Biology subjects

Blanc, A.

Publications and source records attributed to Blanc, A..

15 recordsLinked to original sources

bioRxiv2026

MIRA: an open source and user-friendly software to automate counting and sizing of fungal spores

BackgroundThe quantification of fungal spores constitutes a fundamental metric in phytopathology, serving as the primary variable for inoculum standardization and being used as a proxy for disease severity. Historically, spore quantification has relied on manual hemocytometry, which remains the most precise counting process to date, where chambers such as the Malassez slide are used to count a subsample of the inoculum. However, this method applied manually is highly labor-intensive, time-consuming, and can be prone to operator-dependent variability. To overcome these limitations, we introduce MIRA (Microscopy Image Recognition & Analysis), a novel open-source software integrating You Only Look Once (YOLO) deep learning algorithms. Featuring a user-friendly graphical interface, MIRA is adaptable to multiple camera systems and supports advanced object detection models, including YOLOv11 and YOLOv26. ResultsWe demonstrate that MIRA can be used to accurately detect and count spores from several phytopathogenic fungi, automatically measure spore surface area, and to differentiate spores across different genera. In an exhaustive comparative analysis using Pyricularia oryzae spores as an example, MIRA was benchmarked against manual gold-standard counting slides (Malassez and Kova) and indirect spectrophotometric methods (SPARK). The P. oryzae model loaded via MIRA achieved a strong correlation (R = 0.96) with manual gold standards while reducing processing time by over 90% for high-concentration samples (10 spores/mL). Beyond this benchmark, we also successfully tested specific YOLO models designed to recognize macro- and microconidia of Fusarium oxysporum f. sp. cubense, a model for Pseudocercospora fijiensis, and a single multiclass model capable of identifying six different rice pathogenic fungi. We provide comprehensive tutorials for operating the software and training custom detection models for free using Roboflow and Google Colab. MIRA is available both as open-source Python code and as standalone executables for Windows and Linux. ConclusionsMIRA provides a rapid, accurate, and highly reproducible alternative to manual spore counting, effectively removing a major bottleneck in phytopathology workflows. By combining advanced YOLO-based deep learning with an accessible interface and comprehensive training resources, MIRA makes accessible automated image analysis for researchers without programming expertise. Moreover, MIRA drastically improves the efficiency of high-throughput disease phenotyping and can be adapted for a wide range of microscopic quantification tasks across various biological disciplines.

plant biology↗

bioRxiv2025

A role for prefrontal dopamine signaling in control of goal-directed actions

Impairments in behavioral flexibility commonly found across psychiatric disorders have often been attributed to a dysfunction of the medial prefrontal cortex (mPFC), notably in relation with altered dopamine transmission. However, how dopamine transmission shapes neuronal activity in the mPFC to allow adapting to a changing situation remains largely elusive. Here, we show that dopamine dynamics in the mPFC encode the value of instrumental actions, in particular during reversal learning. Such signal shapes the activity of mPFC dopaminoceptive neurons during reversal learning through the recruitment of heteromers formed by dopamine D1 and D2 receptors with NMDA receptors. In accordance, blockade of either D1/NMDA or D2/NMDA heteromers in the mPFC selectively impairs reversal of outcome identities but not learning and expression of initial action-outcome associations. Our findings provide mechanistic evidence for a central role of dopamine in the mPFC to allow updating goal-directed actions.

neuroscience↗

bioRxiv2024

Experimental Phage Evolution Results in Expanded Host Ranges Against MDR and XDR Klebsiella pneumoniae Isolates

