Agnese Callegari

Invited Presentation by B. Zufiria-Gerbolés at SPIE-ETAI, San Diego, 7 August 2025

Memory capacity in aging. A Brain reservoir computing architecture with uniform random signals applied to all nodes. *(Image from the article.)*

Computational memory capacity predicts aging and cognitive decline
Blanca Zufiria-Gerbolés, Mite Mijalkov, Ludvig Storm, Dániel Veréb, Zhilei Xu, Anna Canal-Garcia, Jiawei Sun, Yu-Wei Chang, Hang Zhao, Emiliano Gómez-Ruiz, Massimiliano Passaretti, Sara Garcia-Ptacek, Miia Kivipelto, Per Svenningsson, Henrik Zetterberg, Heidi Jacobs, Kathy Lüdge, Daniel Brunner, Bernhard Mehlig, Giovanni Volpe, Joana B. Pereira
SPIE-ETAI, San Diego, CA, USA, 3 – 7 August 2025
Date: 7 August 2025
Time: 11:15 AM – 11:45 AM PDT
Place: Conv. Ctr. Room 4

Using reservoir computing and diffusion-weighted imaging, we explored changes in brain connectivity patterns and their impact on cognition during aging. We found that whole-brain networks perform optimally at low densities, with performance decreasing as network density increases, particularly in regions with weaker connections. This decline was strongly associated with age and cognitive performance. Our results suggest that a core network of anatomical hubs is essential for optimal brain function, while peripheral connections are more vulnerable to aging. This study highlights the potential of reservoir computing for understanding age-related cognitive decline.

Reference
Mite Mijalkov, Ludvig Storm, Blanca Zufiria-Gerbolés, Dániel Veréb, Zhilei Xu, Anna Canal-Garcia, Jiawei Sun, Yu-Wei Chang, Hang Zhao, Emiliano Gómez-Ruiz, Massimiliano Passaretti, Sara Garcia-Ptacek, Miia Kivipelto, Per Svenningsson, Henrik Zetterberg, Heidi Jacobs, Kathy Lüdge, Daniel Brunner, Bernhard Mehlig, Giovanni Volpe, Joana B. Pereira, Computational memory capacity predicts aging and cognitive decline
Nature Communications 16, 2748 (2025)

Presentation by Anoop C. Patil at SPIE-ETAI, San Diego, 6 August 2025

In this work, we present an unsupervised deep learning framework using Variational Autoencoders (VAEs) to decode stress-specific biomolecular fingerprints directly from Raman spectral data across multiple plant species and genotypes. *(Image by the Authors of the manuscript. A part of the image was designed using Biorender.com.)*

From Spectra to Stress: Unsupervised Deep Learning for Plant Health Monitoring
Anoop C. Patil, Benny Jian Rong Sng, Yu-Wei Chang, Joana B. Pereira, Chua Nam-Hai, Rajani Sarojam, Gajendra Pratap Singh, In-Cheol Jang, and Giovanni Volpe
Date: 6 August 2025
Time: 10:30 AM – 11:00 AM
Place: Conv. Ctr. Room 4

Plants experience a wide variety of stresses, from light and temperature fluctuations to bacterial infections. Each stress has a biomolecular fingerprint, but detecting and interpreting these signatures across species can be challenging. This work presents a deep learning-based approach using Variational Autoencoders (VAEs) to uncover how plants respond to light stress, shade avoidance, temperature stress, and bacterial infection — all without requiring any human intervention in spectral processing. By encoding Raman spectral data into an intuitive latent space, this method automatically categorizes and visualizes stress-specific biomolecular shifts, offering a powerful, unsupervised tool for stress phenotyping in crops.

