Biology subjects

Gilson, M.

Publications and source records attributed to Gilson, M..

At least 19 recordsLinked to original sources

bioRxiv2026

Towards a scalable effective connectivity estimation in fMRI

Characterizing directed connections between brain regions is fundamental for understanding the brains network organization and how it supports cognitive processes. However, a reliable methodology allowing assessment of effective connectivity (EC) at the whole-brain level and across high-resolution parcellations has remained elusive. This limitation reflects a trade-off between model detail and scalability. Mechanistic approaches such as Dynamic Causal Modeling (DCM) provide a principled description of directed interactions but become impractical for large parcellations, whereas scalable methods based on linear-system inversion often fail to reliably recover key features of the underlying effective connectivity. Here, using data from the Human Connectome Project S1200 release, we show that the multivariate Ornstein-Uhlenbeck (mOU) approach for EC estimation can be adapted to yield EC estimates that are in broad agreement with those of an advanced DCM, at least when the spatial variability of the hemodynamic response falls within the physiological range characteristic of the healthy adult brain. Furthermore, thanks to its low computational cost, mOU can be easily scaled up to large neuroimaging cohorts and high-resolution parcellations. High-resolution, mOU-based EC outperforms functional connectivity for single-subject connectome fingerprinting. In addition, it enables a high-resolution characterization of the brains hierarchical organization and modular architecture by identifying directed interactions within and between communities and revealing driver and receiver regions. Overall, our methodology provides a reliable and scalable EC estimation pipeline of wide applicability to cognitive and network neuroscience.

neuroscience↗

bioRxiv2026

Functional and Structural Brain Age Gaps as Complementary Predictive Markers for Cognitive Performance

Cognitive decline is characterized by substantial inter-individual variability across the lifespan, even among clinically healthy adults. However, it is unclear how structural and functional brain organization jointly relate to differences in cognitive performance within the typical populations. We considered a multimodal brain age framework by predicting for each subject its age from either structural MRI or resting-state functional MRI, after training on the 1000BRAINS cohort (901 participants, 18-85 years). Following, a structural and a functional brain age gap markers were defined as the deviations with respect to the real chronological age. We then used the markers to characterize the interplay between structure and function in shaping cognitive performances. Our findings reveal that the functional marker is as important to predict global cognitive performance as the structural marker. The most accurate prediction was found for global cognition, executive functions and episodic memory, followed by working memory, attention and language. In particular, the structural marker was contributing to all domains, except for working memory, whereas the functional marker had the highest contribution. Overall, structural and functional brain age models together provide a meaningful subject-level summary of brain information for predicting cognition, supporting a multidimensional view of typical aging.

neuroscience↗

bioRxiv2024

ProteinAligner: A Multi-modal Pretraining Framework for Protein Foundation Models

Protein foundation models, particularly protein language models, have demonstrated strong success in learning meaningful representations of proteins using transformer architectures pretrained on large-scale protein datasets with self-supervised learning. These representations have been highly effective for downstream tasks such as predicting protein functions and properties. However, most current protein foundation models focus on pretraining with amino acid sequences, often neglecting additional modalities like protein structures and related literature, both of which provide valuable insights. To address this gap, we propose a multi-modal pretraining approach that integrates three key modalities - protein sequences, structures, and literature text. In our framework, the protein sequence modality serves as the anchor, with the other two modalities aligned to it, enhancing the models capacity to capture more comprehensive protein information. ProteinAligner out-performed state-of-the-art protein foundation models in predicting protein functions and properties across diverse down-stream tasks.

bioinformatics↗

bioRxiv2024

Multi-Modal Large Language Model Enables All-Purpose Prediction of Drug Mechanisms and Properties

Accurately predicting the mechanisms and properties of potential drug molecules is essential for advancing drug discovery. However, traditional methods often require the development of specialized models for each specific prediction task, resulting in inefficiencies in both model training and integration into work-flows. Moreover, these approaches are typically limited to predicting pharmaceutical attributes represented as discrete categories, and struggle with predicting complex attributes that are best described in free-form texts. To address these challenges, we introduce DrugChat, a multi-modal large language model (LLM) designed to provide comprehensive predictions of molecule mechanisms and properties within a unified framework. DrugChat analyzes the structure of an input molecule along with users queries to generate comprehensive, free-form predictions on drug indications, pharmacodynamics, and mechanisms of action. Moreover, DrugChat supports multi-turn dialogues with users, facilitating interactive and in-depth exploration of the same molecule. Our extensive evaluation, including assessments by human experts, demonstrates that DrugChat significantly outperforms GPT-4 and other leading LLMs in generating accurate free-form predictions, and exceeds state-of-the-art specialized prediction models.

