Manifold Learning for fMRI time-varying FC

Abstract

Whole-brain functional connectivity (FC) measured with functional MRI (fMRI) evolve over time in meaningful ways at temporal scales going from years (e.g., development) to seconds (e.g., within-scan time-varying FC (tvFC)). Yet, our ability to explore tvFC is severely constrained by its large dimensionality (several thousands). To overcome this difficulty, researchers seek to generate low dimensional representations (e.g., 2D and 3D scatter plots) expected to retain its most informative aspects (e.g., relationships to behavior, disease progression). Limited prior empirical work suggests that manifold learning techniques (MLTs)--namely those seeking to infer a low dimensional non-linear surface (i.e., the manifold) where most of the data lies--are good candidates for accomplishing this task. Here we explore this possibility in detail. First, we discuss why one should expect tvFC data to lie on a low dimensional manifold. Second, we estimate what is the intrinsic dimension (i.e., minimum number of latent dimensions; ID) of tvFC data manifolds. Third, we describe the inner workings of three state-of-the-art MLTs: Laplacian Eigenmaps (LE), T-distributed Stochastic Neighbor Embedding (T-SNE), and Uniform Manifold Approximation and Projection (UMAP). For each method, we empirically evaluate its ability to generate neuro-biologically meaningful representations of tvFC data, as well as their robustness against hyper-parameter selection. Our results show that tvFC data has an ID that ranges between 4 and 26, and that ID varies significantly between rest and task states. We also show how all three methods can effectively capture subject identity and task being performed: UMAP and T-SNE can capture these two levels of detail concurrently, but LE could only capture one at a time. We observed substantial variability in embedding quality across MLTs, and within-MLT as a function of hyper-parameter selection. To help alleviate this issue, we provide heuristics that can inform future studies. Finally, we also demonstrate the importance of feature normalization when combining data across subjects and the role that temporal autocorrelation plays in the application of MLTs to tvFC data. Overall, we conclude that while MLTs can be useful to generate summary views of labeled tvFC data, their application to unlabeled data such as resting-state remains challenging.

Manifold learning uncovers nonlinear interactions between the adolescent brain and environment that predict emotional and behavioral problems

“…Benchmarking and robustness analyses showed that: 1) in comparing E-PHATE to other dimensionality reduction methods, E-PHATE embeddings significantly out-perform PCA (43,44) and UMAP (45,46) (Supplemental Methods S7 and Figure S3); 2) the increased sensitivity of E-PHATE was attributable to added information specifically about the environment, versus an increase in the quantity of data about each participant ("E-PHATE control") (Supplemental Methods S8 and Figure S4); 3) the increased sensitivity of E-PHATE was attributable to the nonlinear combination of brain and environment ("PHATE + environment"); and 4) E-PHATE matrices built solely upon neighborhood disadvantage (ADI) or family conflict (47,48) improved associations relative to no environmental information, yet neither afforded as great of an improvement as the five-feature environment view (Supplemental Methods S9 and Figure S4).…”

Analyzing asymmetry in brain hierarchies with a linear state-space model of resting-state fMRI data

“…Multiple perspectives have been utilized to provide insights into this domain. These range from data-driven approaches, which seek to describe the time variability of functional connectivity (referred to as dynamical FC or dFC) [1, 59, 36], to more complex dynamical systems models that aim to replicate the dynamic structure of the state space [7, 78].…”