Searcharxiv⌕ Search

arXiv subjects

J. Ramírez

Publications and source records attributed to J. Ramírez.

3 recordsLinked to original sources

Heterogeneous ice nucleation with TIP4P/Ice: the role of lattice mismatch

Ice formation, a crucial phase transition for life on Earth, mainly occurs through heterogeneous nucleation on a solid substrate. Despite the importance of this phenomenon, the correlation between the structure of the substrate and its ability to induce ice nucleation remains poorly understood. In this work, we use molecular dynamics simulations with an atomistic model of water to investigate how the lattice mismatch affects the ability of the substrate to nucleate ice. To unambiguously quantify the mismatch, we use substrates with a stretched ice structure. We put the substrate in contact with supercooled water and determine the temperature at which ice spontaneously nucleates during the course of the simulations. We find that the nucleation temperature decreases by $\sim$ 5-6 K for each percent by which the substrate lattice is stretched relative to that of ice. This trend is similar to that previously obtained with a coarse-grained water model [Camarillo et al., J. Chem. Phys. 160, 134505 (2024)]. \textcolor{black}{The substrate induces a fluctuating network of subcritical ice-like clusters, from which a spherical-cap-shaped supercritical nucleus eventually emerges stochastically.} Using Classical Nucleation Theory, combine previously reported homogeneous nucleation rates with heterogeneous nucleation rates computed in this work to estimate key heterogeneous nucleation parameters, including the nucleation barrier, the kinetic pre-factor, and the contact angle. The latter is consistent with a microscopic visualization of nucleation trajectories, supporting the applicability of Classical Nucleation Theory to describe a microscopic phenomenon such as heterogeneous ice nucleation.

cond-mat.stat-mech↗

An explainable framework for the relationship between dementia and glucose metabolism patterns

High-dimensional neuroimaging data presents challenges for assessing neurodegenerative diseases due to complex non-linear relationships. Variational Autoencoders (VAEs) can encode scans into lower-dimensional latent spaces capturing disease-relevant features. We propose a semi-supervised VAE framework with a flexible similarity regularization term that aligns selected latent variables with clinical or biomarker measures of dementia progression. This allows adapting the similarity metric and supervised variables to specific goals or available data. We demonstrate the approach using PET scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI), guiding the first latent dimension to align with a cognitive score. Using this supervised latent variable, we generate average reconstructions across levels of cognitive impairment. Voxel-wise GLM analysis reveals reduced metabolism in key regions, mainly the hippocampus, and within major Resting State Networks, particularly the Default Mode and Central Executive Networks. The remaining latent variables encode affine transformations and intensity variations, capturing confounds such as inter-subject variability and site effects. Our framework effectively extracts disease-related patterns aligned with established Alzheimer's biomarkers, offering an interpretable and adaptable tool for studying neurodegenerative progression.

cs.LG↗

Revealing Patterns of Symptomatology in Parkinson's Disease: A Latent Space Analysis with 3D Convolutional Autoencoders

This work proposes the use of 3D convolutional variational autoencoders (CVAEs) to trace the changes and symptomatology produced by neurodegeneration in Parkinson's disease (PD). In this work, we present a novel approach to detect and quantify changes in dopamine transporter (DaT) concentration and its spatial patterns using 3D CVAEs on Ioflupane (FPCIT) imaging. Our approach leverages the power of deep learning to learn a low-dimensional representation of the brain imaging data, which then is linked to different symptom categories using regression algorithms. We demonstrate the effectiveness of our approach on a dataset of PD patients and healthy controls, and show that general symptomatology (UPDRS) is linked to a d-dimensional decomposition via the CVAE with R2>0.25. Our work shows the potential of representation learning not only in early diagnosis but in understanding neurodegeneration processes and symptomatology.

eess.IV↗