pharmacology and toxicology↗

bioRxiv2024

Directed neural interactions in fMRI: a comparison between Granger Causality and Effective Connectivity

A key challenge for network neuroscience is to understand the role of interactions between brain regions and how they contribute to the encoding and broadcasting of information within cognitive processes. This requires computational tools to infer directional relations between brain regions from neural time series. For fMRI, the most common approaches are based on Granger causality (GC) analysis and effective connectivity (EC) models. Despite their different conceptual framing, GC and EC models for fMRI are based on similar mathematical assumptions, grounded on continuous- and discrete-time linear stochastic models. Based on a mapping between multivariate Ornstein-Uhlenbeck (MOU) and multivariate autoregressive (MVAR) processes, we analytically obtain an approximately quadratic relation between EC and GC, after rescaling the GC to compensate for unequal noise variances of the source and target. Simulations show that these relations can be observed in finite time series only if a large amount of data is available, implying that they may emerge only at a group level in real fMRI experiments. We verified this prediction by systematically comparing EC and GC in a large-scale fMRI data set from the Human Connectome Project. Overall, our findings can provide methodological and interpretational guidance in the usage of GC and EC for brain network reconstruction by clearly elucidating what is common between the two methods, but also their respective specificities and biases.

neuroscience↗

bioRxiv2023

Identifying the seizure onset zone with intracranial EEG: A systematic comparison between power and connectivity-based biomarkers

BackgroundIctal stereo-encephalography (sEEG) biomarkers for seizure onset zone (SOZ) localization can be classified depending on whether they target abnormalities in signal power or functional connectivity between signals, and they may depend on the frequency band and time window at which they are estimated. New methodThis work aimed to compare and optimize the performance of a power and a connectivity-based biomarkers to identify SOZ contacts from ictal sEEG recordings. To do so, we used a previously introduced power-based measure, the normalized mean activation (nMA), which quantifies the ictal average power activation. Similarly, we defined the normalized mean strength (nMS), to quantify the ictal mean functional connectivity of every contact with the rest. The optimal frequency bands and time windows were selected based on optimizing AUC and F2-score. ResultsThe analysis was performed on a dataset of 67 seizures from 10 patients with pharmacoresistant temporal lobe epilepsy. Our results suggest that the power-based biomarker generally performs better for the detection of SOZ than the connectivity-based one. However, an equivalent performance level can be achieved when both biomarkers are independently optimized. Optimal performance was achieved in the beta and lower-gamma range for the power biomarker and in the higher-gamma range for connectivity, both using a 30 s period after seizure onset. ConclusionsThe results of this study highlight the importance of this optimization step over frequency and time windows when comparing different SOZ discrimination biomarkers. This information should be considered when training SOZ classifiers on retrospective patients data for clinical applications.

neuroscience↗

bioRxiv2023

Phase-dependent stimulation response is shaped by the brain's transient collective dynamics.

Exogenous stimulation is a promising tool for investigating and altering cognitive processes in the brain, with potential clinical applications. Following experimental observations, we hypothesise that the effect of stimulation crucially depends on the endogenous dynamics of the brain. Our study explores how local and global dynamical properties, like the stimulation phase of regional oscillatory activity and the transient network states, modulate the effect of single pulse stimulation in a large-scale network. Our findings demonstrate that the effect of stimulation strongly depends on the interplay between stimulated phase, transient network state, and brain region. Importantly, we show that stimulation is not only state-dependent but can also induce global state switching. Lastly, predicting the effect of stimulation by using machine learning shows that state-aware measures can increase the performance by up to 40%. Our results suggest that a fine characterisation of the complex brain dynamics in experimental setups is essential for improving the reliability of exogenous stimulation.

