Biocide and other semi-volatile organic compound concentrations in settled indoor dust of CRESPI daycare centers and implication for public health.

This study investigates the presence of biocides and other semi-volatile organic compounds (SVOCs) in cleaning products used in daycare centers and health impact through ingestion of settled dust by young children. In Paris metropolitan area, 106 daycares area were investigated between 2019-2022. Fifteen substances were analyzed in settled indoor dust by gas chromatography-tandem mass spectrometry. Detection rates and concentrations ranged from 5 to 100%, and

What is a Rhythm for the Brain? The Impact of Contextual Temporal Variability on Auditory Perception.

Temporal predictions can be formed and impact perception when sensory timing is fully predictable: for instance, the discrimination of a target sound is enhanced if it is presented on the beat of an isochronous rhythm. However, natural sensory stimuli, like speech or music, are not entirely predictable, but still possess statistical temporal regularities. We investigated whether temporal expectations can be formed in non-fully predictable contexts, and how the temporal variability of sensory contexts affects auditory perception. Specifically, we asked how "rhythmic" an auditory stimulation needs to be in order to observe temporal predictions effects on auditory discrimination performances. In this behavioral auditory oddball experiment, participants listened to auditory sound sequences where the temporal interval between each sound was drawn from gaussian distributions with distinct standard deviations. Participants were asked to discriminate sounds with a deviant pitch in the sequences. Auditory discrimination performances, as measured with deviant sound discrimination accuracy and response times, progressively declined as the temporal variability of the sound sequence increased. Moreover, both global and local temporal statistics impacted auditory perception, suggesting that temporal statistics are promptly integrated to optimize perception. Altogether, these results suggests that temporal predictions can be set up quickly based on the temporal statistics of past sensory events and are robust to a certain amount of temporal variability. Therefore, temporal predictions can be built on sensory stimulations that are not purely periodic nor temporally deterministic.

Changes in Gene Expression After Exposing Arabidopsis thaliana Plants to Nanosecond High Amplitude Electromagnetic Field Pulses.

The biological effects of exposure to electromagnetic fields due to wireless technologies and connected devices are a subject of particular research interest. Ultrashort high-amplitude electromagnetic field pulses delivered to biological samples using immersed electrodes in a dedicated cuvette have widely demonstrated their effectiveness in triggering several cell responses including increased cytosolic calcium concentration and reactive oxygen species (ROS) production. In contrast, the effects of these pulses are poorly documented when electromagnetic pulses are delivered through an antenna. Here we exposed Arabidopsis thaliana plants to 30,000 pulses (237 kV m-1 , 280 ps rise-time, duration of 500 ps) emitted through a Koshelev antenna and monitored the consequences of electromagnetic fields exposure on the expression levels of several key genes involved in calcium metabolism, signal transduction, ROS, and energy status. We found that this treatment was mostly unable to trigger significant changes in the messenger RNA accumulation of calmodulin, Zinc-Finger protein ZAT12, NADPH oxidase/respiratory burst oxidase homolog (RBOH) isoforms D and F, Catalase (CAT2), glutamate-cystein ligase (GSH1), glutathione synthetase (GSH2), Sucrose non-fermenting-related Kinase 1 (SnRK1) and Target of rapamycin (TOR). In contrast, Ascorbate peroxidases APX-1 and APX-6 were significantly induced 3 h after the exposure. These results suggest that this treatment, although quite strong in amplitude, is mostly ineffective in inducing biological effects at the transcriptional level when delivered by an antenna. © 2023 The Authors. Bioelectromagnetics published by Wiley Periodicals LLC on behalf of Bioelectromagnetics Society.

Envisaging a global infrastructure to exploit the potential of digitised collections.

Tens of millions of images from biological collections have become available online over the last two decades. In parallel, there has been a dramatic increase in the capabilities of image analysis technologies, especially those involving machine learning and computer vision. While image analysis has become mainstream in consumer applications, it is still used only on an artisanal basis in the biological collections community, largely because the image corpora are dispersed. Yet, there is massive untapped potential for novel applications and research if images of collection objects could be made accessible in a single corpus. In this paper, we make the case for infrastructure that could support image analysis of collection objects. We show that such infrastructure is entirely feasible and well worth investing in.

Estimating Compositions and Nutritional Values of Seed Mixes Based on Vision Transformers.

