A Systematic Evaluation of Machine Learning-Based Biomarkers for Major Depressive Disorder.

Importance: Biological psychiatry aims to understand mental disorders in terms of altered neurobiological pathways. However, for one of the most prevalent and disabling mental disorders, major depressive disorder (MDD), no informative biomarkers have been identified. Objective: To evaluate whether machine learning (ML) can identify a multivariate biomarker for MDD. Design, Setting, and Participants: This study used data from the Marburg-Münster Affective Disorders Cohort Study, a case-control clinical neuroimaging study. Patients with acute or lifetime MDD and healthy controls aged 18 to 65 years were recruited from primary care and the general population in Münster and Marburg, Germany, from September 11, 2014, to September 26, 2018. The Münster Neuroimaging Cohort (MNC) was used as an independent partial replication sample. Data were analyzed from April 2022 to June 2023. Exposure: Patients with MDD and healthy controls. Main Outcome and Measure: Diagnostic classification accuracy was quantified on an individual level using an extensive ML-based multivariate approach across a comprehensive range of neuroimaging modalities, including structural and functional magnetic resonance imaging and diffusion tensor imaging as well as a polygenic risk score for depression. Results: Of 1801 included participants, 1162 (64.5%) were female, and the mean (SD) age was 36.1 (13.1) years. There were a total of 856 patients with MDD (47.5%) and 945 healthy controls (52.5%). The MNC replication sample included 1198 individuals (362 with MDD [30.1%] and 836 healthy controls [69.9%]). Training and testing a total of 4 million ML models, mean (SD) accuracies for diagnostic classification ranged between 48.1% (3.6%) and 62.0% (4.8%). Integrating neuroimaging modalities and stratifying individuals based on age, sex, treatment, or remission status does not enhance model performance. Findings were replicated within study sites and also observed in structural magnetic resonance imaging within MNC. Under simulated conditions of perfect reliability, performance did not significantly improve. Analyzing model errors suggests that symptom severity could be a potential focus for identifying MDD subgroups. Conclusion and Relevance: Despite the improved predictive capability of multivariate compared with univariate neuroimaging markers, no informative individual-level MDD biomarker-even under extensive ML optimization in a large sample of diagnosed patients-could be identified.

Human fertilization in vivo and in vitro requires the CatSper channel to initiate sperm hyperactivation.

The infertility of many couples rests on an enigmatic dysfunction of the man's sperm. To gain insight into the underlying pathomechanisms, we assessed the function of the sperm-specific multisubunit CatSper-channel complex in the sperm of almost 2,300 men undergoing a fertility workup, using a simple motility-based test. We identified a group of men with normal semen parameters but defective CatSper function. These men or couples failed to conceive naturally and upon medically assisted reproduction via intrauterine insemination and in vitro fertilization. Intracytoplasmic sperm injection (ICSI) was, ultimately, required to conceive a child. We revealed that the defective CatSper function was caused by variations in CATSPER genes. Moreover, we unveiled that CatSper-deficient human sperm were unable to undergo hyperactive motility and, therefore, failed to penetrate the egg coat. Thus, our study provides the experimental evidence that sperm hyperactivation is required for human fertilization, explaining the infertility of CatSper-deficient men and the need of ICSI for medically assisted reproduction. Finally, our study also revealed that defective CatSper function and ensuing failure to hyperactivate represents the most common cause of unexplained male infertility known thus far and that this sperm channelopathy can readily be diagnosed, enabling future evidence-based treatment of affected couples.

Therapy-induced modulation of tumor vasculature and oxygenation in a murine glioblastoma model quantified by deep learning-based feature extraction.

Glioblastoma presents characteristically with an exuberant, poorly functional vasculature that causes malperfusion, hypoxia and necrosis. Despite limited clinical efficacy, anti-angiogenesis resulting in vascular normalization remains a promising therapeutic approach. Yet, fundamental questions concerning anti-angiogenic therapy remain unanswered, partly due to the scale and resolution gap between microscopy and clinical imaging and a lack of quantitative data readouts. To what extend does treatment lead to vessel regression or vessel normalization and does it ameliorate or aggravate hypoxia? Clearly, a better understanding of the underlying mechanisms would greatly benefit the development of desperately needed improved treatment regimens. Here, using orthotopic transplantation of Gli36 cells, a widely used murine glioma model, we present a mesoscopic approach based on light sheet fluorescence microscopic imaging of wholemount stained tumors. Deep learning-based segmentation followed by automated feature extraction allowed quantitative analyses of the entire tumor vasculature and oxygenation statuses. Unexpectedly in this model, the response to both cytotoxic and anti-angiogenic therapy was dominated by vessel normalization with little evidence for vessel regression. Equally surprising, only cytotoxic therapy resulted in a significant alleviation of hypoxia. Taken together, we provide and evaluate a quantitative workflow that addresses some of the most urgent mechanistic questions in anti-angiogenic therapy.