neuroscience↗

bioRxiv2023

Structure-Activity Relationships, Tolerability and Efficacy of Microtubule-Active 1,2,4-Triazolopyrimidines as Potential Candidates to Treat Human African Trypanosomiasis

Tubulin and microtubules (MTs) are potential protein targets to treat parasitic infections and our previous studies have shown that the triazolopyrimidine (TPD) class of MT- active compounds hold promise as antitrypanosomal agents. MT-targeting TPDs include structurally related but functionally diverse congeners that interact with mammalian tubulin at either one or two distinct interfacial binding sites; namely, the seventh and vinca sites, which are found within or between ,{beta}-tubulin heterodimers, respectively. Evaluation of the activity of 123 TPD congeners against cultured Trypanosoma brucei enabled a robust quantitative structure-activity relationship (QSAR) model and the prioritization of two congeners for in vivo pharmacokinetics (PK), tolerability and efficacy studies. Treatment of T. brucei-infected mice with tolerable doses of TPDs 3 and 4 significantly decreased blood parasitemia within 24 h. Further, two once-weekly doses of 4 at 10 mg/kg significantly extended the survival of infected mice relative to infected animals treated with vehicle. Further optimization of dosing and/or the dosing schedule of these CNS-active TPDs may provide alternative treatments for human African trypanosomiasis.

microbiology↗

bioRxiv2021

Posterior integration and thalamo-frontotemporal broadcasting are impaired in disorders of consciousness

The study of the brains static and dynamical activity is opening a valuable source of assistance for the clinical assessment of patients with disorders of consciousness. For example, glucose uptake and dysfunctional spread of naturalistic and synthetic stimuli has proven useful to characterize hampered consciousness. However, understanding of the mechanisms behind loss of consciousness following brain injury is still missing. Here, we study the propagation of endogenous and in-silico exogenous perturbations in patients with disorders of consciousness, based upon directed and causal interactions estimated from resting-state fMRI. We found that patients with disorders of consciousness suffer decreased capacity for neural propagation and responsiveness to events, and that this can be related to glucose metabolism as measured with [18F]FDG-PET. In particular, we show that loss of consciousness is related to the malfunctioning of two neural circuits: the posterior cortical regions failing to convey information, in conjunction with reduced broadcasting of information from subcortical, temporal, parietal and frontal regions. These results seed light on the mechanisms behind disorders of consciousness, triangulating network function with basic measures of brain integrity and behavior. HighlightsO_LIPropagation of neural events and network responses are disrupted in patients with DoC. C_LIO_LILoss of consciousness is related to the malfunctioning of two neural circuits. C_LIO_LIPosterior cortical regions lack to integrate information in altered consciousness. C_LIO_LIBreakdown of information broadcasting of subcortical cortical areas in DoC. C_LIO_LILoss of network responses in DoC patients is related to glucose metabolism. C_LI

neuroscience↗

bioRxiv2021

Meditation-induced effects on whole-brain structural and effective connectivity

In the past decades, there has been a growing scientific interest in characterizing neural correlates of meditation training. Nonetheless, the mechanisms underlying meditation remain elusive. In the present work, we investigated meditation-related changes in structural and functional connectivities (SC and FC, respectively). For this purpose, we scanned experienced meditators and control (naive) subjects using magnetic resonance imaging (MRI) to acquire structural and functional data during two conditions, resting-state and meditation (focused attention on breathing). In this way, we aimed to characterize and distinguish both short-term and long-term modifications in the brains structure and function. First, we performed a network-based analysis of anatomical connectivity. Then, to analyze the fMRI data, we calculated whole-brain effective connectivity (EC) estimates, relying on a dynamical network model to replicate BOLD signals spatio-temporal structure, akin to FC with lagged correlations. We compared the estimated EC, FC, and SC links as features to train classifiers to predict behavioral conditions and group identity. The whole-brain SC analysis revealed strengthened anatomical connectivity across large-scale networks for meditators compared to controls. We found that differences in SC were reflected in the functional domain as well. We demonstrated through a machine-learning approach that EC features were more informative than FC and SC solely. Using EC features we reached high performance for the condition-based classification within each group and moderately high accuracies when comparing the two groups in each condition. Moreover, we showed that the most informative EC links that discriminated between meditators and controls involved the same large-scale networks previously found to have increased anatomical connectivity. Overall, the results of our whole-brain model-based approach revealed a mechanism underlying meditation by providing causal relationships at the structure-function level.