The cultivation of seed mixtures for local pastures is a traditional mixed cropping technique of cereals and legumes for producing, at a low production cost, a balanced animal feed in energy and protein in livestock systems. By considerably improving the autonomy and safety of agricultural systems, as well as reducing their impact on the environment, it is a type of crop that responds favorably to both the evolution of the European regulations on the use of phytosanitary products and the expectations of consumers who wish to increase their consumption of organic products. However, farmers find it difficult to adopt it because cereals and legumes do not ripen synchronously and the harvested seeds are heterogeneous, making it more difficult to assess their nutritional value. Many efforts therefore remain to be made to acquire and aggregate technical and economical references to evaluate to what extent the cultivation of seed mixtures could positively contribute to securing and reducing the costs of herd feeding. The work presented in this paper proposes new Artificial Intelligence techniques that could be transferred to an online or smartphone application to automatically estimate the nutritional value of harvested seed mixes to help farmers better manage the yield and thus engage them to promote and contribute to a better knowledge of this type of cultivation. For this purpose, an original open image dataset has been built containing 4,749 images of seed mixes, covering 11 seed varieties, with which 2 types of recent deep learning models have been trained. The results highlight the potential of this method and show that the best-performing model is a recent state-of-the-art vision transformer pre-trained with self-supervision (Bidirectional Encoder representation from Image Transformer). It allows an estimation of the nutritional value of seed mixtures with a coefficient of determination R2 score of 0.91, which demonstrates the interest of this type of approach, for its possible use on a large scale.

Validating instructional design and predicting student performance in histology education: Using machine learning via virtual microscopy.

As a part of modern technological environments, virtual microscopy enriches histological learning, with support from large institutional investments. However, existing literature does not supply empirical evidence of its role in improving pedagogy. Virtual microscopy provides fresh opportunities for investigating user behavior during the histology learning process, through digitized histological slides. This study establishes how students' perceptions and user behavior data can be processed and analyzed using machine learning algorithms. These also provide predictive data called learning analytics that enable predicting students' performance and behavior favorable for academic success. This information can be interpreted and used for validating instructional designs. Data on the perceptions, performances, and user behavior of 552 students enrolled in a histology course were collected from the virtual microscope, Cytomine®. These data were analyzed using an ensemble of machine learning algorithms, the extra-tree regression method, and predictive statistics. The predictive algorithms identified the most pertinent histological slides and descriptive tags, alongside 10 types of student behavior conducive to academic success. We used these data to validate our instructional design, and align the educational purpose, learning outcomes, and evaluation methods of digitized histological slides on Cytomine®. This model also predicts students' examination scores, with an error margin of <0.5 out of 20 points. The results empirically demonstrate the value of a digital learning environment for both students and teachers of histology.

Atovaquone exposure and Pneumocystis jirovecii cytochrome b mutations: French data and review of the literature.

Pneumocystis jirovecii is a transmissible fungus responsible for severe pneumonia (Pneumocystis pneumonia [PCP]) in immunocompromised patients. Missense mutations due to atovaquone selective pressure have been identified on cytochrome b (CYB) gene of P. jirovecii. It was recently shown that atovaquone prophylaxis can lead to the selection of specific P. jirovecii CYB mutants potentially resistant to atovaquone among organ transplant recipients. In this context, our objectives were to provide data on P. jirovecii CYB mutants and the putative selective pressure exerted by atovaquone on P. jirovecii organisms in France. A total of 123 patients (124 P. jirovecii specimens) from four metropolitan hospitals and two overseas hospitals were retrospectively enrolled. Fourteen patients had prior exposure to atovaquone, whereas 109 patients did not at the time of P. jirovecii detection. A 638 base-pair fragment of the CYB gene of P. jirovecii was amplified and sequenced. A total of 10 single nucleotide polymorphisms (SNPs) were identified. Both missense mutations C431T (Ala144Val) and C823T (Leu275Phe), located at the Qo active site of the enzyme, were significantly associated with prior atovaquone exposure, these mutations being conversely incidental in the absence of prior atovaquone exposure (P < 0.001). Considering that the aforementioned hospitals may be representative of the national territory, these findings suggest that the overall presence of P. jirovecii CYB mutants remains low in France.

The mutations C431T (Ala144Val) and C823T (Leu275Phe) at the cytochrome b (CYB) active site of Pneumocystis jirovecii are associated with patient prior exposure to atovaquone. Conversely, these mutations are incidental in the absence of exposure. Overall, the presence of P. jirovecii CYB mutants remains low in France.

