DySCo: a general framework for dynamic Functional Connectivity DOI Creative Commons
Giuseppe de Alteriis,

Oliver Sherwood,

Alessandro Ciaramella

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Июнь 13, 2024

A crucial challenge in neuroscience involves characterising brain dynamics from high-dimensional recordings. Dynamic Functional Connectivity (dFC) is an analysis paradigm that aims to address this challenge. dFC consists of a time-varying matrix (dFC matrix) expressing how pairwise interactions across areas change with time. However, the main approaches have been developed and applied mostly empirically, lacking unifying theoretical framework, general interpretation, common set measures quantify matrices properties. Moreover, field has ad-hoc algorithms compute process efficiently. This prevented show its full potential datasets and/or real time applications. With paper, we introduce Symmetric Matrix framework (DySCo), associated repository. DySCo approach allows study signals at different spatio-temporal scales, down voxel level, computationally ultrafast. unifies single most employed matrices, which share mathematical structure. Doing so it allows: 1) new interpretation further justifies use capture spatiotemporal patterns data form easily translatable imaging modalities. 2) The introduction Recurrence EVD store eigenvectors eigenvalues all types efficent manner orders magnitude faster than naive algorithms, without loss information. 3) To simply define quantities interest for dynamic analyses such as: amount connectivity (norm similarity between their informational complexity. methodology here validated on both synthetic dataset rest/N-back task experimental - fMRI Human Connectome Project dataset. We demonstrate proposed are highly sensitive changes configurations. illustrate computational efficiency toolbox, perform voxel-level, very demanding afforded by RMEVD algorithm.

Язык: Английский

Is Ockham’s razor losing its edge? New perspectives on the principle of model parsimony DOI Creative Commons
Marina Dubova, Suyog Chandramouli, Gerd Gigerenzer

и другие.

Proceedings of the National Academy of Sciences, Год журнала: 2025, Номер 122(5)

Опубликована: Янв. 27, 2025

The preference for simple explanations, known as the parsimony principle, has long guided development of scientific theories, hypotheses, and models. Yet recent years have seen a number successes in employing highly complex models inquiry (e.g., 3D protein folding or climate forecasting). In this paper, we reexamine principle light these technological advancements. We review developments, including surprising benefits modeling with more parameters than data, increasing appreciation context-sensitivity data misspecification models, new tools. By integrating insights, reassess utility proxy desirable model traits, such predictive accuracy, interpretability, effectiveness guiding research, resource efficiency. conclude that are sometimes essential progress, discuss ways which complexity can play complementary roles practice.

Язык: Английский

Процитировано

2
Molecular causality in the advent of foundation models DOI Creative Commons
Sebastian Lobentanzer, Pablo Rodríguez-Mier, Stefan Bauer

и другие.

Molecular Systems Biology, Год журнала: 2024, Номер 20(8), С. 848 - 858

Опубликована: Июнь 18, 2024

Язык: Английский

Процитировано

7
DySCo: A general framework for dynamic functional connectivity DOI Creative Commons
Giuseppe de Alteriis,

Oliver Sherwood,

Alessandro Ciaramella

и другие.

PLoS Computational Biology, Год журнала: 2025, Номер 21(3), С. e1012795 - e1012795

Опубликована: Март 7, 2025

A crucial challenge in neuroscience involves characterising brain dynamics from high-dimensional recordings. Dynamic Functional Connectivity (dFC) is an analysis paradigm that aims to address this challenge. dFC consists of a time-varying matrix (dFC matrix) expressing how pairwise interactions across areas change over time. However, the main approaches have been developed and applied mostly empirically, lacking common theoretical framework clear view on interpretation results derived matrices. Moreover, community has not using most efficient algorithms compute process matrices efficiently, which prevented showing its full potential with datasets and/or real-time applications. In paper, we introduce Symmetric Matrix (DySCo), associated repository. DySCo presents commonly used measures language implements them computationally way. This allows study activity at different spatio-temporal scales, down voxel level. provides single to: (1) Use as tool capture interaction patterns data form easily translatable imaging modalities. (2) Provide comprehensive set quantify properties evolution time: amount connectivity, similarity between matrices, their informational complexity. By combining it possible perform analysis. (3) Leverage Temporal Covariance EVD algorithm (TCEVD) store eigenvectors values then also EVD. Developing eigenvector space orders magnitude faster more memory than naïve space, without loss information. The methodology here validated both synthetic dataset rest/N-back task experimental fMRI Human Connectome Project dataset. We show all proposed are sensitive changes configurations consistent time subjects. To illustrate computational efficiency toolbox, performed level, demanding but afforded by TCEVD.