neuroscience↗

bioRxiv2021

Covariance-based information processing in reservoir computing systems

In biological neuronal networks, information representation and processing are achieved through plasticity learning rules that have been empirically characterized as sensitive to second and higher-order statistics in spike trains. However, most models in both computational neuroscience and machine learning aim to convert diverse statistical properties in inputs into first-order statistics in outputs, like in modern deep learning networks. In the context of classification, such schemes have merit for inputs like static images, but they are not well suited to capture the temporal structure in time series. In contrast, the recently developed covariance perceptron uses second-order statistics by mapping input covariances to output covariances in a consistent fashion. Here, we explore the applicability of covariance-based perceptron readouts in reservoir computing networks to classify synthetic multivariate time series structured at different statistical orders (first and second). We show that the second-order framework outperforms or matches the classical mean paradigm in terms of accuracy. Our results highlight a nontrivial relationship between input and reservoir properties in generating the output reservoir activity, which suggests an important role for recurrent connectivity in transforming information representations in biologically inspired architectures. Finally, we solve a speech recognition task for the classification of spoken digits to further demonstrate the potential of covariance-based decoding for real data.

neuroscience↗

bioRxiv2020

Prefrontal Transcranial Direct Current Stimulation globally improves learning, but does not selectively potentiate the benefits of Targeted Memory Reactivation on awake memory consolidation

Targeted Memory Reactivation (TMR) and transcranial Direct Current Stimulation (tDCS) can enhance memory consolidation. It is currently unknown whether TMR reinforced by simultaneous tDCS has superior efficacy. In this study, we investigated the complementary effect of TMR and bilateral tDCS on the consolidation of emotionally neutral and negative declarative memories. Participants learned neutral and negative word pairs. Each word pair was presented with an emotionally compatible sound. Following learning, participants spent a 20-minutes retention interval awake under 4 possible conditions: (1) TMR alone (i.e. replay of 50% of the associated sounds), (2) TMR combined with anodal stimulation of the left DLPFC, (3) TMR combined with anodal stimulation of the right DLPFC and (4) TMR with sham tDCS. Results evidenced selective memory enhancement for the replayed stimuli in the TMR-only and TMR-sham conditions, which confirms a specific effect of TMR on memory. However, memory was enhanced at higher levels for all learned items (irrespective of TMR) in the TMR-anodal right and TMR-anodal left tDCS conditions, suggesting that the beneficial effects of tDCS overshadow the specific effects of TMR. Emotionally negative memories were not modulated by tDCS hemispheric polarity. We conclude that electrical stimulation of the DLPFC during post-learning period globally benefits memory consolidation, but does not potentiate the specific benefits of TMR.

neuroscience↗

bioRxiv2019

Whole-genome and RNA sequencing reveal variation and transcriptomic coordination in the developing human prefrontal cortex

Variation in gene expression underlies neurotypical development, while genomic variants contribute to neuropsychiatric disorders. BrainVar is a unique resource of paired whole-genome sequencing and bulk-tissue RNA-sequencing from the human dorsolateral prefrontal cortex of 176 neurotypical individuals across prenatal and postnatal development, providing the opportunity to assay genomic and transcriptomic variation in tandem. Leveraging this resource, we identified rare premature stop codons with commensurate reduced and allele-specific expression of corresponding genes, and common variants that alter gene expression (expression quantitative trait loci, eQTLs). Categorizing eQTLs by prenatal and postnatal effect, genes affected by temporally-specific eQTLs, compared to constitutive eQTLs, are enriched for haploinsufficiency, protein-protein interactions, and neuropsychiatric disorder risk loci. Expression levels of over 12,000 genes rise or fall in a concerted late-fetal transition, with the transitional genes enriched for cell type specific genes and neuropsychiatric disorder loci, underscoring the importance of cataloguing developmental trajectories in understanding cortical physiology and pathology.\n\nHighlightsO_LIWhole-genome and RNA-sequencing across human prefrontal cortex development in BrainVar\nC_LIO_LIGene-specific developmental trajectories characterize the late-fetal transition\nC_LIO_LIIdentification of constitutive, prenatal-specific, postnatal-specific, and rare eQTLs\nC_LIO_LIIntegrated analysis reveals genetic and developmental influences on CNS traits and disorders\nC_LI

neuroscience↗

bioRxiv2019

The covariance perceptron: A new framework for classification and processing of time series in recurrent neural networks