[1,2,4]triazolo[4,3-a]quinoxaline as Novel Scaffold in the Imiqualines Family: Candidates with Cytotoxic Activities on Melanoma Cell Lines.

Cutaneous melanoma is one of the most aggressive human cancers and is the deadliest form of skin cancer, essentially due to metastases. Novel therapies are always required, since cutaneous melanoma develop resistance to oncogenic pathway inhibition treatment. The Imiqualine family is composed of heterocycles diversely substituted around imidazo[1,2-a]quinoxaline, imidazo[1,2-a]pyrazine, imidazo[1,5-a]quinoxaline, and pyrazolo[1,5-a]quinoxaline scaffolds, which display interesting activities on a panel of cancer cell lines, especially melanoma cell lines. We have designed and prepared novel compounds based on the [1,2,4]triazolo[4,3-a]quinoxaline scaffold through a common synthetic route, using 1-chloro-2-hydrazinoquinoxaline and an appropriate aldehyde. Cyclization is ensured by an oxidation-reduction mechanism using chloranil. The substituents on positions 1 and 8 were chosen based on previous structure-activity relationship (SAR) studies conducted within our heterocyclic Imiqualine family. Physicochemical parameters of all compounds have also been predicted. A375 melanoma cell line viability has been evaluated for 16 compounds. Among them, three novel [1,2,4]triazolo[4,3-a]quinoxalines display cytotoxic activities. Compounds 16a and 16b demonstrate relative activities in the micromolar range (respectively, 3158 nM and 3527 nM). Compound 17a shows the best EC50 of the novel series (365 nM), even if EAPB02303 remains the lead of the entire Imiqualine family (3 nM).

Exposure to Disinfectants and Cleaning Products and Respiratory Health of Workers and Children in Daycares: The CRESPI Cohort Protocol.

Although cleaning tasks are frequently performed in daycare, no study has focused on exposures in daycares in relation to respiratory health. The CRESPI cohort is an epidemiological study among workers (n~320) and children (n~540) attending daycares. The purpose is to examine the impact of daycare exposures to disinfectants and cleaning products (DCP) on the respiratory health of workers and children. A sample of 108 randomly selected daycares in the region of Paris has been visited to collect settled dust to analyze semi-volatile organic compounds and microbiota, as well as sample indoor air to analyze aldehydes and volatile organic compounds. Innovative tools (smartphone applications) are used to scan DCP barcodes in daycare and inform their use; a database then matches the barcodes with the products' compositions. At baseline, workers/parents completed a standardized questionnaire, collecting information on DCP used at home, respiratory health, and potential confounders. Follow-up regarding children's respiratory health (monthly report through a smartphone application and biannual questionnaires) is ongoing until the end of 2023. Associations between DCP exposures and the respiratory health of workers/children will be evaluated. By identifying specific environments or DCP substances associated with the adverse respiratory health of workers and children, this longitudinal study will contribute to the improvement of preventive measures.

Non thermal 2.45 GHz electromagnetic exposure causes rapid changes in Arabidopsis thaliana metabolism.

Numerous studies report different types of responses following exposure of plants to high frequency electromagnetic fields (HF-EMF). While this phenomenon is related to tissue heating in animals, the situation is much less straightforward in plants where metabolic changes seem to occur without tissue temperature increase. We have set up an exposure system allowing reliable measurements of tissue heating (using a reflectometric probe and thermal imaging) after a long exposure (30 min) to an electromagnetic field of 2.45 GHz transmitted through a horn antenna (about 100 V m-1 at the plant level). We did not observe any heating of the tissues, but we detected rapid increases (60 min) in the accumulation of transcripts of stress-related genes (TCH1 and ZAT12 transcription factor) or involved in ROS metabolism (RBOHF and APX1). At the same time, the amounts of hydrogen peroxide and dehydroascorbic acid increased while glutathione (reduced and oxidized forms), ascorbic acid, and lipid peroxidation remained stable. Therefore, our results unambiguously show that molecular and biochemical responses occur rapidly (within 60min) in plants after exposure to an electromagnetic field, in absence of tissue heating.

Reactive sputtering onto an ionic liquid, a new synthesis route for bismuth-based nanoparticles.

