Biology subjects

Schmaal, L.

Publications and source records attributed to Schmaal, L..

17 recordsLinked to original sources

bioRxiv2025

Image-Based Meta- and Mega-Analysis (IBMMA): A Unified Framework for Large-Scale, Multi-Site, Neuroimaging Data Analysis

The increasing scale and complexity of neuroimaging datasets aggregated from multiple study sites present substantial analytic challenges, as existing statistical analysis tools struggle to handle missing voxel-data, suffer from limited computational speed and inefficient memory allocation, and are restricted in the types of statistical designs they are able to model. We introduce Image-Based Meta- & Mega-Analysis (IBMMA), a novel software package implemented in R and Python that provides a unified framework for analyzing diverse neuroimaging features, efficiently handles large-scale datasets through parallel processing, offers flexible statistical modeling options, and properly manages missing voxel-data commonly encountered in multi-site studies. IBMMA produced stronger effect sizes and revealed findings in brain regions that traditional software overlooked due to missing voxel-data resulting in gaps in brain coverage. IBMMA has the potential to accelerate discoveries in neuroscience and enhance the clinical utility of neuroimaging findings.

neuroscience↗

bioRxiv2023

A Brain Model of Disturbed Self-Appraisal in Social Anxiety Disorder

The brains default mode network has a central role in the processing of information concerning oneself. Dysfunction in this self-referential processing represents a key component of multiple mental health conditions, including social anxiety disorder (SAD). This case-control study aimed to clarify alterations to network dynamics present during self-appraisal in SAD participants. A total of 38 adolescents and young adults with SAD and 72 healthy control participants underwent a self-referential processing fMRI task. The task involved two primary conditions of interest: direct self-appraisal (thinking about oneself) and reflected self-appraisal (thinking about how others might think about oneself). Dynamic causal modelling and parametric empirical Bayes were then used to explore differences in the effective connectivity of the default mode network between groups. We observed connectivity differences between SAD and healthy control participants in the reflected self-appraisal but not the direct self-appraisal condition. Specifically, SAD participants exhibited greater excitatory connectivity from the posterior cingulate cortex (PCC) to medial prefrontal cortex (MPFC) and greater inhibitory connectivity from the inferior parietal lobule (IPL) to MPFC. In contrast, in the absence of task modulation, SAD participants exhibited reduced intrinsic connectivity, with reduced excitatory connectivity from the PCC to MPFC and reduced inhibitory connectivity from the IPL to MPFC. As such, participants with SAD showed changes to afferent connections to the MPFC which occurred during both reflected self-appraisal as well as intrinsically. The presence of connectivity differences in reflected and not direct self-appraisal is consistent with the characteristic fear of negative social evaluation that is experienced by people with SAD.

neuroscience↗

bioRxiv2022

A Large-Scale ENIGMA Multisite Replication Study of Brain Age in Depression

BackgroundSeveral studies have evaluated whether depressed persons have older appearing brains than their nondepressed peers. However, the estimated neuroimaging-derived "brain age gap" has varied from study to study, likely driven by differences in training and testing sample (size), age range, and used modality/features. To validate our previously developed ENIGMA brain age model and the identified brain age gap, we aim to replicate the presence and effect size estimate previously found in the largest study in depression to date (N=2,126 controls & N=2,675 cases; +1.08 years [SE 0.22], Cohens d=0.14, 95% CI: 0.08-0.20), in independent cohorts that were not part of the original study. MethodsA previously trained brain age model (www.photon-ai.com/enigma_brainage) based on 77 FreeSurfer brain regions of interest was used to obtain unbiased brain age predictions in 751 controls and 766 persons with depression (18-75 years) from 13 new cohorts collected from 20 different scanners. ResultsOur ENIGMA MDD brain age model generalized reasonably well to controls from the new cohorts (predicted age vs. age: r = 0.73, R2=0.47, MAE=7.50 years), although the performance varied from cohort to cohort. In these new cohorts, on average, depressed persons showed a significantly higher brain age gap of +1 year (SE 0.35) (Cohens d{square}={square}{square}.15, 95% CI: 0.05-0.25) compared with controls, highly similar to our previous finding. ConclusionsThis study further validates our previously developed ENIGMA brain age algorithm. Importantly, we replicated the brain age gap in depression with a comparable effect size. Thus, two large-scale independent mega-analyses across in total 32 cohorts and >3,400 patients and >2,800 controls worldwide show reliable but subtle effects of brain aging in adult depression.