Deep learning predicts therapy-relevant genetics in acute myeloid leukemia from Pappenheim-stained bone marrow smears.

ABSTRACT: The detection of genetic aberrations is crucial for early therapy decisions in acute myeloid leukemia (AML) and recommended for all patients. Because genetic testing is expensive and time consuming, a need remains for cost-effective, fast, and broadly accessible tests to predict these aberrations in this aggressive malignancy. Here, we developed a novel fully automated end-to-end deep learning pipeline to predict genetic aberrations directly from single-cell images from scans of conventionally stained bone marrow smears already on the day of diagnosis. We used this pipeline to compile a multiterabyte data set of >2 000 000 single-cell images from diagnostic samples of 408 patients with AML. These images were then used to train convolutional neural networks for the prediction of various therapy-relevant genetic alterations. Moreover, we created a temporal test cohort data set of >444 000 single-cell images from further 71 patients with AML. We show that the models from our pipeline can significantly predict these subgroups with high areas under the curve of the receiver operating characteristic. Potential genotype-phenotype links were visualized with 2 different strategies. Our pipeline holds the potential to be used as a fast and inexpensive automated tool to screen patients with AML for therapy-relevant genetic aberrations directly from routine, conventionally stained bone marrow smears already on the day of diagnosis. It also creates a foundation to develop similar approaches for other bone marrow disorders in the future.

Trail using ants follow idiosyncratic routes in complex landscapes.

A large volume of research on individually navigating ants has shown how path integration and visually guided navigation form a major part of the ant navigation toolkit for many species and are sufficient mechanisms for successful navigation. One of the behavioural markers of the interaction of these mechanisms is that experienced foragers develop idiosyncratic routes that require that individual ants have personal and unique visual memories that they use to guide habitual routes between the nest and feeding sites. The majority of ants, however, inhabit complex cluttered environments and social pheromone trails are often part of the collective recruitment, organisation and navigation of these foragers. We do not know how individual navigation interacts with collective behaviour along shared trails in complex natural environments. We thus asked here if wood ants that forage through densely cluttered woodlands where they travel along shared trails repeatedly follow the same routes or if they choose a spread of paths within the shared trail. We recorded three long homing trajectories of 20 individual wood ants in their natural woodland habitat. We found that wood ants follow idiosyncratic routes when navigating along shared trails through highly complex visual landscapes. This shows that ants rely on individual memories for habitual route guidance even in cluttered environments when chemical trail information is available. We argue that visual cues are likely to be the dominant sensory modality for the idiosyncratic routes. These experiments shed new light on how ants, or insects in general, navigate through complex multimodal environments.

Event-driven adaptive optical neural network.

We present an adaptive optical neural network based on a large-scale event-driven architecture. In addition to changing the synaptic weights (synaptic plasticity), the optical neural network's structure can also be reconfigured enabling various functionalities (structural plasticity). Key building blocks are wavelength-addressable artificial neurons with embedded phase-change materials that implement nonlinear activation functions and nonvolatile memory. Using multimode focusing, the activation function features both excitatory and inhibitory responses and shows a reversible switching contrast of 3.2 decibels. We train the neural network to distinguish between English and German text samples via an evolutionary algorithm. We investigate both the synaptic and structural plasticity during the training process. On the basis of this concept, we realize a large-scale network consisting of 736 subnetworks with 16 phase-change material neurons each. Overall, 8398 neurons are functional, highlighting the scalability of the photonic architecture.

Immersive training of clinical decision making with AI driven virtual patients - a new VR platform called medical tr.AI.ning.