Resistance to antibiotics is approaching crisis levels for organisms such as the ESKAPEE pathogens (includes Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Enterobacter spp., and Escherichia coli) that often are acquired in hospitals. These organisms sometimes have acquired plasmids that confer resistance to most if not all beta-lactam antibiotics such as those produced by Carbapenem Resistant Enterobacterales (CREs). We have been developing alternative means for dealing with antibiotic resistant microbes that cause infections in humans by developing viruses (bacteriophages) that attack and kill them. We have been working with one of the ESKAPEE pathogens, K. pneumoniae, that has one of the highest propensities for antimicrobial resistance, to develop phages that target and kill it. We identified a number of phages that have lytic capacity against only a few clinical isolates, and through experimental evolution over the course of 30 days, were able to vastly expand the host ranges of these phages to kill a broader range of clinical K. pneumoniae isolates including MDR (multi-drug resistant) and XDR (extensively-drug resistant) isolates. Most interestingly, they were capable of inhibiting growth of clinical isolates both on solid and in liquid medium over extended periods. That we were able to extend the host ranges of multiple naive MDR and XDR K. pneumoniae through experimental phage evolution suggests that such a technique may be applicable to other antibiotic-resistant organisms to help stem the tide of antibiotic resistance and offer further options for medical treatments. ImportanceBacterial pathogens are becoming greater threats given the rise in antibiotic resistance, where traditional therapies may no longer work to cure some infections. Chief amongst these multidrug resistant infections (MDR) and extensively drug-resistant infections (XDR) is Klebsiella pneumoniae, which is known to sometimes harbor genetic elements that render it incredibly difficult to treat with conventional antibiotics. Treatments like bacteriophages have not had much success against such pathogens because resistance to the phages used often develops rapidly. We adapted a co-evolutionary technique to develop K. pneumoniae phages to be highly active longitudinally against K. pneumoniae clinical isolates. In as few as 30 days, we were able to vastly expand the host ranges of K. pneumoniae phages against MDR and XDR clinical isolates and that maintain their infectivity over clinically relevant time periods. By adapting these established techniques to clinical MDR and XDR K. pneumoniae isolates, we believe we can establish similar techniques for expanding phage host ranges against most antibiotic-resistant bacteria. As such, phages can be viable alternatives to antibiotics when antibiotic resistance exists in hospitals and communities.

microbiology↗

bioRxiv2024

Identification of a large cohort of Enterobacter jumbo phages with broad host ranges across pathogenic Gammaproteobacteria

ESKAPE pathogens cause most hospital-acquired infections globally and often carry antibiotic resistance. Many of them have been the target of bacteriophage therapies. Enterobacter is an ESKAPE pathogen but is less frequently a target for phage therapy due to a relative lack of available phages. We isolated eight jumbo phages with genomes ranging from 223 to 366 kbp targeting Enterobacter spp. and found that they belonged to separate phage clades. Six of them formed nucleus-like structures confirmed by DAPI-staining, and were phylogenetically related to Chimalliviridae. Two jumbo phages did not form nucleus-like structures and did not cluster with Chimalliviridae. Although these jumbo phages were found on Enterobacter, many were closely related to phages with non-Enterobacter hosts. To test whether these phages may have had expanded host ranges, we examined 14 pathogenic Gammaproteobacteria and found that these phages were capable of creating plaques on 8 of them. These species included Escherichia coli, Klebsiella aerogenes, Serratia marcescens, Salmonella spp., Shigella spp., Providencia spp., Citrobacter spp., and Cronobacter sakazakii. We verified that there was phage amplification in these microbes rather than lysis from without by performing qPCR to confirm DNA replication in each species. Phages typically have narrow host ranges, a benefit for microbiome-sparing compared to antibiotics. However, the broad host ranges of these Gammaproteobacteria jumbo phages suggests that not all phages have the same risk/benefit ratios. While this broad range could aid their development as antibiotic alternatives, further study is needed to assess potential microbiome disruption. SignificanceWith the growing threat of antibiotic resistant bacteria, alternative treatments like bacteriophages have emerged. Bacteriophages typically have narrow host ranges, a disadvantage compared to antibiotics. We discovered eight jumbo phages that kill antibiotic resistant Enterobacter, but were diverse phylogenetically. Six of them formed nucleus-like structures and were members of the Chimalliviridae family and the other two did not form nucleus-like structures. We confirmed that each phage had broad host ranges capable of lysing at least eight different human Gammaproteobacteria including pathogens such as E. coli, K. aerogenes, and S. marcescens. By identifying broad host range jumbo phages that attack pathogens, we may have identified phages with spectrums of activity more similar to antibiotics than have been traditionally attributed to phages.

microbiology↗

bioRxiv2024

Investigating the genetic architecture of biotic stress response in stone fruit tree orchards under natural infections with a multi-environment GWAS approach