neuroscience↗

bioRxiv2022

Global brain health modulates the impact of lesion damage on post-stroke sensorimotor outcomes

Sensorimotor performance after stroke is strongly related to focal injury measures such as corticospinal tract lesion load. However, the role of global brain health is less clear. Here, we examined the impact of brain age, a measure of neurobiological aging derived from whole brain structural neuroimaging, on sensorimotor outcomes. We hypothesized that stroke lesion damage would result in older brain age, which would in turn be associated with poorer sensorimotor outcomes. We also expected that brain age would mediate the impact of lesion damage on sensorimotor outcomes and that these relationships would be driven by post-stroke secondary atrophy (e.g., strongest in the ipsilesional hemisphere in chronic stroke). We further hypothesized that structural brain resilience, which we define in the context of stroke as the brains ability to maintain its global integrity despite focal lesion damage, would differentiate people with better versus worse outcomes. We analyzed cross-sectional high-resolution brain MRI and outcomes data from 963 people with stroke from 38 cohorts worldwide using robust linear mixed-effects regressions to examine the relationship between sensorimotor behavior, lesion damage, and brain age. We used a mediation analysis to examine whether brain age mediates the impact of lesion damage on stroke outcomes and if associations are driven by ipsilesional measures in chronic ([≥]180 days) stroke. We assessed the impact of brain resilience on sensorimotor outcome using logistic regression with propensity score matching on lesion damage. Stroke lesion damage was associated with older brain age, which in turn was associated with poorer sensorimotor outcomes. Brain age mediated the impact of corticospinal tract lesion load on sensorimotor outcomes most strongly in the ipsilesional hemisphere in chronic stroke. Greater brain resilience, as indexed by younger brain age, explained why people have better versus worse sensorimotor outcomes when lesion damage was fixed. We present novel evidence that global brain health is associated with superior post-stroke sensorimotor outcomes and modifies the impact of focal damage. This relationship appears to be due to post-stroke secondary degeneration. Brain resilience provides insight into why some people have better outcomes after stroke, despite similar amounts of focal injury. Inclusion of imaging-based assessments of global brain health may improve prediction of post-stroke sensorimotor outcomes compared to focal injury measures alone. This investigation is important because it introduces the potential to apply novel therapeutic interventions to prevent or slow brain aging from other fields (e.g., Alzheimers disease) to stroke.

neuroscience↗

bioRxiv2022

Molecular and connectomic vulnerability shape cross-disorder cortical abnormalities

Numerous brain disorders demonstrate structural brain abnormalities, which are thought to arise from molecular perturbations or connectome miswiring. The unique and shared contributions of these molecular and connectomic vulnerabilities to brain disorders remain unknown, and has yet to be studied in a single multi-disorder framework. Using MRI morphometry from the ENIGMA consortium, we construct maps of cortical abnormalities for thirteen neurodevelopmental, neurological, and psychiatric disorders from N = 21 000 patients and N = 26 000 controls, collected using a harmonized processing protocol. We systematically compare cortical maps to multiple micro-architectural measures, including gene expression, neurotransmitter density, metabolism, and myelination (molecular vulnerability), as well as global connectomic measures including number of connections, centrality, and connection diversity (connectomic vulnerability). We find that regional molecular vulnerability and macroscale brain network architecture interact to drive the spatial patterning of cortical abnormalities in multiple disorders. Local attributes, particularly neurotransmitter receptor profiles, constitute the best predictors of both disorder-specific cortical morphology and cross-disorder similarity. Finally, we find that cross-disorder abnormalities are consistently subtended by a small subset of network epicentres in bilateral sensory-motor, medial temporal lobe, precuneus, and superior parietal cortex. Collectively, our results highlight how local biological attributes and global connectivity jointly shape cross-disorder cortical abnormalities.

neuroscience↗

bioRxiv2021

Multilevel neural gradients reflect transdiagnostic effects of major psychiatric conditions on cortical morphology

