Biology subjects

Genc, E.

Publications and source records attributed to Genc, E..

At least 19 recordsLinked to original sources

bioRxiv2026

Distinct and overlapping correlates of fear acquisition and extinction across different neuroimaging modalities

Interindividual differences in fear acquisition and extinction have been related to variation in specific brain correlates. However, variability in experimental setups complicates the integration of findings. Here, we present a combined fear acquisition (n = 101) and extinction (n = 88) study from which we obtained distinct microstructural, macrostructural, and connectivity brain properties with magnetic resonance imaging in healthy, young participants. The properties included regional brain volume, cortical surface area and thickness, neurite density and orientation dispersion, and nodal efficiency of structural and functional connectivity. Fear responses were quantified as changes in skin conductance. Data from 360 cortical and 16 subcortical brain regions as well as the efficiency of network connections between them were used as independent variables in bootstrapped and cross-validated regularized regression models. Results show that numerous brain regions, spanning the so-called 'fear and extinction network' and extending beyond it, contribute to fear acquisition, to extinction, or dynamically shift between both phases. For several brain regions, data from multiple imaging modalities showed a high degree of concordance for several brain regions. These findings call for further research to examine the potential interplay between brain correlates shaping fear acquisition and extinction, as opposed to studying the imaging modalities in isolation.

neuroscience↗

bioRxiv2026

Testing the reliability of novel Voxel Placement approaches for Magnetic Resonance Spectroscopy

BackgroundSingle-voxel magnetic resonance spectroscopy (MRS) is a non-invasive method for measuring clinically and cognitively relevant metabolites. Reliable measurements require precise voxel placement across sessions and participants. We developed a scanner-console-based approach to improve voxel placement precision. MethodsIn a crossover design (n=7; six sessions each), we compared test-retest reliability of three voxel placement methods in a reference benchmark (left parietal cortex) and a technically challenging region (left ventromedial prefrontal cortex). Methods included (1) conventional anatomy-based placement, (2) mask-guided real-time positioning (MGRP), and (3) semiautomated session-locked voxel repositioning (SSVR). Resting-state MRS data were acquired using PRESS and MEGA-PRESS. Within-subject reliability of voxel placement and metabolite concentrations, namely, total N-acetylaspartate (tNAA), total Creatine (tCr), GABA (gamma-aminobutyric acid), and Glx (glutamate + glutamine) are reported using the coefficient of variation (CV), the intraclass correlation coefficient (ICC), minimal detectable change (MDC), and the spatial overlap. ResultsSSVR markedly improved voxel placement reliability, increasing spatial overlap (up to 88%) and achieving near-perfect geometric reproducibility (ICC = 0.99) compared to conventional anatomy-based placement and MGRP. SSVR improved tissue composition consistency and reduced metabolite variability in the technically challenging region (variability reduction of [~]70% tCr, [~]59% tNAA, and [~]51% Glx) while further refining already stable measurements in the benchmark region (tNAA from [~]15% to [~]10%). ConclusionBoth MGRP and SSVR improved voxel placement and metabolite measurement reproducibility compared with conventional anatomy-based placement. SSVR further enhanced within-subject reproducibility across repeated sessions, particularly in the technically challenging region, providing a robust approach for longitudinal single-voxel MRS studies.

neuroscience↗

bioRxiv2026

Determining Reliable Neural Targets for enhancing Motor Sequence Learning: A preparatory study for focal tDCS

