Multi-view manifold learning of human brain state trajectories

Busch, Erica L.; Huang, Jessie; Benz, Andrew; Wallenstein, Tom; Lajoie, Guillaume; Wolf, Guy; Krishnaswamy, Smita; Turk‐Browne, Nicholas B.

doi:10.1101/2022.05.03.490534

Cited by 10 publications

(19 citation statements)

References 78 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This work is limited not only in terms of data size ( 20 scans) and evaluation metrics (clustering and classification), but also in terms of the breath of methods being evaluated. For example, manifold estimation can also be accomplished via multidimensional scaling (Kruskal, 1964), ISOMAP (Tenenbaum et al, 2000), diffusion maps (Coifman et al, 2005) or T-PHATE (Busch et al, 2022), to name a few additional MLTs not considered here. All these other methods have been previously applied to fMRI data using either regional levels of activity (Busch et al, 2022; Gao et al, 2021) or static FC (Ioannis K Gallos et al, 2021; Ioannis K. Gallos et al, 2021) as inputs.…”

Section: Discussionmentioning

confidence: 99%

“…For example, manifold estimation can also be accomplished via multidimensional scaling (Kruskal, 1964), ISOMAP (Tenenbaum et al, 2000), diffusion maps (Coifman et al, 2005) or T-PHATE (Busch et al, 2022), to name a few additional MLTs not considered here. All these other methods have been previously applied to fMRI data using either regional levels of activity (Busch et al, 2022; Gao et al, 2021) or static FC (Ioannis K Gallos et al, 2021; Ioannis K. Gallos et al, 2021) as inputs. Future research should evaluate their efficacy on tvFC data.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Manifold Learning for fMRI time-varying FC

Gonzalez-Castillo

Fernandez

Lam

et al. 2023

Preprint

View full text Add to dashboard Cite

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.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Manifold Learning for fMRI time-varying FC

Gonzalez-Castillo

Fernandez

Lam

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Additionally, we only used PCA to define the manifold. However, many nonlinear methods exist 7,41,42 which may better identify task co-activation patterns. Further advances through nonlinear techniques may reveal complementary results.…”

Section: Limitationsmentioning

confidence: 99%

Connectome caricatures: removing large-amplitude co-activation patterns in resting-state fMRI emphasizes individual differences

Rodriguez,

Noble,

Camp

et al. 2024

Preprint

View full text Add to dashboard Cite

High-amplitude co-activation patterns are sparsely present during resting-state fMRI but drive functional connectivity1–5. Further, they resemble task activation patterns and are well-studied3,5–10. However, little research has characterized the remaining majority of the resting-state signal. In this work, we introduced caricaturing—a method to project resting-state data to a subspace orthogonal to a manifold of co-activation patterns estimated from the task fMRI data. Projecting to this subspace removes linear combinations of these co-activation patterns from the resting-state data to create Caricatured connectomes. We used rich task data from the Human Connectome Project (HCP)11and the UCLA Consortium for Neuropsychiatric Phenomics12to construct a manifold of task co-activation patterns. Caricatured connectomes were created by projecting resting-state data from the HCP and the Yale Test-Retest13datasets away from this manifold. Like caricatures, these connectomes emphasized individual differences by reducing between-individual similarity and increasing individual identification14. They also improved predictive modeling of brain-phenotype associations. As caricaturing removes group-relevant task variance, it is an initial attempt to remove task-like co-activations from rest. Therefore, our results suggest that there is a useful signal beyond the dominating co-activations that drive resting-state functional connectivity, which may better characterize the brain’s intrinsic functional architecture.

show abstract

“…The difficult answers and unappealing tradeoffs associated with these questions have spurred clever solutions optimized for designing around all of this complexity. Such approaches include calculating intersubject correlations (ISC; Hasson et al, 2004 , 2008b ), or borrowing other methods from resting-state fMRI, dynamic causal analyses (e.g., Granger causality or DCM methods), or introducing other advanced tools to decipher entangled brain responses ( Di and Biswal, 2020 ; Van Der Meer et al, 2020 ; Busch et al, 2022 ).…”

Section: The Arrow Of Causality: From Media Content To Reception Resp...mentioning

confidence: 99%

Integrating media content analysis, reception analysis, and media effects studies

Schmälzle

Huskey

2023

Front. Neurosci.

View full text Add to dashboard Cite

Every day, the world of media is at our fingertips, whether it is watching movies, listening to the radio, or browsing online media. On average, people spend over 8 h per day consuming messages from the mass media, amounting to a total lifetime dose of more than 20 years in which conceptual content stimulates our brains. Effects from this flood of information range from short-term attention bursts (e.g., by breaking news features or viral ‘memes’) to life-long memories (e.g., of one’s favorite childhood movie), and from micro-level impacts on an individual’s memory, attitudes, and behaviors to macro-level effects on nations or generations. The modern study of media’s influence on society dates back to the 1940s. This body of mass communication scholarship has largely asked, “what is media’s effect on the individual?” Around the time of the cognitive revolution, media psychologists began to ask, “what cognitive processes are involved in media processing?” More recently, neuroimaging researchers started using real-life media as stimuli to examine perception and cognition under more natural conditions. Such research asks: “what can media tell us about brain function?” With some exceptions, these bodies of scholarship often talk past each other. An integration offers new insights into the neurocognitive mechanisms through which media affect single individuals and entire audiences. However, this endeavor faces the same challenges as all interdisciplinary approaches: Researchers with different backgrounds have different levels of expertise, goals, and foci. For instance, neuroimaging researchers label media stimuli as “naturalistic” although they are in many ways rather artificial. Similarly, media experts are typically unfamiliar with the brain. Neither media creators nor neuroscientifically oriented researchers approach media effects from a social scientific perspective, which is the domain of yet another species. In this article, we provide an overview of approaches and traditions to studying media, and we review the emerging literature that aims to connect these streams. We introduce an organizing scheme that connects the causal paths from media content → brain responses → media effects and discuss network control theory as a promising framework to integrate media content, reception, and effects analyses.

show abstract

Multi-view manifold learning of human brain state trajectories

Cited by 10 publications

References 78 publications

Manifold Learning for fMRI time-varying FC

Manifold Learning for fMRI time-varying FC

Connectome caricatures: removing large-amplitude co-activation patterns in resting-state fMRI emphasizes individual differences

Integrating media content analysis, reception analysis, and media effects studies

Contact Info

Product

Resources

About