Learning in neuronal networks has developed in many directions, in particular to reproduce cognitive tasks like image recognition and speech processing. Implementations have been inspired by stereotypical neuronal responses like tuning curves in the visual system, where, for example, ON/OFF cells fire or not depending on the contrast in their receptive fields. Classical models of neuronal networks therefore map a set of input signals to a set of activity levels in the output of the network. Each category of inputs is thereby predominantly characterized by its mean. In the case of time series, fluctuations around this mean constitute noise in this view. For this paradigm, the high variability exhibited by the cortical activity may thus imply limitations or constraints, which have been discussed for many years. For example, the need for averaging neuronal activity over long periods or large groups of cells to assess a robust mean and to diminish the effect of noise correlations. To reconcile robust computations with variable neuronal activity, we here propose a conceptual change of perspective by employing variability of activity as the basis for stimulus-related information to be learned by neurons, rather than merely being the noise that corrupts the mean signal. In this new paradigm both afferent and recurrent weights in a network are tuned to shape the input-output mapping for covariances, the second-order statistics of the fluctuating activity. When including time lags, covariance patterns define a natural metric for time series that capture their propagating nature. We develop the theory for classification of time series based on their spatio-temporal covariances, which reflect dynamical properties. We demonstrate that recurrent connectivity is able to transform information contained in the temporal structure of the signal into spatial covariances. Finally, we use the MNIST database to show how the covariance perceptron can capture specific second-order statistical patterns generated by moving digits. Author summaryThe dynamics in cortex is characterized by highly fluctuating activity: Even under the very same experimental conditions the activity typically does not reproduce on the level of individual spikes. Given this variability, how then does the brain realize its quasi-deterministic function? One obvious solution is to compute averages over many cells, assuming that the mean activity, or rate, is actually the decisive signal. Variability across trials of an experiment is thus considered noise. We here explore the opposite view: Can fluctuations be used to actually represent information? And if yes, is there a benefit over a representation using the mean rate? We find that a fluctuation-based scheme is not only powerful in distinguishing signals into several classes, but also that networks can efficiently be trained in the new paradigm. Moreover, we argue why such a scheme of representation is more consistent with known forms of synaptic plasticity than rate-based network dynamics.

neuroscience↗

bioRxiv2019

MOU-EC: model-based whole-brain effective connectivity to extract biomarkers for brain dynamics from fMRI data and study distributed cognition

Neuroimaging techniques are increasingly used to study brain cognition in humans. Beyond their individual activation, the functional associations between brain areas have become a standard proxy to describe how information is distributed across the brain network. Among the many analysis tools available, dynamic models of brain activity have been developed to overcome the limitations of original connectivity measures such as functional connectivity. In particular, much effort has been devoted to the assessment of directional interactions between brain areas from their observed activity. This paper summarizes our recent approach to analyze fMRI data based on our whole-brain effective connectivity referred to as MOU-EC, while discussing the pros and cons of its underlying assumptions with respect to other established approaches. Once tuned, the model provides a connectivity measure that reflects the dynamical state of BOLD activity obtained using fMRI, which can be used to explore the brain cognition. We focus on two important applications. First, as a connectivity measure, MOU-EC can be used to extract biomarkers for task-specific brain coordination, understood as the patterns of areas exchanging information. The multivariate nature of connectivity measures raises several challenges for whole-brain analysis, for which machine-learning tools presents some advantages over statistical testing. Second, we show how to interpret changes in MOU-EC connections in a collective and model-based manner, bridging with network analysis. To illustrate our framework, we use a dataset where subjects were recorded in two conditions, watching a movie and a black screen (referred to as rest). Our framework provides a comprehensive set of tools that open exciting perspectives for the study of distributed cognition, as well as neuropathologies.

neuroscience↗

bioRxiv2018

Comparing Task-Relevant Information Across Different Methods of Extracting Functional Connectivity