AO_SCPLOWBSTRACTC_SCPLOWIt is increasingly recognized that multiple psychiatric conditions are underpinned by shared neural pathways, affecting similar brain systems. Here, we assessed i) shared dimensions of alterations in cortical morphology across six major psychiatric conditions (autism spectrum disorder, attention deficit/hyperactivity disorder, major depression, obsessive-compulsive disorder, bipolar disorder, schizophrenia) and ii) carried out a multiscale neural contextualization, by cross-referencing shared anomalies against cortical myeloarchitecture and cytoarchitecture, as well as connectome and neurotransmitter organization. Pooling disease-related effects on MRI-based cortical thickness measures across six ENIGMA working groups, including a total of 28,546 participants (12,876 patients and 15,670 controls), we computed a shared disease dimension on cortical morphology using principal component analysis that described a sensory-fugal pattern with paralimbic regions showing the most consistent abnormalities across conditions. The shared disease dimension was closely related to cortical gradients of microstructure and intrinsic connectivity, as well as neurotransmitter systems, specifically serotonin and dopamine. Our findings embed the shared effects of major psychiatric conditions on brain structure in multiple scales of brain organization and may provide novel insights into neural mechanisms into transdiagnostic vulnerability.

neuroscience↗

bioRxiv2021

ENIGMA HALFpipe: Interactive, reproducible, and efficient analysis for resting-state and task-based fMRI data

The reproducibility crisis in neuroimaging has led to an increased demand for standardized data processing workflows. Within the ENIGMA consortium, we developed HALFpipe (Harmonized AnaLysis of Functional MRI pipeline), an open-source, containerized, user-friendly tool that facilitates reproducible analysis of task-based and resting-state fMRI data through uniform application of preprocessing, quality assessment, single-subject feature extraction, and group-level statistics. It provides state-of-the-art preprocessing using fMRIPrep without the requirement for input data in Brain Imaging Data Structure (BIDS) format. HALFpipe extends the functionality of fMRIPrep with additional preprocessing steps, which include spatial smoothing, grand mean scaling, temporal filtering, and confound regression. HALFpipe generates an interactive quality assessment (QA) webpage to assess the quality of key preprocessing outputs and raw data in general. HALFpipe features myriad post-processing functions at the individual subject level, including calculation of task-based activation, seed-based connectivity, network-template (or dual) regression, atlas-based functional connectivity matrices, regional homogeneity (ReHo), and fractional amplitude of low frequency fluctuations (fALFF), offering support to evaluate a combinatorial number of features or preprocessing settings in one run. Finally, flexible factorial models can be defined for mixed-effects regression analysis at the group level, including multiple comparison correction. Here, we introduce the theoretical framework in which HALFpipe was developed, and present an overview of the main functions of the pipeline. HALFpipe offers the scientific community a major advance toward addressing the reproducibility crisis in neuroimaging, providing a workflow that encompasses preprocessing, post-processing, and QA of fMRI data, while broadening core principles of data analysis for producing reproducible results. Instructions and code can be found at https://github.com/HALFpipe/HALFpipe.

neuroscience↗

bioRxiv2021

Accommodating site variation in neuroimaging data using hierarchical and Bayesian models

AO_SCPLOWBSTRACTC_SCPLOWThe potential of normative modeling to make individualized predictions from neuroimaging data has enabled inferences that go beyond the case-control approach. However, site effects are often confounded with variables of interest in a complex manner and can bias estimates of normative models, which has impeded the application of normative models to large multi-site neuroimaging data sets. In this study, we suggest accommodating for these site effects by including them as random effects in a hierarchical Bayesian model. We compared the performance of a linear and a non-linear hierarchical Bayesian model in modeling the effect of age on cortical thickness. We used data of 570 healthy individuals from the ABIDE (autism brain imaging data exchange) data set in our experiments. In addition, we used data from individuals with autism to test whether our models are able to retain clinically useful information while removing site effects. We compared the proposed single stage hierarchical Bayesian method to several harmonization techniques commonly used to deal with additive and multiplicative site effects using a two stage regression, including regressing out site and harmonizing for site with ComBat, both with and without explicitly preserving variance related to age and sex as biological variation of interest. In addition, we made predictions from raw data, in which site has not been accommodated for. The proposed hierarchical Bayesian method showed the best predictive performance according to multiple metrics. Beyond that, the resulting z-scores showed little to no residual site effects, yet still retained clinically useful information. In contrast, performance was particularly poor for the regression model and the ComBat model in which age and sex were not explicitly modeled. In all two stage harmonization models, predictions were poorly scaled, suffering from a loss of more than 90 % of the original variance. Our results show the value of hierarchical Bayesian regression methods for accommodating site variation in neuroimaging data, which provides an alternative to harmonization techniques. While the approach we propose may have broad utility, our approach is particularly well suited to normative modelling where the primary interest is in accurate modelling of inter-subject variation and statistical quantification of deviations from a reference model. 1 HighlightsO_LIDevelopment and presentation of normative modeling approach based on hierarchical Bayesian modeling that can be applied to large multi-site neuroimaging data sets. C_LIO_LIComparison of performance of Hierarchical Bayesian model including site as predictor to several common ways to harmonize for multi-site effects. C_LIO_LIPresentation of normative modeling as site correction tool. C_LI