The mapping and introduction of sustainable plant immunity to pests and diseases in fruit tree is still a major challenge in modern breeding. This study aims at deciphering the genetic architecture underlying resistance or tolerance across environments for major pests and diseases in peach (P. persica) and apricot (P. armeniaca). We set up a multi-environment trial (MET) approach by studying two core collections of 206 peach and 150 apricot accessions deployed under low phytosanitary conditions in respectively three and two environmentally contrasted locations in South-East of France. To capture the complex dynamics of pest and disease spread in naturally infected orchards, visual scoring of symptoms was repeated within and between 3 years, for five and two pests and diseases respectively for peach and apricot, resulting in the maximum of damage score and the AUDPC. These traits were used as phenotypic inputs in our genome-wide association studies (GWAS) strategy, and leading to the identification of: i) non-additive genotype-phenotype associations, ii) environment-shared QTLs iii) environment-specific QTLs, and iv) interactive QTLs which changes in direction ( antagonist) or intensity ( differential) according to the environment. By conducting GWAS with multiple methods, we successfully identified a total of 60 high confidence QTLs, leading to the identification of 87 candidate genes, the majority belonging to the Leucine-rich repeat containing receptors (LRR-CRs) family gene. Finally, we provided a comparative analysis of our results on peach and apricot, two closely related species. The present results contribute to the development of genomics-assisted breeding to improve biotic resilience in Prunus varieties.

plant biology↗

bioRxiv2024

Enhanced Suppression of Stenotrophomonas maltophilia by a Three-Phage Cocktail: Genomic Insights and Kinetic Profiling

In our era of rising antibiotic resistance, Stenotrophomonas maltophilia (STM) is an understudied, gram-negative, aerobic bacterium widespread in the environment and increasingly causing opportunistic infections. Treating STM infections remains difficult, leading to an increase in disease severity and higher hospitalization rates in people with Cystic Fibrosis (pwCF), cancer, and other immunocompromised health conditions. The lack of effective antibiotics has led to renewed interest in phage therapy; however, there is a need for well-characterized phages. In response to an oncology patient with a respiratory infection, we collected 18 phages from Southern California wastewater influent that exhibit different plaque morphology against STM host strain B28B, cultivated from a blood sample. Here, we characterize the genomes and life cycle kinetics of our STM phage collection. We hypothesize that genetically distinct phages give rise to unique lytic life cycles that can enhance bacterial killing when combined into a phage cocktail compared to the individual phages alone. We identified three genetically distinct clusters of phages, and a representative from each group was screened for potential therapeutic use and investigated for infection kinetics. The results demonstrated that the three-phage cocktail significantly suppressed bacterial growth compared to individual phages when observed for 48 hours. We also assessed the lytic impacts of our three-phage cocktail against a collection of 46 STM strains to determine if a multi-phage cocktail can expand the host range of individual phages. Our phages remained strain-specific and infect >50% of tested strains. The multi-phage cocktail maintains bacterial growth suppression and prevents the emergence of phage-resistant strains throughout our 40-hour assay. These findings suggest specialized phage cocktails may be an effective avenue of treatment for recalcitrant STM infections resistant to current antibiotics. IMPORTANCEPhage therapy could provide a vital strategy in the fight against antimicrobial resistance (AMR) bacterial infections; however, significant knowledge gaps remain. This study investigates phage cocktail development for the opportunistic pathogen Stenotrophomonas maltophilia (STM). Our findings contribute novel phages, their lytic characteristics, and limitations when exposed to an array of clinically relevant STM strains. Eighteen bacteriophages were isolated from wastewater influent from Escondido, California, and subjected to genomic analysis. We investigated genetically distinct phages to establish their infection kinetics and developed them into a phage cocktail. Our findings suggest that a genetically distinct STM phage cocktail provides an effective strategy for bacterial suppression of host strain B28B and five other clinically relevant STM strains. Phage therapy against STM remains poorly understood, as only 39 phages have been previously isolated. Future research into the underlying mechanism of how phage cocktails overwhelm the host bacteria will provide essential information that could aid in optimizing phage applications and impact alternative treatment options.

microbiology↗

bioRxiv2024

In-vitro resensitization of multidrug resistant clinical isolates of Enterococcus faecium and faecalis through phage-antibiotic synergy

