the other combinations.