bioinformatics↗

bioRxiv2020

Controlling for effects of confounding variables on machine learning predictions

Machine learning predictive models are being used in neuroimaging to predict information about the task or stimuli or to identify potentially clinically useful biomarkers. However, the predictions can be driven by confounding variables unrelated to the signal of interest, such as scanner effect or head motion, limiting the clinical usefulness and interpretation of machine learning models. The most common method to control for confounding effects is regressing out the confounding variables separately from each input variable before machine learning modeling. However, we show that this method is insufficient because machine learning models can learn information from the data that cannot be regressed out. Instead of regressing out confounding effects from each input variable, we propose controlling for confounds post-hoc on the level of machine learning predictions. This allows partitioning of the predictive performance into the performance that can be explained by confounds and performance independent of confounds. This approach is flexible and allows for parametric and non-parametric confound adjustment. We show in real and simulated data that this method correctly controls for confounding effects even when traditional input variable adjustment produces false-positive findings.

bioinformatics↗

bioRxiv2020

Individual deviations from normative models of brain structure in a large cross-sectional schizophrenia cohort

BackgroundThe heterogeneity of schizophrenia has defied efforts to derive reproducible and definitive anatomical maps of structural brain changes associated with the disorder. We aimed to map deviations from normative ranges of brain structure for individual patients and evaluate whether the loci of individual deviations recapitulated group-average brain maps of schizophrenia pathology. MethodsFor each of 48 white matter tracts and 68 cortical regions, normative percentiles of variation in fractional anisotropy (FA) and cortical thickness (CT) were established using diffusion-weighted and structural MRI from healthy adults (n=195). Individuals with schizophrenia (n=322) were classified as either within the normative range for healthy individuals of the same age and sex (5-95% percentiles), infra-normal (<5% percentile) or supra-normal (>95% percentile). Repeating this classification for each tract and region yielded a deviation map for each individual. ResultsCompared to the healthy comparison group, the schizophrenia group showed widespread reductions in FA and CT, involving virtually all white matter tracts and cortical regions. Paradoxically, however, no more than 15-20% of patients deviated from the normative range for any single tract or region, whereas 79% of patients showed infra-normal deviations for at least one locus (healthy individuals: 59{+/-}2%, p<0.001). Higher polygenic risk for schizophrenia associated with a greater number of regions with infra-normal deviations in CT (r=-0.17, p=0.006). ConclusionsAnatomical loci of schizophrenia-related changes are highly heterogeneous across individuals to the extent that group-consensus pathological maps are not representative of most individual patients. Normative modeling can aid in parsing schizophrenia heterogeneity and guiding personalized interventions.

neuroscience↗

bioRxiv2019

Structural Brain Imaging Studies Offer Clues about the Effects of the Shared Genetic Etiology among Neuropsychiatric Disorders

BackgroundGenomewide association studies have found significant genetic correlations among many neuropsychiatric disorders. In contrast, we know much less about the degree to which structural brain alterations are similar among disorders and, if so, the degree to which such similarities have a genetic etiology. MethodsFrom the Enhancing Neuroimaging Genetics through Meta-Analysis (ENIGMA) consortium, we acquired standardized mean differences (SMDs) in regional brain volume and cortical thickness between cases and controls. We had data on 41 brain regions for: attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), bipolar disorder (BD), epilepsy, major depressive disorder (MDD), obsessive compulsive disorder (OCD) and schizophrenia (SCZ). These data had been derived from 24,360 patients and 37,425 controls. ResultsThe SMDs were significantly correlated between SCZ and BD, OCD, MDD, and ASD. MDD was positively correlated with BD and OCD. BD was positively correlated with OCD and negatively correlated with ADHD. These pairwise correlations among disorders were correlated with the corresponding pairwise correlations among disorders derived from genomewide association studies (r = 0.49). ConclusionsOur results show substantial similarities in sMRI phenotypes among neuropsychiatric disorders and suggest that these similarities are accounted for, in part, by corresponding similarities in common genetic variant architectures.