Bacteriophages are an increasingly attractive option for the treatment of antibiotic resistant infections, but their efficacy is difficult to discern due to confounding effects of antibiotics. Phages are generally delivered in conjunction with antibiotics, and thus, when patients improve its unclear whether the phages, antibiotics or both are responsible. This question is particularly relevant for enterococcus infections, as limited data suggest phages might restore antibiotic efficacy against resistant strains. Enterococci can develop high-level resistance to vancomycin, a primary treatment. We assessed clinical and laboratory isolates of Enterococcus faecium and Enterococcus faecalis to determine whether we could observe synergistic interactions between phages and antibiotics. We identified synergy between multiple phages and antibiotics including linezolid, ampicillin, and vancomycin. Notably, antibiotic susceptibility did not predict synergistic interactions with phages. Vancomycin resistant isolates (n=6) were eradicated by the vancomycin-phage combination as effectively as vancomycin susceptible isolates (n=2). Transcriptome analysis revealed significant gene expression changes under antibiotic-phage conditions, especially for linezolid and vancomycin, with upregulated genes involved in nucleotide and protein biosynthesis and downregulated stress response and prophage-related genes. While our results do not conclusively determine the etiology of the observed synergistic interactions between antibiotics and phages, they do confirm and build upon previous research that observed these synergistic interactions. Our work highlights how using phages can restore the effectiveness of vancomycin against resistant isolates. This finding provides a promising, although unexpected, strategy for moving forward with phage treatments for Vancomycin Resistant Enterococcus infections.

microbiology↗

bioRxiv2024

Statistical signature of subtle behavioural changes inlarge-scale behavioural assays

The central nervous system can generate various behaviours, including motor responses, which we can observe through video recordings. Recent advancements in genetics, automated behavioural acquisition at scale, and machine learning enable us to link behaviours to their underlying neural mechanisms causally. Moreover, in some animals, such as the Drosophila larva, this mapping is possible at unprecedented scales of millions of animals and single neurons, allowing us to identify the neural circuits generating particular behaviours. These high-throughput screening efforts are invaluable, linking the activation or suppression of specific neurons to behavioural patterns in millions of animals. This provides a rich dataset to explore how diverse nervous system responses can be to the same stimuli. However, challenges remain in identifying subtle behaviours from these large datasets, including immediate and delayed responses to neural activation or suppression, and understanding these behaviours on a large scale. We introduce several statistically robust methods for analyzing behavioural data in response to these challenges: 1) A generative physical model that regularizes the inference of larval shapes across the entire dataset. 2) An unsupervised kernel-based method for statistical testing in learned behavioural spaces aimed at detecting subtle deviations in behaviour. 3) A generative model for larval behavioural sequences, providing a benchmark for identifying complex behavioural changes. 4) A comprehensive analysis technique using suffix trees to categorize genetic lines into clusters based on common action sequences. We showcase these methodologies through a behavioural screen focused on responses to an air puff, analyzing data from 280,716 larvae across 568 genetic lines. Author SummaryThere is a significant gap in understanding between the architecture of neural circuits and the mechanisms of action selection and behaviour generation.Drosophila larvae have emerged as an ideal platform for simultaneously probing behaviour and the underlying neuronal computation [1]. Modern genetic tools allow efficient activation or silencing of individual and small groups of neurons. Combining these techniques with standardized stimuli over thousands of individuals makes it possible to relate neurons to behaviour causally. However, extracting these relationships from massive and noisy recordings requires the development of new statistically robust approaches. We introduce a suite of statistical methods that utilize individual behavioural data and the overarching structure of the behavioural screen to deduce subtle behavioural changes from raw data. Given our studys extensive number of larvae, addressing and preempting potential challenges in body shape recognition is critical for enhancing behaviour detection. To this end, we have adopted a physics-informed inference model. Our first group of techniques enables robust statistical analysis within a learned continuous behaviour latent space, facilitating the detection of subtle behavioural shifts relative to reference genetic lines. A second array of methods probes for subtle variations in action sequences by comparing them to a bespoke generative model. Together, these strategies have enabled us to construct representations of behavioural patterns specific to a lineage and identify a roster of "hit" neurons with the potential to influence behaviour subtly.

neuroscience↗

bioRxiv2024

LarvaTagger: Manual and automatic tagging of Drosophila larval behaviour

MotivationAs more behavioural assays are carried out in large-scale experiments on Drosophila larvae, the definitions of the archetypal actions of a larva are regularly refined. In addition, video recording and tracking technologies constantly evolve. Consequently, automatic tagging tools for Drosophila larval behaviour must be retrained to learn new representations from new data. However, existing tools cannot transfer knowledge from large amounts of previously accumulated data. We introduce LarvaTagger, a piece of software that combines a pre-trained deep neural network, providing a continuous latent representation of larva actions for stereotypical behaviour identification, with a graphical user interface to manually tag the behaviour and train new automatic taggers with the updated ground truth. ResultsWe reproduced results from an automatic tagger with high accuracy, and we demonstrated that pre-training on large databases accelerates the training of a new tagger, achieving similar prediction accuracy using less data. AvailabilityAll the code is free and open source. Docker images are also available. See git-lab.pasteur.fr/nyx/LarvaTagger.jl.