BackgroundMotor sequence learning (MSL) underlies the acquisition of everyday skilled movements, and transcranial direct current stimulation (tDCS) has shown promise for enhancing it, although reported effects vary considerably. Before such interventions can be applied and interpreted reliably, the behavioural and neural signatures of the underlying task should be shown to be robust. This study therefore characterised an fMRI-based Serial Reaction Time Task (SRTT) protocol and evaluated the consistency of its behavioural and fMRI-based neural signatures in repeated measurements to identify potential network targets for tDCS. MethodsTwenty healthy young- to middle-aged adults (10 females; mean age 24, SD 4) performed the SRTT during 3T fMRI in two sessions about one week apart, each using one of two matched, non-overlapping task versions in counterbalanced order. We examined block-wise behaviour, task-related and sequence-specific brain activation (general task activation relative to baseline and the SEQ > RND contrast), and functional connectivity using two complementary approaches: the task and temporal dynamics of seed-based connectivity (SBC) within an a priori-defined motor network (M1, SMA, PMC, putamen, and cerebellum), and whole-brain generalised psychophysiological interaction (gPPI) for the SEQ > RND contrast. ResultsReaction time decreased steadily across sequence blocks and returned toward baseline during random blocks, confirming implicit sequence learning; this pattern was consistent across both sessions and both task versions. Test-retest reliability was excellent for random-block reaction time (ICC {approx} .80 - .85), and modest for sequence blocks (ICC {approx} .29 - .53). General task activation (sequence and random performance relative to baseline) engaged the expected sensorimotor network, including the primary motor cortex (M1), premotor cortex, and SMA, with M1 more strongly engaged during sequence than random blocks. The sequence-specific contrast (SEQ > RND) additionally revealed six clusters spanning the left superior parietal cortex, bilateral supplementary and cingulate motor areas, bilateral secondary somatosensory cortex, and left ventral visual cortex together with cerebellar Crus I. The a priori defined motor network was engaged across all task phases and was anchored throughout by strong M1-SMA coupling, with overall network coupling following a non-monotonic course across the session, consistent with successive stages of resource allocation, consolidation, and automatisation. The gPPI analysis identified four sequence-specific connectivity clusters, dominated by a right-lateralised visuomotor network and a somatosensory-visual integration channel. ConclusionsTogether, these findings indicate a two-component architecture for SRTT performance: a left hemispheric cortico-striato-cerebellar network that supports performance and acquisition of the sequence, and a right hemispheric network associated with monitoring its predictable structure. Within this protocol, the SRTT produced robust and stable group-level signatures of implicit motor learning, yielding well-defined network targets for future tDCS interventions..

neuroscience↗

bioRxiv2026

BOLD signals of learning dynamics across and within trials beyond the classical fear and extinction network

Human functional magnetic resonance imaging studies of fear conditioning often average neural responses across trials, potentially obscuring transient activations that vary across learning. In this study with 139 participants, we examined finer temporal dynamics of conditioned responding by analyzing three 2s segments within the conditioned stimulus (CS) presentation period across each quarter of fear acquisition and extinction training. This approach revealed distinct, time-specific engagement of regions within fear- and safety-related networks, both within and across trials. In particular, different activation patterns emerged across the three trial segments during CS presentation, indicating that neural responses were not limited to CS onset. We observed a more classical activation pattern at 0s relative to CS onset that diverged in later trial segments, most notably involving the amygdala, hippocampus, and prefrontal cortex (PFC) structures such as vmPFC present exclusively in 2s and 4s trial segments. We also found sustained activations consistent across all blocks of trials, such as right vlPFC activation 4s after CS onset across all fear acquisition quarters. These findings suggest that conditioned fear and safety processing unfold as dynamic spatiotemporal cascades and highlight the importance of modeling later responses following CS onset rather than focusing exclusively on onset-related activation.

neuroscience↗

bioRxiv2026

Optimizing Network-Level TMS-fMRI: Benchmarking a Novel TMS-Compatible "Sushi" MR Coil

Concurrent TMS-fMRI can map how stimulation affects both the targeted cortex and connected brain-wide networks, but this requires MR receive hardware that allows TMS coil placement while preserving reliable whole-brain BOLD sensitivity. We developed and benchmarked a practical TMS-compatible "Sushi" MR receive setup assembled from two flexible 18-channel body arrays. Across six experiments, we tested functional readout validity, signal quality, and active TMS-fMRI compatibility. Resting-state fMRI (n = 12) and verbal N-back task-fMRI (n = 8) were acquired with Sushi, a commercially available 2x7-channel Surface setup, and a standard 64-channel head/neck array. Functional similarity to the 64-channel reference was quantified with spatial overlap, and multi-echo combination (MEcomb) was tested as a post-acquisition signal optimization strategy. Sushi recovered subject-specific resting-state networks that more closely matched the 64-channel reference than Surface, with no significant difference from the 64-channel test-retest reference. For task-fMRI, MEcomb increased task-map similarity for Sushi, whereas setup comparisons within each pipeline were not significant. MEcomb also improved resting-state similarity and increased temporal signal-to-noise ratio (tSNR) across receive setups. In phantom measurements, TMS coil placement produced spatially graded tSNR reductions relative to the no-TMS-coil condition, strongest near the coil. In one participant, active interleaved single-pulse TMS-fMRI over two cortical sites showed no detectable pulse-locked image artifacts; whole-brain MEcomb tSNR during active TMS-fMRI was reduced by 2.6-6.9% relative to the no-coil/no-stimulation reference. Together, Sushi and MEcomb provide complementary hardware and processing tools for TMS-compatible whole-brain fMRI. This combination supports network-level functional readouts while preserving feasibility for active interleaved TMS-fMRI.