Metallic bismuth and Bi-oxyfluoride nanoparticles (NPs) are successfully synthesized by non-reactive and reactive sputtering of a Bi target onto 1-butyl-3-methylimidazolium bis(trifluoro-methylsulfonyl)imide ([BMIM][TFSI]) ionic liquid (IL). Non-reactive sputtering is realized in pure Ar plasma, where isotropic, well crystallized and dispersed Bi NPs of 3-7 nm are obtained. The diameter and the size distribution of these NPs do not significantly vary with the power, gas pressure, and sputtering time; but these sputtering parameters seem to influence the NP concentration. Then, the introduction of O2 and CF4 gases in addition to Ar enables the reaction of radicals from plasma with Bi clusters at the liquid's top surface to form Bi-oxyfluoride NPs of 3-12 nm in diameter with photocatalytic activity. Hence, the reactive sputtering onto an IL is an efficient, original and promising method for synthesizing Bi-based compound NPs. Finally, we propose a mechanism based on reactions of species from plasma at the IL surface to explain the formation of Bi-compounds by reactive sputtering.

Recent Advances on Small-Molecule Antagonists Targeting TLR7.

Toll-like receptor 7 (TLR7) is a class of pattern recognition receptors (PRRs) recognizing the pathogen-associated elements and damage and as such is a major player in the innate immune system. TLR7 triggers the release of pro-inflammatory cytokines or type-I interferons (IFN), which is essential for immunoregulation. Increasing reports also highlight that the abnormal activation of endosomal TLR7 is implicated in various immune-related diseases, carcinogenesis as well as the proliferation of human immunodeficiency virus (HIV). Hence, the design and development of potent and selective TLR7 antagonists based on small molecules or oligonucleotides may offer new tools for the prevention and management of such diseases. In this review, we offer an updated overview of the main structural features and therapeutic potential of small-molecule antagonists of TLR7. Various heterocyclic scaffolds targeting TLR7 binding sites are presented: pyrazoloquinoxaline, quinazoline, purine, imidazopyridine, pyridone, benzanilide, pyrazolopyrimidine/pyridine, benzoxazole, indazole, indole, and quinoline. Additionally, their structure-activity relationships (SAR) studies associated with biological activities and protein binding modes are introduced.

Apparent Absence of Selective Pressure on Pneumocystis jirovecii Organisms in Patients with Prior Methotrexate Exposure.

Pneumocystis jirovecii infections occur in patients treated with methotrexate (MTX) because of immunosuppressive effects of this highly potent dihydrofolate reductase (DHFR) inhibitor. Conversely, MTX may act as an anti-P. jirovecii drug and consequently may exert a selective pressure on this fungus. In this context, we compared the sequences of the dhfr gene of P. jirovecii isolates obtained from two groups of patients with P. jirovecii infections. The first group, with systemic diseases or malignancies, had prior exposure to MTX (21 patients), whereas the second group (22 patients), the control group, did not. Three single nucleotide polymorphisms (SNPs) were observed at positions 278, 312, and 381. The first one was located in the intronic region and the two others were synonymous. Based on these SNPs, three P. jirovecii dhfr alleles, named A, B, and C, were specified. Allele A was the most frequent, as it was observed in 18 patients (85.7%) and in 16 patients (72.7%) of the first and second groups, respectively. No significant difference in P. jirovecii dhfr gene diversity in the two patient groups was observed. In conclusion, these original results suggest that MTX does not exert an overt selective pressure on P. jirovecii organisms.

Domain Decomposition Spectral Method Applied to Modal Method: Direct and Inverse Spectral Transforms.

We introduce a Domain Decomposition Spectral Method (DDSM) as a solution for Maxwell's equations in the frequency domain. It will be illustrated in the framework of the Aperiodic Fourier Modal Method (AFMM). This method may be applied to compute the electromagnetic field diffracted by a large-scale surface under any kind of incident excitation. In the proposed approach, a large-size surface is decomposed into square sub-cells, and a projector, linking the set of eigenvectors of the large-scale problem to those of the small-size sub-cells, is defined. This projector allows one to associate univocally the spectrum of any electromagnetic field of a problem stated on the large-size domain with its footprint on the small-scale problem eigenfunctions. This approach is suitable for parallel computing, since the spectrum of the electromagnetic field is computed on each sub-cell independently from the others. In order to demonstrate the method's ability, to simulate both near and far fields of a full three-dimensional (3D) structure, we apply it to design large area diffractive metalenses with a conventional personal computer.