Reference:
Patil, A.C. et al. Deep-Learning Investigation of Vibrational Raman Spectra for Plant-Stress Analysis. arXiv preprint arXiv:2507.15772v1 (2025). URL https://arxiv.org/abs/2507.15772

Presentation by M. Selin at SPIE-ETAI, San Diego, 5 August 2025

Illustration of three different experiments autonomously performed by the SmartTrap system: DNA pulling experiments (top), red blood cell stretching (bottom left), and particle-particle interaction measurements (bottom right). *(Image by M. Selin.)*

Advanced autonomous optical tweezers experiments
Martin Selin, A. Ciarlo, G. Pesce, L. Bengtsson, J. Camuñas-Soler, V. Sundar Rajan, F. Westerlund, L. M. Wilhelmsson, I. Pastor, F. Ritort, S. B. Smith, C. Bustamante, G. Volpe
Date: 5 August 2025
Time: 4:30 PM – 4:45 PM PDT
Place: Conv. Ctr. Room 4

Single-molecule studies are vital for elucidating fundamental biological processes, including protein folding, DNA transcription, and replication. However, performing these experiments manually on individual molecules is notoriously time-consuming and costly. To address this challenge, we have developed a fully autonomous single-molecule force spectroscopy platform by integrating a custom-built optical tweezers instrument with real-time deep-learning-based image analysis and adaptive control protocols. Our system achieves human-level throughput in terms of experiments per hour while remaining robust enough to operate continuously for hours without intervention. We demonstrate the versatility of our platform by having it perform DNA pulling experiments on both lambda DNA and DNA hairpins fully autonomously. These results push the boundaries of high-throughput data collection in single-molecule biophysics, paving the way for merging single-molecule studies with large-scale, data-driven approaches—ultimately enabling new insights into the dynamic, transient states of complex biological systems.

Invited Presentation by B. Yeroshenko at SPIE-ETAI, San Diego, 5 August 2025

Kymographs of DNA inside Channel II. *(Image from 10.1038/s41592-022-01491-6. by Barbora Špačková)*

Pushing the limits of label-free single-molecule characterization by nanofluidic scattering microscopy
Bohdan Yeroshenko, Henrik Klein Moberg, Leyla Beckerman, Joachim Fritzsche, David Albinsson, Barbora Špačková, Daniel Midtvedt, Giovanni Volpe, Christoph Langhammer
SPIE-ETAI, San Diego, CA, USA, 3 – 7 August 2025
Date: 5 August 2025
Time: 8:45 AM – 9:15 AM PDT
Place: Conv. Ctr. Room 4

Nanofluidic Scattering Microscopy (NSM) is a label-free characterization method that leverages the interference of light scattered by nanochannels and single molecules within them. This technique enables accurate determination of molecular weight and hydrodynamic radius based solely on scattering, without requiring prior molecular knowledge. However, standard analysis methods limit NSM’s sensitivity to 66 kDa for proteins. In this presentation, I will demonstrate how we push this detection limit an order of magnitude further by integrating ultrasmall geometry with an advanced machine learning analysis approach, all while maintaining the same input laser power intensity.

Poster by A. Callegari at SPIE-OTOM, San Diego, 4 August 2025

One exemplar of the HEXBUGS used in the experiment. (Image by the Authors of the manuscript.)

Experimenting with macroscopic active matter
Angelo Barona Balda, Aykut Argun, Agnese Callegari, Giovanni Volpe
SPIE-OTOM, San Diego, CA, USA, 3 – 7 August 2025
Date: 4 August 2025
Time: 5:30 PM – 7:30 PM PDT
Place: Conv. Ctr. Exhibit Hall A

Presenter: Giovanni Volpe
Contribution submitted by Agnese Callegari

Active matter is based on concepts of nonequilibrium thermodynamics applied to the most diverse disciplines. A key concept is the active Brownian particle, which, unlike passive ones, extracts energy from its environment to generate complex motion and emergent behaviors. Despite its significance, active matter remains absent from standard curricula. This work presents macroscopic experiments using commercially available Hexbugs to demonstrate active matter phenomena. We show how Hexbugs can be modified to perform both regular and chiral active Brownian motion and interact with passive objects, inducing movement and rotation. By introducing obstacles, we sort Hexbugs based on motility and chirality. Finally, we demonstrate a Casimir-like attraction effect between planar objects in the presence of active particles.