bioinformatics↗

bioRxiv2024

Asymmetric representation of symmetric semantic information in the human brain

Specific pairs of semantic entities have symmetric relationships, such as word pairs with opposite meanings (e.g., "intelligent" and "stupid"; "human" and "mechanical"). Such semantic symmetry is a key feature of semantic information. However, the representation of symmetric semantic information in the brain is not yet understood. Additionally, it is unclear whether symmetric pairs of semantic information do have symmetric representations in the brain? We addressed this question in a data-driven manner by using the voxelwise modeling of movie-evoked cortical response measured by functional magnetic resonance imaging. In this modeling, response in each voxel was predicted from semantic labels designated for each movie scene. The semantic labels consisted of 30 different items, including 15 pairs of semantically symmetric items. Each item was manually evaluated using a 5-point scale. By localizing the semantic representation associated with each item based on the voxelwise accuracy of brain-response predictions, we found that semantic representations of symmetric item pairs are broadly distributed but with little overlap in the cortex. Additionally, the weight of voxelwise models revealed highly complex, various patterns of cortical representations for each item pair. These results suggest that symmetric semantic information is rather asymmetric and heterogeneous representations in the human brain. Significance statementThis study aimed to investigate if symmetric pairs of semantic information have symmetric representations in the human brain using statistical modeling of functional magnetic resonance imaging signals evoked by naturalistic movies. We built a model based on movie labeling of symmetric semantic items to quantify the cortical representations of symmetric semantic entities. These findings showed that symmetric pairs of semantic entities are represented in widespread cortical regions; however, still exhibit little overlap in localization and heterogeneous representations. These results offer significant insights into the cortical representations of semantic symmetry and advance the understanding of representational structure of semantic information in the human brain.

neuroscience↗

bioRxiv2023

Feeding-state dependent modulation of reciprocally interconnected inhibitory neurons biases sensorimotor decisions in Drosophila

Animals feeding state changes behavioral priorities and thus influences even non-feeding related decisions. How is the feeding state information transmitted to non-feeding related circuits and what are the circuit mechanisms involved in biasing non-feeding related decisions remains an open question. By combining calcium imaging, neuronal manipulations, behavioral analysis and computational modeling, we determined that the competition between different aversive responses to mechanical cues is biased by feeding state changes. We found that this is achieved by differential modulation of two different types of reciprocally connected inhibitory neurons promoting opposing actions. This modulation results in a more frequent active type of response and less frequently a protective type of response if larvae are fed sugar compared to when they are fed a balanced diet. The information about the internal state is conveyed to the inhibitory neurons through homologues of the vertebrate neuropeptide Y known to be involved in regulating feeding behavior.

neuroscience↗

bioRxiv2023

A simple solid media assay for detection of synergy between bacteriophages and antibiotics

The emergence of antibiotic resistant bacteria (ARB) has necessitated the development of alternative therapies to deal with this global threat. Bacteriophages (viruses that target bacteria) that kill ARB are one such alternative. While phages have been used clinically for decades with inconsistent results, a number of recent advances in phage selection, propagation and purification have enabled a reevaluation of their utility in contemporary clinical medicine. In most phage therapy cases, phages are administered in combination with antibiotics to ensure that patients receive the standard-of-care treatment. Some phages may work cooperatively with antibiotics to eradicate ARB, as often determined using non-standardized broth assays. We sought to develop a solid media-based assay to assess cooperativity between antibiotics and phages to offer a standardized platform for such testing. We modeled the interactions that occur between antibiotics and phages on solid medium to measure additive, antagonistic, and synergistic interactions. We then tested the method using different bacterial isolates, and identified a number of isolates where synergistic interactions were identified. These interactions were not dependent on the specific organism, phage family, or antibiotic used. A priori susceptibility to the antibiotic or the specific phage were not requirements to observe synergistic interactions. Our data also confirm the potential for the restoration of vancomycin to treat Vancomycin Resistant Enterococcus (VRE) when used in combination with phages. Solid media assays for the detection of cooperative interactions between antibiotics and phages can be an accessible technique adopted by clinical laboratories to evaluate antibiotic and phage choices in phage therapy.

microbiology↗

bioRxiv2022

Decoding Individual Differences in Mental Information from Human Brain Response Predicted by Convolutional Neural Networks