Язык: Английский

Процитировано

1
Assistive sensory-motor perturbations influence learned neural representations DOI Creative Commons
Pavithra Rajeswaran, Alexandre Payeur, Guillaume Lajoie

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2024, Номер unknown

Опубликована: Март 20, 2024

Task errors are used to learn and refine motor skills. We investigated how task assistance influences learned neural representations using Brain-Computer Interfaces (BCIs), which map activity into movement via a decoder. analyzed cortex as monkeys practiced BCI with decoder that adapted improve or maintain performance over days. The dimensionality of the population neurons controlling remained constant increased learning, counter expected trends from learning. Yet, time, information was contained in smaller subset modes. Moreover, ultimately stored modes occupied small fraction variance. An artificial network model suggests adaptive decoders contribute forming these compact representations. Our findings show assistive manipulate error for long-term learning computations, like credit assignment, informs our understanding has implications designing real-world BCIs.

Язык: Английский

Процитировано

5
Is Ockham’s razor losing its edge? New perspectives on the principle of model parsimony DOI Open Access
Marina Dubova, Suyog Chandramouli, Gerd Gigerenzer

и другие.

Опубликована: Март 19, 2024

The preference for simpler explanations, known as the parsimony principle, has long guided development of scientific theories, hypotheses, and models. Yet recent years have seen a number successes in employing highly complex models inquiry (e.g., 3D protein folding or climate forecasting). In this paper, we re-examine principle light these technological advancements. We review developments, including surprising benefits modeling with more parameters than data, increasing appreciation context-sensitivity data misspecification models, new tools. By integrating insights, reassess utility proxy desirable model traits, such predictive accuracy, interpretability, effectiveness guiding research, resource efficiency. conclude that are sometimes essential progress, discuss ways which complexity can play complementary roles practice.

Язык: Английский

Процитировано

4
Evaluating Melatonin's Effects on Hepatocyte Lipidome: A Critique of Analytical Methods DOI
Yoshiyasu Takefuji

Journal of Pineal Research, Год журнала: 2025, Номер 77(3)

Опубликована: Апрель 1, 2025

Язык: Английский

Процитировано

0
Reevaluating analytical approaches in systemic sclerosis research: challenges of PCA and logistic regression DOI
Yoshiyasu Takefuji

Annals of the Rheumatic Diseases, Год журнала: 2025, Номер unknown

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0
Reevaluating Feature Selection in Machine Learning-Based Radiomics for Hepatocellular Carcinoma: Bridging the Gap Between Predictive Accuracy and Biological Relevance DOI
Yoshiyasu Takefuji

Journal of Hepatology, Год журнала: 2025, Номер unknown

Опубликована: Май 1, 2025

Язык: Английский

Процитировано

0
Unmanned aerial vehicle-based prediction of cold tolerance indicators in sugarcane (Saccharum spp. hybrids) varieties DOI Creative Commons
Minori Uchimiya,

André Fróes de Borja Reis,

Bruno Cocco Lago

и другие.

Industrial Crops and Products, Год журнала: 2025, Номер 232, С. 121289 - 121289

Опубликована: Июнь 5, 2025

Язык: Английский

Процитировано

0
Arousal as a universal embedding for spatiotemporal brain dynamics DOI Creative Commons
Ryan V. Raut, Zachary P. Rosenthal, Xiaodan Wang

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2023, Номер unknown

Опубликована: Ноя. 8, 2023

Neural activity in awake organisms shows widespread and spatiotemporally diverse correlations with behavioral physiological measurements. We propose that this covariation reflects part the dynamics of a unified, multidimensional arousal-related process regulates brain-wide physiology on timescale seconds. By framing interpretation within dynamical systems theory, we arrive at surprising prediction: single, scalar measurement arousal (e.g., pupil diameter) should suffice to reconstruct continuous evolution multidimensional, spatiotemporal measurements large-scale brain physiology. To test hypothesis, perform multimodal, cortex-wide optical imaging monitoring mice. demonstrate neuronal calcium, metabolism, blood-oxygen can be accurately parsimoniously modeled from low-dimensional state-space reconstructed time history diameter. Extending framework electrophysiological Allen Brain Observatory, ability integrate experimental data into unified generative model via mappings an intrinsic manifold. Our results support hypothesis spontaneous, spatially structured fluctuations physiology-widely interpreted reflect regionally-specific neural communication-are large reflections process. This enriched view has broad implications for interpreting observations brain, body, behavior as measured across modalities, contexts, scales.

Язык: Английский

Процитировано

7