Reference
Angelo Barona Balda, Aykut Argun, Agnese Callegari, Giovanni Volpe
Playing with Active Matter, Americal Journal of Physics 92, 847–858 (2024)

Poster by A. Callegari at SPIE-ETAI, San Diego, 4 August 2025

Focused rays scattered by an ellipsoidal particles (left). Optical torque along y calculated in the x-y plane using ray scattering with a grid of 1600 rays (up, right) and using a trained neural network (down, right). (Image by the Authors of the manuscript.)

Dense neural networks for geometrical optics
David Bronte Ciriza, Alessandro Magazzù, Agnese Callegari, Gunther Barbosa, Antonio A. R. Neves, Maria Antonia Iatì, Giovanni Volpe, and Onofrio M. Maragò
SPIE-ETAI, San Diego, CA, USA, 3 – 7 August 2025
Date: 4 August 2025
Time: 5:30 PM – 7:30 PM PDT
Place: Conv. Ctr. Exhibit Hall A

Presenter: Giovanni Volpe
Contribution submitted by Agnese Callegari

Light can trap and manipulate microscopic objects through optical forces and torques, as seen in optical tweezers. Predicting these forces is crucial for experiments and setup design. This study focuses on the geometrical optics regime, which applies to particles much larger than the light’s wavelength. In this model, a beam is represented by discrete rays that undergo multiple reflections and refractions, transferring momentum and angular momentum. However, the choice of ray discretization affects the balance between computational speed and accuracy. We demonstrate that neural networks overcome this limitation, enabling faster and even more precise simulations. Using an optically trapped spherical particle with an analytical solution as a benchmark, we validate our method and apply it to study complex systems that would otherwise be computationally hard.

Reference
David Bronte Ciriza, Alessandro Magazzù, Agnese Callegari, Gunther Barbosa, Antonio A. R. Neves, Maria A. Iatì, Giovanni Volpe, Onofrio M. Maragò, Faster and more accurate geometrical-optics optical force calculation using neural networks, ACS Photonics 10, 234–241 (2023)

Quantitative evaluation of methods to analyze motion changes in single-particle experiments published on Nature Communications

Rationale for the challenge organization. The interactions of biomolecules in complex environments, such as the cell membrane, regulate physiological processes in living systems. These interactions produce changes in molecular motion that can be used as a proxy to measure interaction parameters. Time-lapse single-molecule imaging allows us to visualize these processes with high spatiotemporal resolution and, in combination with single-particle tracking methods, provide trajectories of individual molecules. *(Image by the Authors of the manuscript.)*

Quantitative evaluation of methods to analyze motion changes in single-particle experiments
Gorka Muñoz-Gil, Harshith Bachimanchi, Jesús Pineda, Benjamin Midtvedt, Gabriel Fernández-Fernández, Borja Requena, Yusef Ahsini, Solomon Asghar, Jaeyong Bae, Francisco J. Barrantes, Steen W. B. Bender, Clément Cabriel, J. Alberto Conejero, Marc Escoto, Xiaochen Feng, Rasched Haidari, Nikos S. Hatzakis, Zihan Huang, Ignacio Izeddin, Hawoong Jeong, Yuan Jiang, Jacob Kæstel-Hansen, Judith Miné-Hattab, Ran Ni, Junwoo Park, Xiang Qu, Lucas A. Saavedra, Hao Sha, Nataliya Sokolovska, Yongbing Zhang, Giorgio Volpe, Maciej Lewenstein, Ralf Metzler, Diego Krapf, Giovanni Volpe, Carlo Manzo
Nature Communications 16, 6749 (2025)
arXiv: 2311.18100
doi: https://doi.org/10.1038/s41467-025-61949-x

