Probabilistic Inference on Virtual Brain Models of Disorders DOI Creative Commons
Meysam Hashemi, Abolfazl Ziaeemehr, Marmaduke Woodman

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 23, 2024

Abstract Connectome-based models, also known as Virtual Brain Models (VBMs), have been well established in network neuroscience to investigate pathophysiological causes underlying a large range of brain diseases. The integration an individual’s imaging data VBMs has improved patient-specific predictivity, although Bayesian estimation spatially distributed parameters remains challenging even with state-of-the-art Monte Carlo sampling. imply latent nonlinear state space models driven by noise and input, necessitating advanced probabilistic machine learning techniques for widely applicable estimation. Here we present Simulation-Based Inference on (SBI-VBMs), demonstrate that training deep neural networks both spatio-temporal functional features allows accurate generative disorders. systematic use stimulation provides effective remedy the non-identifiability issue estimating degradation intra-hemispheric connections. By prioritizing model structure over data, show hierarchical SBI-VBMs renders inference more effective, precise biologically plausible. This approach could broadly advance precision medicine enabling fast reliable prediction

Language: Английский

Simulation-based Inference on Virtual Brain Models of Disorders DOI Creative Commons
Meysam Hashemi, Abolfazl Ziaeemehr, Marmaduke Woodman

et al.

Machine Learning Science and Technology, Journal Year: 2024, Volume and Issue: 5(3), P. 035019 - 035019

Published: July 11, 2024

Abstract Connectome-based models, also known as virtual brain models (VBMs), have been well established in network neuroscience to investigate pathophysiological causes underlying a large range of diseases. The integration an individual’s imaging data VBMs has improved patient-specific predictivity, although Bayesian estimation spatially distributed parameters remains challenging even with state-of-the-art Monte Carlo sampling. imply latent nonlinear state space driven by noise and input, necessitating advanced probabilistic machine learning techniques for widely applicable estimation. Here we present simulation-based inference on (SBI-VBMs), demonstrate that training deep neural networks both spatio-temporal functional features allows accurate generative disorders. systematic use stimulation provides effective remedy the non-identifiability issue estimating degradation limited smaller subset connections. By prioritizing model structure over data, show hierarchical SBI-VBMs renders more effective, precise biologically plausible. This approach could broadly advance precision medicine enabling fast reliable prediction

Language: Английский

Citations

5
HUMA: Heterogeneous, Ultra Low-Latency Model Accelerator for The Virtual Brain on a Versal Adaptive SoC DOI Creative Commons

Amirreza Movahedin,

Lennart P. L. Landsmeer, Christos Strydis

et al.

Published: Feb. 26, 2025

Brain modeling can occur at different levels of abstraction, each aimed a purpose. The Virtual (TVB) is an open-source platform for constructing and simulating personalized brain-network models, favoring whole-brain macro-scales while reducing micro-level detail. Among other purposes, TVB used to build patient-specific, digital, brain twins that be in clinical settings, such as the study treatment epilepsy. However, fitting patient-specific models requires large number successive time-consuming simulations. By studying internal structure TVB, we observed heterogeneous computation needs its which could leveraged accelerate In this work, designed implemented HUMA, heterogeneous, ultra low-latency, dataflow architecture on AMD Versal Adaptive SoC patient-brain makeups. Our solution runs about 27× faster compared modern-day, server-class, 32-core CPU consuming fraction power. Additionally, it delivers average 14× lower latency, 1.7× better power efficiency order-of-magnitude energy consumption when against high-performance GPU version TVB. achieved latency savings reveal significant potential model-fitting individual patients well closed-loop biohybrid experiments.

Language: Английский

Citations

0
Probabilistic Inference on Virtual Brain Models of Disorders DOI Creative Commons
Meysam Hashemi, Abolfazl Ziaeemehr, Marmaduke Woodman

et al.

bioRxiv (Cold Spring Harbor Laboratory), Journal Year: 2024, Volume and Issue: unknown

Published: Feb. 23, 2024

Abstract Connectome-based models, also known as Virtual Brain Models (VBMs), have been well established in network neuroscience to investigate pathophysiological causes underlying a large range of brain diseases. The integration an individual’s imaging data VBMs has improved patient-specific predictivity, although Bayesian estimation spatially distributed parameters remains challenging even with state-of-the-art Monte Carlo sampling. imply latent nonlinear state space models driven by noise and input, necessitating advanced probabilistic machine learning techniques for widely applicable estimation. Here we present Simulation-Based Inference on (SBI-VBMs), demonstrate that training deep neural networks both spatio-temporal functional features allows accurate generative disorders. systematic use stimulation provides effective remedy the non-identifiability issue estimating degradation intra-hemispheric connections. By prioritizing model structure over data, show hierarchical SBI-VBMs renders inference more effective, precise biologically plausible. This approach could broadly advance precision medicine enabling fast reliable prediction

Language: Английский

Citations

0