neuroscience↗

bioRxiv2025

Predicting individual differences of fear and cognitive learning and extinction

The abilities to acquire new information and to modify previously learned knowledge are critical in an ever-changing world. However, the efficacy of learning is notably variable among individuals, with extinction learning being the epitome of such variability. Abundant studies have identified a core network of brain regions including amygdala, hippocampus, dorsal anterior cingulate cortex (ACC), ventromedial prefrontal cortex (PFC) and, more recently, the cerebellum, as key players in learning and extinction. Yet, the precise interactions within this network and their relationship to individual learning abilities and extinction have remained largely unexplored. In the present study, we examined how functional (FC), effective (EC), and structural (SC) connectivity patterns in the core learning network allow predicting individual differences in the efficacy of learning, extinction, and renewal. Analysing a large dataset of over 500 participants across a multitude of paradigms, our results revealed that FC predicted better acquisition, with a central role of ACC and hippocampus, whereas SC, involving ACC and amygdala, predicted higher levels of extinction learning. EC results suggested a predominantly inhibitory coupling among core learning network nodes, with paradigm-specific EC connectivity patterns predicting learning. Our predictions not only generalised between fear and cognitive predictive learning paradigms but were also successful in predicting learning from task-related FC and simulated data. Together, these results describe the multimodal neural determinants of learning, extinction, and renewal, and may inform individualised interventions for affective disorders based on neural connectivity patterns.

neuroscience↗

bioRxiv2025

Neural correlates of appetitive extinction learning: An fMRI study with actively participating pigeons

AbstractExtinction learning is an important learning process that enables adaptive and flexible behavior. Human neuroimaging studies show that the neural basis of extinction learning consists of a neural network that includes the hippocampus, amygdala, and subcomponents of the prefrontal cortex, but also extends beyond them. The limitations of applying fMRI to actively participating animals have so far restricted the identification of the entire extinction network in non-human animals. Here, we present the first fMRI study of extinction in awake and actively participating pigeons, using a Go/NoGo operant paradigm with a water reward. Our study revealed an extensive and largely left hemispheric telencephalic network of sensory, limbic, executive, and motor areas that slowly ceased to be active during the process of extinction learning. We propose that the onset of extinction ignites a neuronal updating of the associated consequences of own actions within a large telencephalic neural network until a new association is established which inhibits the previously acquired operant response.

neuroscience↗

bioRxiv2025

Multi-modal brain properties are associated with interindividual differences in fear acquisition and extinction

Interindividual differences in fear acquisition and extinction have been related to variation in specific brain correlates. However, variability in experimental setups complicates the integration of findings. Here, we present a combined fear acquisition (n = 101) and extinction (n = 88) experiment in which both phenomena were related to brain correlates obtained via functional magnetic resonance imaging in healthy, young participants. Correlates included regional brain volume, cortical surface area and thickness, neurite density and orientation dispersion, structural and functional connectivity. Fear responses were quantified as changes in skin conductance. Data from 376 brain areas and 70,500 network connections were used as independent variables in regularized regression models. Regression models of fear acquisition could be obtained for all modalities but regional brain volume. There were 284 predictors of which 77 appeared in exactly two models and 19 in exactly three. The latter primarily included brain areas from the somatosensory, insular, cingulate, and frontal cortices. Fear extinction yielded regression models based on neurite density, structural connectivity, and functional connectivity with 112 predictors in total. Two predictors, located in the dorsolateral prefrontal cortex, replicated across exactly two regression models (neurite density and structural connectivity). This study is the first to investigate the neural correlates of both fear acquisition and extinction in an explorative, multi-modal fMRI approach. Results show that numerous brain regions contribute to fear conditioning, some of them via more than one correlate. These findings call for further research to examine the potential interplay between brain correlates shaping fear conditioning.