The analysis of live-cell single-molecule imaging experiments can reveal valuable information about the heterogeneity of transport processes and interactions between cell components. These characteristics are seen as motion changes in the particle trajectories. Despite the existence of multiple approaches to carry out this type of analysis, no objective assessment of these methods has been performed so far. Here, we report the results of a competition to characterize and rank the performance of these methods when analyzing the dynamic behavior of single molecules. To run this competition, we implemented a software library that simulates realistic data corresponding to widespread diffusion and interaction models, both in the form of trajectories and videos obtained in typical experimental conditions. The competition constitutes the first assessment of these methods, providing insights into the current limitations of the field, fostering the development of new approaches, and guiding researchers to identify optimal tools for analyzing their experiments.

Deep-Learning Investigation of Vibrational Raman Spectra for Plant-Stress Analysis on ArXiv

Detecting stress in plants is crucial for both open-farm and controlled-environment agriculture. Biomolecules within plants serve as key stress indicators, offering vital markers for continuous health monitoring and early disease detection. Raman spectroscopy provides a powerful, non-invasive means to quantify these biomolecules through their molecular vibrational signatures. However, traditional Raman analysis relies on customized data-processing workflows that require fluorescence background removal and prior identification of Raman peaks of interest-introducing potential biases and inconsistencies. Here, we introduce DIVA (Deep-learning-based Investigation of Vibrational Raman spectra for plant-stress Analysis), a fully automated workflow based on a variational autoencoder. Unlike conventional approaches, DIVA processes native Raman spectra-including fluorescence backgrounds-without manual preprocessing, identifying and quantifying significant spectral features in an unbiased manner. We applied DIVA to detect a range of plant stresses, including abiotic (shading, high light intensity, high temperature) and biotic stressors (bacterial infections). By integrating deep learning with vibrational spectroscopy, DIVA paves the way for AI-driven plant health assessment, fostering more resilient and sustainable agricultural practices.

Xinwen Zhang joins the Soft Matter Lab

Xinwen Zhang started her PhD at the Physics Department of Gothenburg University on 7 July 2025.

Xinwen holds a master’s degree in Physics (biophysics) from the University of Science and Technology of China (USTC), Hefei, China.

During her PhD, she will focus on label-free optical microscopy combined with deep learning, aiming to characterize nanoparticles and uncover their physical mechanisms.

Seminar by G. Volpe and C. Manzo at CIG, Makerere University, Kampala, Uganda, 3 July 2025 (Online)

GAUDI leverages a hierarchical graph-convolutional variational autoencoder architecture, where an encoder progressively compresses the graph into a low-dimensional latent space, and a decoder reconstructs the graph from the latent embedding. *(Image by M. Granfors and J. Pineda.)*

Cutting Training Data Needs through Inductive Bias & Unsupervised Learning
Giovanni Volpe and Carlo Manzo
Computational Intelligence Group (CIG), Weekly Reading Session
Date: 3 July 2025
Time: 17:00
Place: Makerere University, Kampala, Uganda (Online)

Graphs provide a powerful framework for modeling complex systems, but their structural variability makes analysis and classification challenging. To address this, we introduce GAUDI (Graph Autoencoder Uncovering Descriptive Information), a novel unsupervised geometric deep learning framework that captures both local details and global structure. GAUDI employs an innovative hourglass architecture with hierarchical pooling and upsampling layers, linked through skip connections to preserve essential connectivity information throughout the encoding–decoding process. By mapping different realizations of a system — generated from the same underlying parameters — into a continuous, structured latent space, GAUDI disentangles invariant process-level features from stochastic noise. We demonstrate its power across multiple applications, including modeling small-world networks, characterizing protein assemblies from super-resolution microscopy, analyzing collective motion in the Vicsek model, and capturing age-related changes in brain connectivity. This approach not only improves the analysis of complex graphs but also provides new insights into emergent phenomena across diverse scientific domains.

Youtube: Global graph features unveiled by unsupervised geometric deep learning