Background: Medical students need to be prepared for various situations in clinical decision-making that cannot be systematically trained with real patients without risking their health or integrity. To target system-related limitations of actor-based training, digital learning methods are increasingly used in medical education, with virtual reality (VR)- training seeming to have high potential. Virtually generated training scenarios allow repetitive training of highly relevant clinical skills within a protected, realistic learning environment. Thanks to Artificial Intelligence (AI), face-to-face interaction with virtual agents is feasible. Combining this technology with VR-simulations offers a new way of situated context-based, first-person training for medical students. Project goal and method: The authors' aim is to develop a modular digital training platform for medical education with virtual, interactable agents and to integrate this platform into the medical curriculum. The medical tr.AI.ning platform will provide veridical simulation of clinical scenarios with virtual patients, augmented with highly realistic medical pathologies within a customizable, realistic situational context. Medical tr.AI.ning is scaled to four complementary developmental steps with different scenarios that can be used separately and so each outcome can successively be integrated early within the project. Every step has its own focus (visual, movement, communication, combination) and extends an author toolbox through its modularity. The modules of each step will be specified and designed together with medical didactics experts. Perspective: To ensure constant improvement of user experience, realism, and medical validity, the authors will perform regular iterative evaluation rounds.Furthermore, integration of medical tr.AI.ning into the medical curriculum will enable long-term and large-scale detection of benefits and limitations of this approach, providing enhanced alternative teaching paradigms for VR technology.

Inverse design of nanophotonic devices using dynamic binarization.

The complexity of applications addressed with photonic integrated circuits is steadily rising and poses increasingly challenging demands on individual component functionality, performance and footprint. Inverse design methods have recently shown great promise to address these demands using fully automated design procedures that enable access to non-intuitive device layouts beyond conventional nanophotonic design concepts. Here we present a dynamic binarization method for the objective-first algorithm that lies at the core of the currently most successful inverse design algorithms. Our results demonstrate significant performance advantages over previous implementations of objective first algorithms, which we show for a fundamental TE00 to TE20 waveguide mode converter both in simulation and in experiments with fabricated devices.

CATER: Combined Animal Tracking & Environment Reconstruction.

Quantifying the behavior of small animals traversing long distances in complex environments is one of the most difficult tracking scenarios for computer vision. Tiny and low-contrast foreground objects have to be localized in cluttered and dynamic scenes as well as trajectories compensated for camera motion and drift in multiple lengthy recordings. We introduce CATER, a novel methodology combining an unsupervised probabilistic detection mechanism with a globally optimized environment reconstruction pipeline enabling precision behavioral quantification in natural environments. Implemented as an easy to use and highly parallelized tool, we show its application to recover fine-scale motion trajectories, registered to a high-resolution image mosaic reconstruction, of naturally foraging desert ants from unconstrained field recordings. By bridging the gap between laboratory and field experiments, we gain previously unknown insights into ant navigation with respect to motivational states, previous experience, and current environments and provide an appearance-agnostic method applicable to study the behavior of a wide range of terrestrial species under realistic conditions.

An overview and a roadmap for artificial intelligence in hematology and oncology.

BACKGROUND: Artificial intelligence (AI) is influencing our society on many levels and has broad implications for the future practice of hematology and oncology. However, for many medical professionals and researchers, it often remains unclear what AI can and cannot do, and what are promising areas for a sensible application of AI in hematology and oncology. Finally, the limits and perils of using AI in oncology are not obvious to many healthcare professionals. METHODS: In this article, we provide an expert-based consensus statement by the joint Working Group on "Artificial Intelligence in Hematology and Oncology" by the German Society of Hematology and Oncology (DGHO), the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), and the Special Interest Group Digital Health of the German Informatics Society (GI). We provide a conceptual framework for AI in hematology and oncology. RESULTS: First, we propose a technological definition, which we deliberately set in a narrow frame to mainly include the technical developments of the last ten years. Second, we present a taxonomy of clinically relevant AI systems, structured according to the type of clinical data they are used to analyze. Third, we show an overview of potential applications, including clinical, research, and educational environments with a focus on hematology and oncology. CONCLUSION: Thus, this article provides a point of reference for hematologists and oncologists, and at the same time sets forth a framework for the further development and clinical deployment of AI in hematology and oncology in the future.

Volumetric imaging reveals VEGF-C-dependent formation of hepatic lymph vessels in mice.