Diagnosis and treatment of pudendal and inferior cluneal nerve entrapment syndrome: a narrative review.

AIM: Pudendal and inferior cluneal nerve entrapment can cause a neuropathic pain syndrome in the sensitive areas innervated by these nerves. Diagnosis is challenging and patients often suffer several years before diagnosis is made. The purpose of the review was to inform healthcare workers about this disease and to provide a basis of anatomy and physiopathology, to inform about diagnostic tools and invasive or non-invasive treatment modalities and outcome. METHODS: A description of pudendal and inferior cluneal nerve anatomy is given. Physiopathology for entrapment is explained. Diagnostic criteria are described, and all non-invasive and invasive treatment options are discussed. RESULTS: The Nantes criteria offer a solid basis for diagnosing this rare condition. Treatment should be offered in a pluri-disciplinary setting and consists of avoidance of painful stimuli, physiotherapy, psychotherapy, pharmacological treatment led by tricyclic antidepressants and anticonvulsants. Nerve blocks are efficient at short term and serve mainly as a diagnostic tool. Pulsed radiofrequency (PRF) is described as a successful treatment option for pudendal neuralgia in patients non-responding to non-invasive treatment. If all other treatments fail, surgery can be offered. Different surgical procedures exist but only the open transgluteal approach has proven its efficacy compared to medical treatment. The minimal-invasive ENTRAMI technique offers the possibility to combine nerve release with pudendal neuromodulation. CONCLUSIONS: Pudendal and inferior cluneal nerve entrapment syndrome are a challenge not only for diagnosis but also for treatment. Different non-invasive and invasive treatment options exist and should be offered in a pluri-disciplinary setting.

Dynamic Species Distribution Modeling Reveals the Pivotal Role of Human-Mediated Long-Distance Dispersal in Plant Invasion.

Plant invasions generate massive ecological and economic costs worldwide. Predicting their spatial dynamics is crucial to the design of effective management strategies and the prevention of invasions. Earlier studies highlighted the crucial role of long-distance dispersal in explaining the speed of many invasions. In addition, invasion speed depends highly on the duration of its lag phase, which may depend on the scaling of fecundity with age, especially for woody plants, even though empirical proof is still rare. Bayesian dynamic species distribution models enable the fitting of process-based models to partial and heterogeneous observations using a state-space modeling approach, thus offering a tool to test such hypotheses on past invasions over large spatial scales. We use such a model to explore the roles of long-distance dispersal and age-structured fecundity in the transient invasion dynamics of Plectranthus barbatus, a woody plant invader in South Africa. Our lattice-based model accounts for both short and human-mediated long-distance dispersal, as well as age-structured fecundity. We fitted our model on opportunistic occurrences, accounting for the spatio-temporal variations of the sampling effort and the variable detection rates across datasets. The Bayesian framework enables us to integrate a priori knowledge on demographic parameters and control identifiability issues. The model revealed a massive wave of spatial spread driven by human-mediated long-distance dispersal during the first decade and a subsequent drastic population growth, leading to a global equilibrium in the mid-1990s. Without long-distance dispersal, the maximum population would have been equivalent to 30% of the current equilibrium population. We further identified the reproductive maturity at three years old, which contributed to the lag phase before the final wave of population growth. Our results highlighted the importance of the early eradication of weedy horticultural alien plants around urban areas to hamper and delay the invasive spread.

The dural vascular plexus in subdural hematoma: Illustration through a case of dural arteriovenous fistula.

Background: The initiation of chronic subdural hematoma (cSDH) is traditionally explained by rupture of bridging veins. Recent descriptions of the embryology and anatomy of the meninges and their vascularization, however, point to the dural vascular plexus (DVP) as a plausible origin of cSDH. This dural plexus is supplied by meningeal arteries. Their endovascular occlusion is efficient in cSDH treatment. Dural arteriovenous fistulae (dAVF) may also present with subdural hematoma. Case Description: A 65-year-old female patient presented with parietal parasagittal dAVF and bilateral cSDH requiring surgical disconnection followed by complete clinical and imaging resolution of dAVF and cSDH. Conclusion: In common cSDH, pressure in the DVP may be normal and subdural bleeding may occur due to mechanical traction on the DVP. In the setting of dAVF, it may be the increase in pressure due to the fistula, within the DVP, that causes subdural hematoma. The DVP, supplied by meningeal arteries, thus not only allows for convergent pathophysiological explanation of subdural bleeding in both cSDH and dAVF but may also be the actual target of the emergent endovascular treatment of cSDH trough meningeal artery embolization.