neuroscience↗

bioRxiv2025

Electrophysiological resting-state signatures link polygenic scores to general intelligence

Intelligence is associated with important life outcomes. Behavioral, genetic, structural, and functional brain correlates of intelligence have been studied for decades, but questions remain as to how genetics are related to trait expression and what intermediary role brain properties play. This study investigated these mediations in a representative sample of 434 individuals, comprising young and older adults. Polygenic scores (PGS) for intelligence were calculated. Resting-state EEG recordings were analyzed using graph theory quantifying functional connectivity across different frequencies. We tested whether global and local graph metrics like efficiency and clustering mediated the association between PGS and intelligence. PGS significantly predicted variance in intelligence and were related to frequency-specific graph metrics in areas predominantly located in parieto-frontal regions, which in turn were associated with intelligence. These findings, which are based on the first study linking PGS to intelligence using EEG-derived graph metrics, advance our understanding of the neurogenetics of intelligence.

neuroscience↗

bioRxiv2025

Neurite density but not myelination of specific fiber tracts links polygenic scores to general intelligence

White matter is fundamental for efficient and accurate information transfer throughout the human brain and thus crucial for intelligence. Previous studies often demonstrated associations between fractional anisotropy (FA) as a metric of white matter "microstructural integrity" and intelligence, but it is still unclear, whether this relation is due to greater axon density, parallel, homogenous fiber orientation distributions, or greater myelination since all of these measures influence FA. Using neurite orientation dispersion and density imaging (NODDI) and myelin water fraction (MWF) imaging data, we analyzed the microstructural architecture of intelligence in more detail in a sample of 500 healthy young adults. Furthermore, we were interested whether specific white matter microstructural indices play intermediary roles in the pathway that links genetic disposition for intelligence to phenotype. Thus, we conducted for the first time mediation analyses investigating whether neurite density (NDI), orientation dispersion (ODI), and MWF of 64 white matter fiber tracts mediate the effects of polygenic scores for intelligence (PGSGI) on general intelligence. By doing so, we showed that NDI, but not ODI or MWF of white matter fiber tracts was significantly associated with general intelligence and that the NDI of six fiber tracts mediated the relation between genetic variability and g. These findings are a crucial step forward in decoding the neurogenetic underpinnings of general intelligence, as they identify that neurite density of specific fiber tracts relates polygenic variation to g, whereas orientation dispersion and myelination did not.

neuroscience↗

bioRxiv2024

The Simultaneous Measurement of Reversed Phase-Encoding EPI in a Single fMRI Session at 7T: Quantitative and Qualitative Evaluation of Geometric Distortion Correction in Submillimetre fMRI

The high temporal resolution of echo-planar imaging (EPI) has driven its widespread use in functional MRI (fMRI), and recent advancements in EPI have enabled the mapping of cortical layer-specific functional activities. Notwithstanding these technical innovations, geometric distortion correction in EPI remains essential to ensure accurate functional mapping onto anatomical references. A common approach for distortion correction involves the acquisition of an additional scan, with the phase-encoding direction reversed. However, this extra scan necessitates redundant acquisitions of calibration scans, significantly increasing acquisition time and energy deposition. To address this concern, this work presents a novel EPI scheme that simultaneously acquires both the original and reversed phase-encoding data within a single fMRI session. Furthermore, despite the widespread use of distortion correction, methods for qualitative and quantitative evaluation of its impact on submillimetre fMRI analysis have remained largely unexplored. This study acquires submillimetre visual fMRI data (0.73 x 0.73 mm2) at 7T using the suggested EPI scheme and evaluates the impact of distortion correction through various metrics proposed here, including spatial resolution, co-registration accuracy, functional mapping fidelity, and distribution of functional voxels. Our fMRI acquisition scheme effectively reduced redundant acquisition time and total energy deposition to subjects. Distortion-corrected fMRI data demonstrated significant improvements in co-registration with anatomical scans, thereby enhancing functional mapping accuracy. The improvements were achieved without significant degradation of spatial resolution or alteration of the functional activation distribution. These findings were verified through qualitative and quantitative assessments, highlighting the effectiveness of distortion correction in submillimetre fMRI. HighlightsO_LIWe present a novel fMRI scheme that effectively acquires reversed-PE EPI data C_LIO_LIOur fMRI scheme substantially reduces extra acquisition time and energy deposition C_LIO_LIThe impact of EPI distortion correction is evaluated for submillimetre fMRI C_LIO_LIWe propose diverse qualitative and quantitative metrics, enabling thorough analysis C_LIO_LIDistortion correction demonstrates its effectiveness for high-fidelity fMRI mapping C_LI

neuroscience↗

bioRxiv2024

Association between Adolescent Brain Morphometry and Cognitive Function: Insights from a Cross-Sectional Analysis of ABCD Data from 9 to 15 Years Old