neuroscience↗

bioRxiv2019

Neurovegetative symptom subtypes in young people with major depressive disorder and their structural brain correlates

BackgroundDepression is a leading cause of burden of disease among young people. Current treatments are not uniformly effective, in part due to the heterogeneous nature of major depressive disorder (MDD). Refining MDD into more homogeneous subtypes is an important step towards identifying underlying pathophysiological mechanisms and improving treatment of young people. In adults, symptom-based subtypes of depression identified using data-driven methods mainly differed in patterns of neurovegetative symptoms (sleep and appetite/weight). These subtypes have been associated with differential biological mechanisms, including immuno-metabolic markers, genetics and brain alterations (mainly in the ventral striatum and insular cortex).\n\nMethodsK-means clustering was applied to individual depressive symptoms from the Quick Inventory of Depressive Symptoms (QIDS) in 275 young people (15-25 years old) with MDD to identify symptom-based subtypes, and in 244 young people from an independent dataset (a subsample of the STAR*D dataset). Insula surface area and thickness and ventral striatum volume were compared between the subtypes using structural MRI.\n\nResultsThree subtypes were identified in the discovery dataset and replicated in the independent dataset; severe depression with increased appetite, severe depression with decreased appetite and severe insomnia, and moderate depression. The severe increased appetite subtype showed lower surface area in the anterior insula compared to both healthy controls and the moderate subtype.\n\nConclusionsOur findings in young people replicate the previously identified symptom-based depression subtypes in adults. The structural alterations of the anterior insular cortex add to the existing evidence of different pathophysiological mechanisms involved in this subtype.

neuroscience↗

bioRxiv2019

Beyond accuracy: Measures for assessing machine learning models, pitfalls and guidelines

Pattern recognition predictive models have become an important tool for analysis of neuroimaging data and answering important questions from clinical and cognitive neuroscience. Regardless of the application, the most commonly used method to quantify model performance is to calculate prediction accuracy, i.e. the proportion of correctly classified samples. While simple and intuitive, other performance measures are often more appropriate with respect to many common goals of neuroimaging pattern recognition studies. In this paper, we will review alternative performance measures and focus on their interpretation and practical aspects of model evaluation. Specifically, we will focus on 4 families of performance measures: 1) categorical performance measures such as accuracy, 2) rank based performance measures such as the area under the curve, 3) probabilistic performance measures based on quadratic error such as Brier score, and 4) probabilistic performance measures based on information criteria such as logarithmic score. We will examine their statistical properties in various settings using simulated data and real neuroimaging data derived from public datasets. Results showed that accuracy had the worst performance with respect to statistical power, detecting model improvement, selecting informative features and reliability of results. Therefore in most cases, it should not be used to make statistical inference about model performance. Accuracy should also be avoided for evaluating utility of clinical models, because it does not take into account clinically relevant information, such as relative cost of false-positive and false-negative misclassification or calibration of probabilistic predictions. We recommend alternative evaluation criteria with respect to the goals of a specific machine learning model.

neuroscience↗

bioRxiv2019

Subcortical Shape Alterations in Major Depressive Disorder: Findings from the ENIGMA Major Depressive Disorder Working Group