Recent advantages in brain decoding with functional magnetic resonance imaging (fMRI) have enabled us to estimate individual differences in mental information from brain responses to natural stimuli. However, the physical constraints and costs of fMRI measurements prevent brain decoding from achieving real-world applications. We address this issue by building a novel framework to decode individual differences in mental information under natural situations from brain responses predicted using convolutional neural networks (CNNs). Once the CNN-based prediction model is trained using measured response, mental information can be decoded from the predicted responses of individual brains with no need for additional fMRI measurements. The model was found to capture individual-difference patterns consistent with conventional decoding using measured responses in 81/87 items to be decoded. Our framework has great potential to decode personal mental information with low dependence on fMRI measurements, which could help substantially expand the applicability of brain decoding in daily life.

neuroscience↗

bioRxiv2020

Design, expression, purification and characterization of a YFP-tagged 2019-nCoV spike receptor-binding domain construct

2019-nCoV is the causative agent of the serious, still ongoing, worldwide COVID-19 pandemic. High quality recombinant virus proteins are required for research related to the development of vaccines and improved assays, and to the general understanding of virus action. The receptor-binding domain (RBD) of the 2019-nCoV spike (S) protein contains disulfide bonds and N-linked glycosylations, therefore, it is typically produced by secretion. Here, we describe a construct and protocol for the expression and purification of yellow fluorescent protein (YFP) labeled 2019-nCoV spike RBD. The fusion protein, in the vector pcDNA 4/TO, comprises an N-terminal interferon alpha 2 (IFN2) signal peptide, an eYFP, a FLAG-tag, a human rhinovirus 3C protease cleavage site, the RBD of the 2019-nCoV spike protein and a C-terminal 8x His-tag. We stably transfected HEK 293 cells. Following expansion of the cells, the fusion protein was secreted from adherent cells into serum-free medium. Ni-NTA IMAC purification resulted in very high protein purity, based on analysis by SDS-PAGE. The fusion protein was soluble and monodisperse, as confirmed by size-exclusion chromatography (SEC) and negative staining electron microscopy. Deglycosylation experiments confirmed the presence of N-linked glycosylations in the secreted protein. Complex formation with the peptidase domain of human angiotensin-converting enzyme 2 (ACE2), the receptor for the 2019-nCoV spike RBD, was confirmed by SEC, both for the YFP-fused spike RBD and for spike RBD alone, after removal of YFP by proteolytic cleavage. Possible applications for the fusion protein include binding studies on cells or in vitro, fluorescent labeling of potential virus-binding sites on cells, the use as an antigen for immunization studies or as a tool for the development of novel virus- or antibody-detection assays.

biochemistry↗

bioRxiv2020

Behavioral correlates of cortical semantic representations modeled by word vectors

The quantitative modeling of semantic representations in the brain plays a key role in understanding the neural basis of semantic processing. Previous studies have demonstrated that word vectors, which were originally developed for use in the field of natural language processing, provide a powerful tool for such quantitative modeling. However, whether semantic representations revealed by the word vector-based models actually capture our perception of semantic information remains unclear, as there has been no study explicitly examining the behavioral correlates of the modeled semantic representations. To address this issue, we compared the semantic structure of nouns and adjectives estimated from word vector-based brain models with that evaluated from human behavior. The brain models were constructed using voxelwise modeling to predict the functional magnetic resonance imaging (fMRI) response to natural movies from semantic contents in each movie scene through a word vector space. The semantic dissimilarity of word representations was then evaluated using the brain models. Meanwhile, data on human behavior reflecting the perception of semantic dissimilarity between words were collected in psychological experiments. We found a significant correlation between brain model- and behavior-derived semantic dissimilarities of words. This finding suggests that semantic representations in the brain modeled via word vectors appropriately capture our perception of word meanings. Author summeryWord vectors, which have been originally developed in the field of engineering (natural language processing), have been extensively leveraged in neuroscience studies to model semantic representations in the human brain. These studies have attempted to model brain semantic representations by associating them with the meanings of thousands of words via a word vector space. However, there has been no study explicitly examining whether the modeled semantic representations actually capture our perception of semantic information. To address this issue, we compared the semantic representational structure of words estimated from word vector-based brain models with that evaluated from behavioral data in psychological experiments. The results revealed a significant correlation between these model- and behavior-derived semantic representational structures of words. This indicates that the brain semantic representations modeled using word vectors actually reflect the human perception of word meanings. Our findings contribute to the establishment of word vector-based brain modeling as a useful tool in studying human semantic processing.

neuroscience↗