Significant changes occur in brain structure and cognitive abilities during adolescence. Investigating their association can provide insight into brain-based cognitive development, yet previous studies were limited by narrow brain measures, small samples, and lacking focus on age-related variation. Here, we analyzed a large cohort (N = 8,534, age 9-15) from the Adolescent Brain Cognitive Development dataset. Using structural MRI and diffusion imaging, we derived 16 regional structural measures and integrated them via morphometric similarity networks to characterize 16,563 regional, connectivity, and hub features. We applied large-scale computational models to investigate their associations with performance on seven cognitive subtests and general intelligence (g), as well as age-related changes. Brain areas most strongly associated with cognitive ability also showed the greatest age-related variability in these associations, located primarily in the frontal, temporal, and occipital lobes. Structural MRI measures exhibited stronger associations with cognition and greater age-related variability than diffusion-derived metrics, while global hub measures showed stronger and more variable associations than local measures. Overall, our study provides a comprehensive and reliable understanding of brain structure-cognition associations during adolescence.

neuroscience↗

bioRxiv2022

Robust associations between white matter microstructure and general intelligence

Early research on the neural correlates of human intelligence was almost exclusively focused on gray matter properties. The advent of diffusion-weighted imaging led to an exponential growth of white matter brain imaging studies. However, this line of research has yielded mixed observations, especially about the relations between general intelligence and white matter microstructure. We used a multi-center approach to identify white matter regions that show replicable structure-function associations, employing data from four independent samples comprising over 2000 healthy participants. We used tract-based spatial statistics to examine associations between g factor scores and white matter microstructure and identified 188 voxels which exhibited positive associations between g factor scores and fractional anisotropy in all four data sets. Replicable voxels formed three clusters: one located around the forceps minor, crossing with extensions of the anterior thalamic radiation, the cingulum-cingulate gyrus, and the inferior fronto-occipital fasciculus in the left hemisphere, one located around the left-hemispheric superior longitudinal fasciculus, and one located around the left-hemispheric cingulum-cingulate gyrus, crossing with extensions of the anterior thalamic radiation and the inferior fronto-occipital fasciculus. Our results indicate that individual differences in general intelligence are robustly associated with white matter organization in specific fiber bundles.

neuroscience↗

bioRxiv2022

Structural architecture and brain network efficiency links polygenic scores to intelligence

Intelligence is highly heritable. Genome-wide association studies (GWAS) have shown that thousands of alleles contribute to variation in intelligence with small effect sizes. Polygenic scores (PGS), which combine these effects into one genetic summary measure, are increasingly used to investigate polygenic effects in independent samples. Whereas PGS explain a considerable amount of variance in intelligence, it is largely unknown how brain structure and function mediate this relationship. Here we show that individuals with higher PGS for educational attainment and intelligence had higher scores on cognitive tests, larger surface area, and more efficient fiber connectivity derived by graph theory. Fiber network efficiency as well as surface of brain areas partly located in parieto-frontal regions were found to mediate the relationship between PGS and cognitive performance. These findings are a crucial step forward in decoding the neurogenetic underpinnings of intelligence, as they identify specific regional networks that link polygenic predisposition to intelligence.

genetics↗

bioRxiv2020

Conditioned fear reactions are associated with gray matter density but not cortical microstructure