The concept of brain states, functionally relevant large-scale activity patterns, has become popular in neuroimaging. Not all components of such patterns are equally characteristic for each brain state, but machine learning provides a possibility for extracting and comparing the structure of brain states from functional data. However, their characterization in terms of functional connectivity measures varies widely, from cross-correlation to phase coherence, and the idea that different measures provide similar or coherent information is a common assumption made in neuroimaging. Here, we compare the brain state signatures extracted from of phase coherence, pairwise covariance, correlation, regularized covariance and regularized precision for a dataset of subjects performing five different cognitive tasks. In addition, we compare the classification performance in identifying the tasks for each connectivity measure. The measures are evaluated in their ability to discriminate the five tasks with two types of cross-validation: within-subject cross-validation, which reflects the stability of the signature over time; and between-subject cross-validation, which aims at extracting signatures that generalize across subjects. Secondly, we compare the informative features (connections or links between brain regions/areas) across measures to test the assumption that similar information is obtained about brain state signatures from different connectivity measures. In our results, the different types of cross-validation give different classification performance and emphasize that functional connectivity measures on fMRI require observation windows of sufficient duration. Furthermore, we find that informative links for the classification, meaning changes between tasks that are consistent across subjects, are entirely uncorrelated between BOLD correlations and covariances. These results indicate that the corresponding FC signature can strongly differ across FC methods used and that interpretation is subject to caution in terms of subnetworks related to a task.

neuroscience↗

bioRxiv2018

Continuous Evaluation of Ligand Protein Predictions: A Weekly Community Challenge for Drug Docking

1Docking calculations can be used to accelerate drug discovery by providing predictions of the poses of candidate ligands bound to a targeted protein. However, studies in the literature use varied docking methods, and it is not clear which work best, either in general or for specific protein targets. In addition, a complete docking calculation requires components beyond the docking algorithm itself, such as preparation of the protein and ligand for calculations, and it is difficult to isolate which aspects of a method are most in need of improvement. To address such issues, we have developed the Continuous Evaluation of Ligand Protein Predictions (CELPP), a weekly blinded challenge for automated docking workflows. Participants in CELPP create a workflow to predict protein-ligand binding poses, which is then tasked with predicting 10-100 new (never before released) protein-ligand crystal structures each week. CELPP evaluates the accuracy of each workflows predictions and posts the scores online. CELPP is a new cyberinfrastructure resource to identify the strengths and weaknesses of current approaches, help map docking problems to the algorithms most likely to overcome them, and illuminate areas of unmet need in structure-guided drug design.

biophysics↗

bioRxiv2018

Analysis of brain network dynamics estimated from fMRI data: A new framework based on communicability and flow

Neuroimaging techniques such as MRI have been widely used to explore the associations between brain areas. Structural connectivity (SC) captures the anatomical pathways across the brain and functional connectivity (FC) measures the correlation between the activity of brain regions. These connectivity measures have been much studied using network theory in order to uncover the distributed organization of brain structures, in particular FC for task-specific brain communication. However, the application of network theory to study FC matrices is often "static" despite the dynamic nature of time series obtained from fMRI. The present study aims to overcome this limitation by introducing a network-oriented analysis applied to whole-brain effective connectivity (EC) useful to interpret the brain dynamics. Technically, we tune a multivariate Ornstein-Uhlenbeck (MOU) process to reproduce the statistics of the whole-brain resting-state fMRI signals, which provides estimates for MOU-EC as well as input properties (similar to local excitabilities). The network analysis is then based on the Green function (or network impulse response) that describes the interactions between nodes across time for the estimated dynamics. This model-based approach provides time-dependent graph-like descriptor, named communicability, that characterize the roles that either nodes or connections play in the propagation of activity within the network. They can be used at both global and local levels, and also enables the comparison of estimates from real data with surrogates (e.g. random network or ring lattice). In contrast to classical graph approaches to study SC or FC, our framework stresses the importance of taking the temporal aspect of fMRI signals into account. Our results show a merging of functional communities over time (in which input properties play a role), moving from segregated to global integration of the network activity. Our formalism sets a solid ground for the analysis and interpretation of fMRI data, including task-evoked activity.

neuroscience↗