Alterations in regional subcortical brain volumes have been widely investigated as part of the efforts of an international consortium, ENIGMA, to determine reliable structural brain signatures for Major Depressive Disorder (MDD). Given that subcortical structures are comprised of distinct subfields, we sought to build significantly from prior work to precisely map localized MDD-related differences in subcortical regions using shape analysis. In this meta-analysis of subcortical shape from the ENIGMA-MDD working group, we compared 1,781 patients with MDD and 2,953 healthy controls (CTL) on individual measures of shape metrics (thickness and surface area) on the surface of seven bilateral subcortical structures: nucleus accumbens, amygdala, caudate, hippocampus, pallidum, putamen, and thalamus. Harmonized data processing and statistical analyses were conducted locally at each site, and findings were aggregated by meta-analysis. Relative to CTL, patients with MDD had lower surface area in the subiculum of the hippocampus, the basolateral amygdala, and the nucleus accumbens shell. Relative to CTL, patients with adolescent-onset MDD ([≤] 21 years) had lower thickness and surface area of the subiculum of the hippocampus and the basolateral amygdala. Relative to first-episode MDD, recurrent MDD patients had lower thickness and surface area in the CA1 of the hippocampus and the basolateral amygdala. Our results suggest that previously reported MDD-associated volumetric differences may be localized to specific subfields of these structures that have been shown to be sensitive to the effects of stress, with important implications for mapping treatments to patients based on specific neural targets and key clinical features.

neuroscience↗

bioRxiv2018

Evaluating the evidence for biotypes of depression: attempted replication of Drysdale et.al. 2017

BackgroundPsychiatric disorders are highly heterogeneous, defined based on symptoms with little connection to potential underlying biological mechanisms. A possible approach to dissect biological heterogeneity is to look for biologically meaningful subtypes. A recent study Drysdale et al. (2017) showed promising results along this line by simultaneously using resting state fMRI and clinical data and identified four distinct subtypes of depression with different clinical profiles and abnormal resting state fMRI connectivity. These subtypes were predictive of treatment response to transcranial magnetic stimulation therapy.\n\nObjectiveHere, we attempted to replicate the procedure followed in the Drysdale et al. study and their findings in an independent dataset of a clinically more heterogeneous sample of 187 participants with depression and anxiety. We aimed to answer the following questions: 1) Using the same procedure, can we find a statistically significant and reliable relationship between brain connectivity and clinical symptoms? 2) Is the observed relationship similar to the one found in the original study? 3) Can we identify distinct and reliable subtypes? 4) Do they have similar clinical profiles as the subtypes identified in the original study?\n\nMethodsWe followed the original procedure as closely as possible, including a canonical correlation analysis to find a low dimensional representation of clinically relevant resting state fMRI features, followed by hierarchical clustering to identify subtypes. We extended the original procedure using additional statistical tests, to test the statistical significance of the relationship between resting state fMRI and clinical data, and the existence of distinct subtypes. Furthermore, we examined the stability of the whole procedure using resampling.\n\nResults and ConclusionWe were not able to replicate the findings of the original study. Relationships between brain connectivity and clinical symptoms were not statistically significant and we also did not find clearly distinct subtypes of depression. We argue, that based on our rigorous approach and in-depth review of the original results, that the evidence for the existence of the distinct resting state connectivity based subtypes of depression is weak and should be interpreted with caution.

neuroscience↗

bioRxiv2018

Deep Learning for Quality Control of Subcortical Brain 3D Shape Models

We present several deep learning models for assessing the morphometric fidelity of deep grey matter region models extracted from brain MRI. We test three different convolutional neural net architectures (VGGNet, ResNet and Inception) over 2D maps of geometric features. Further, we present a novel geometry feature augmentation technique based on parametric spherical mapping. Finally, we present an approach for model decision visualization, allowing human raters to see the areas of subcortical shapes most likely to be deemed of failing quality by the machine. Our training data is comprised of 5200 subjects from the ENIGMA Schizophrenia MRI cohorts, and our test dataset contains 1500 subjects from the ENIGMA Major Depressive Disorder cohorts. Our final models reduce human rater time by 46-70%. ResNet outperforms VGGNet and Inception for all of our predictive tasks.

neuroscience↗

bioRxiv2017

Machine Learning for Large-Scale Quality Control of 3D Shape Models in Neuroimaging

As very large studies of complex neuroimaging phenotypes become more common, human quality assessment of MRI-derived data remains one of the last major bottlenecks. Few attempts have so far been made to address this issue with machine learning. In this work, we optimize predictive models of quality for meshes representing deep brain structure shapes. We use standard vertex-wise and global shape features computed homologously across 19 cohorts and over 7500 human-rated subjects, training kernelized Support Vector Machine and Gradient Boosted Decision Trees classifiers to detect meshes of failing quality. Our models generalize across datasets and diseases, reducing human workload by 30-70%, or equivalently hundreds of human rater hours for datasets of comparable size, with recall rates approaching inter-rater reliability.

neuroscience↗