Research has shown that fear acquisition, in reaction to potentially harmful stimuli or situations, is characterized by pronounced interindividual differences. It is likely that such differences are evoked by variability in the macro- and microstructural properties of brain regions involved in the processing of threat or safety signals from the environment. Indeed, previous studies have shown that the strength of conditioned fear reactions is associated with the cortical thickness or volume of various brain regions. However, respective studies were exclusively targeted at single brain regions instead of whole brain networks. Here, we tested 60 young and healthy individuals in a differential fear conditioning paradigm while they underwent fMRI scanning. In addition, we acquired T1-weighted and multi-shell diffusion-weighted images prior to testing. We used task-based fMRI data to define global brain networks which exhibited increased BOLD responses towards CS+ or CS- presentations, respectively. From these networks, we obtained mean values of gray matter density, neurite density, and neurite orientation dispersion. We found that mean gray matter density averaged across the CS+ network was significantly correlated with the strength of conditioned fear reactions quantified via skin conductance response. Measures of neurite architecture were not associated with conditioned fear reaction in any of the two networks. Our results extend previous findings on the relationship between brain morphometry and fear learning. Most importantly, our study is the first to introduce neurite imaging to fear learning research and discusses how its implementation can be improved in future research.

neuroscience↗

bioRxiv2020

Polygenic Scores for Cognitive Abilities and their Association with Different Aspects of General Intelligence - a Deep Phenotyping Approach

Intelligence is a highly polygenic trait and genome-wide association studies (GWAS) have identified thousands of DNA variants contributing with small effects. Polygenic scores (PGS) can aggregate those effects for trait prediction in independent samples. As large-scale light-phenotyping GWAS operationalized intelligence as performance in rather superficial tests, the question arises which intelligence facets are actually captured. We used deep-phenotyping to investigate the molecular determinants of individual differences in cognitive ability. We therefore studied the association between PGS of intelligence (IQ-PGS), cognitive performance (CP-PGS) and educational attainment (EA-PGS) with a wide range of intelligence facets in a sample of 557 healthy adults. IQ-PGS, CP-PGS and EA-PGS had the highest incremental R2s for general (2.71%; 4.27%; 2.06%), verbal (3.30%; 4.64%; 1.61%) and numerical intelligence (3.06%; 3.24%; 1.26%) and the weakest for non-verbal intelligence (0.89%; 1.47%; 0.70%) and memory (0.80%; 1.06%; 0.67%). These results indicate that PGS derived from light-phenotyping GWAS do not reflect different facets of intelligence equally well, and thus should not be interpreted as genetic indicators of intelligence per se. The findings refine our understanding of how PGS are related to other traits or life outcomes.

genetics↗

bioRxiv2020

Fear learning sculpts functional brain connectivity at rest beyond the traditional fear network in humans

Neuroscientific research has identified specific brain networks involved in the acquisition of fear memories. Using fMRI to assess changes in resting-state functional connectivity (RSFC) induced by fear acquisition, single brain regions from these networks have also been linked to fear memory consolidation. However, previous studies only examined RSFC changes within restricted sets of brain regions or without a proper control group, leaving our knowledge about fear consolidation outside of traditional fear networks incomplete. Here, we tested a group of 84 healthy participants in a differential fear conditioning paradigm and quantified RSFC changes between 358 cortical and 16 subcortical brain areas. Subsequent to fear learning, 21 functional connections exhibited significant RSFC changes. Importantly, these connections were not restricted to the traditional fear networks but also comprised various frontal and visual areas. Our findings indicate that fear memory consolidation is a complex process that integrates relevant information across the entire brain.

neuroscience↗

bioRxiv2019

Myelin development in visual scene-network tracts beyond late childhood: A multimethod neuroimaging study

The visual scene-network--comprising the parahippocampal place area (PPA), retrosplenial cortex (RSC), and occipital place area (OPA)--shows a prolonged functional development. Structural development of white matter that underlies the scene-network has not been investigated despite its potential influence on scene-network function. The key factor for white matter maturation is myelination. However, research on myelination using the gold standard method of post-mortem histology is scarce. In vivo alternatives diffusion-weighed imaging (DWI) and myelin water imaging (MWI) so far report broad-scale findings that prohibit inferences concerning the scene-network. Here, we combine MWI, DWI tractography, and fMRI to investigate myelination in scene-network tracts in middle childhood, late childhood, and adulthood. We report increasing myelin from middle childhood to adulthood in right PPA-OPA, and trends towards increases in the left and right RSC-OPA tracts. Investigating tracts to regions highly connected with the scene-network, such as early visual cortex and the hippocampus did not yield any significant age group differences. Our findings indicate that structural development coincides with functional development in the scene-network, possibly enabling structure-function interactions.

neuroscience↗