Deep Species Distribution Modeling From Sentinel-2 Image Time-Series: A Global Scale Analysis on the Orchid Family.

Species distribution models (SDMs) are widely used numerical tools that rely on correlations between geolocated presences (and possibly absences) and environmental predictors to model the ecological preferences of species. Recently, SDMs exploiting deep learning and remote sensing images have emerged and have demonstrated high predictive performance. In particular, it has been shown that one of the key advantages of these models (called deep-SDMs) is their ability to capture the spatial structure of the landscape, unlike prior models. In this paper, we examine whether the temporal dimension of remote sensing images can also be exploited by deep-SDMs. Indeed, satellites such as Sentinel-2 are now providing data with a high temporal revisit, and it is likely that the resulting time-series of images contain relevant information about the seasonal variations of the environment and vegetation. To confirm this hypothesis, we built a substantial and original dataset (called DeepOrchidSeries) aimed at modeling the distribution of orchids on a global scale based on Sentinel-2 image time series. It includes around 1 million occurrences of orchids worldwide, each being paired with a 12-month-long time series of high-resolution images (640 x 640 m RGB+IR patches centered on the geolocated observations). This ambitious dataset enabled us to train several deep-SDMs based on convolutional neural networks (CNNs) whose input was extended to include the temporal dimension. To quantify the contribution of the temporal dimension, we designed a novel interpretability methodology based on temporal permutation tests, temporal sampling, and temporal averaging. We show that the predictive performance of the model is greatly increased by the seasonality information contained in the temporal series. In particular, occurrence-poor species and diversity-rich regions are the ones that benefit the most from this improvement, revealing the importance of habitat's temporal dynamics to characterize species distribution.

The impact of material design on the photocatalytic removal efficiency and toxicity of two textile dyes.

This study deals with the toxicity of the treated solutions of two types of dyes, namely, the anthraquinonic Reactive Bleu 19 dye (RB19) and the bi-azoic Direct Red 227 dye (DR227), which are treated in single and binary mixture systems. The target molecules were removed by the photocatalysis process using ZnO as a catalyst, which was calcined at two temperatures 250 and 420 °C (ZnO250 and ZnO420) prepared in the lab by the one-step calcination method. XRD, TEM, EDX, XPS, FT-IR, BET, RAMAN, and EPR analyses were carried out to characterize the catalyst material. While the phytotoxicity was being conducted using watercress seeds, the cytotoxicity took place using a cell line (raw) and an intestinal cell (caco-2). The XRD analysis showed the partial calcination of ZnO250 and the presence of anhydrous zinc acetate along with the ZnO nanoparticles (NPs). This result was not observed for ZnO420. Despite the complete discoloration (100%) of all the final solutions, ZnO250 exhibited a high cytotoxicity and phytotoxicity against the RB19 dye after the photocatalytic treatment; however, it was not the case of ZnO420 which was selected as an eco-friendly photocatalyst for the degradation of organic dyes based on the results of removal efficiency, cytotoxicity, and phytotoxicity.

Very High Resolution Species Distribution Modeling Based on Remote Sensing Imagery: How to Capture Fine-Grained and Large-Scale Vegetation Ecology With Convolutional Neural Networks?

Species Distribution Models (SDMs) are fundamental tools in ecology for predicting the geographic distribution of species based on environmental data. They are also very useful from an application point of view, whether for the implementation of conservation plans for threatened species or for monitoring invasive species. The generalizability and spatial accuracy of an SDM depend very strongly on the type of model used and the environmental data used as explanatory variables. In this article, we study a country-wide species distribution model based on very high resolution (VHR) (1 m) remote sensing images processed by a convolutional neural network. We demonstrate that this model can capture landscape and habitat information at very fine spatial scales while providing overall better predictive performance than conventional models. Moreover, to demonstrate the ecological significance of the model, we propose an original analysis based on the t-distributed Stochastic Neighbor Embedding (t-SNE) dimension reduction technique. It allows visualizing the relation between input data and species traits or environment learned by the model as well as conducting some statistical tests verifying them. We also analyze the spatial mapping of the t-SNE dimensions at both national and local levels, showing the model benefit of automatically learning environmental variation at multiple scales.