The liver is a major biosynthetic and detoxifying organ in vertebrates, but also generates 25%-50% of the lymph passing through the thoracic duct and is thereby the organ with the highest contribution to lymph flow. In contrast to its metabolic function, the role of the liver for lymph generation and composition is presently severely understudied. We took a rigorous, volume imaging-based approach to describe the microarchitecture and spatial composition of the hepatic lymphatic vasculature with cellular resolution in whole mount immune stained specimen ranging from thick sections up to entire mouse liver lobes. Here, we describe that in healthy adult livers, lymphatic vessels were exclusively located within the portal tracts, where they formed a unique, highly ramified tree. Ragged, spiky initials enmeshed the portal veins along their entire length and communicated with long lymphatic vessels that followed the path of the portal vein in close association with bile ducts. Together these lymphatic vessels formed a uniquely shaped vascular bed with a delicate architecture highly adapted to the histological structure of the liver. Unexpectedly, with the exception of short collector stretches at the porta hepatis, which we identified as exit point of the liver lymph vessels, the entire hepatic lymph vessel system was comprised of capillary lymphatic endothelial cells only. Functional experiments confirmed the space of Disse as the origin of the hepatic lymph and flow via the space of Mall to the portal lymph capillaries. After entry into the lymphatic initials, the lymph drained retrograde to the portal blood flow towards the exit at the liver hilum. Perinatally, the liver undergoes complex changes transforming from the main hematopoietic to the largest metabolic organ. We investigated the time course of lymphatic vessel development and identified the hepatic lymphatics to emerge postnatally in a process that relies on input from the VEGF-C/VERGFR-3 growth factor-receptor pair for formation of the fully articulate hepatic lymph vessel bed.

Perspectives in machine learning for wildlife conservation.

Inexpensive and accessible sensors are accelerating data acquisition in animal ecology. These technologies hold great potential for large-scale ecological understanding, but are limited by current processing approaches which inefficiently distill data into relevant information. We argue that animal ecologists can capitalize on large datasets generated by modern sensors by combining machine learning approaches with domain knowledge. Incorporating machine learning into ecological workflows could improve inputs for ecological models and lead to integrated hybrid modeling tools. This approach will require close interdisciplinary collaboration to ensure the quality of novel approaches and train a new generation of data scientists in ecology and conservation.

An uncertainty-aware, shareable, and transparent neural network architecture for brain-age modeling.

The deviation between chronological age and age predicted from neuroimaging data has been identified as a sensitive risk marker of cross-disorder brain changes, growing into a cornerstone of biological age research. However, machine learning models underlying the field do not consider uncertainty, thereby confounding results with training data density and variability. Also, existing models are commonly based on homogeneous training sets, often not independently validated, and cannot be shared because of data protection issues. Here, we introduce an uncertainty-aware, shareable, and transparent Monte Carlo dropout composite quantile regression (MCCQR) Neural Network trained on N = 10,691 datasets from the German National Cohort. The MCCQR model provides robust, distribution-free uncertainty quantification in high-dimensional neuroimaging data, achieving lower error rates compared with existing models. In two examples, we demonstrate that it prevents spurious associations and increases power to detect deviant brain aging. We make the pretrained model and code publicly available.

PHOTONAI-A Python API for rapid machine learning model development.

PHOTONAI is a high-level Python API designed to simplify and accelerate machine learning model development. It functions as a unifying framework allowing the user to easily access and combine algorithms from different toolboxes into custom algorithm sequences. It is especially designed to support the iterative model development process and automates the repetitive training, hyperparameter optimization and evaluation tasks. Importantly, the workflow ensures unbiased performance estimates while still allowing the user to fully customize the machine learning analysis. PHOTONAI extends existing solutions with a novel pipeline implementation supporting more complex data streams, feature combinations, and algorithm selection. Metrics and results can be conveniently visualized using the PHOTONAI Explorer and predictive models are shareable in a standardized format for further external validation or application. A growing add-on ecosystem allows researchers to offer data modality specific algorithms to the community and enhance machine learning in the areas of the life sciences. Its practical utility is demonstrated on an exemplary medical machine learning problem, achieving a state-of-the-art solution in few lines of code. Source code is publicly available on Github, while examples and documentation can be found at www.photon-ai.com.

Resolving Colliding Larvae by Fitting ASM to Random Walker-Based Pre-Segmentations.

Drosophila melanogaster is an important model organism for research in neuro- and behavioral biology. Automated studies of their locomotion are crucial to link sensory input and neural processing to motor output which has led to numerous vision-based tracking systems. However, most of these approaches share the inability to segment the contours of colliding animals causing identity losses, appearing and disappearing animals, and the absence of posture and motion related measurements during the time of the collision. We present a novel collision resolution algorithm enabling an accurate contour segmentation of multiple touching Drosophila larvae. Our algorithm utilizes an adapted active shape model (ASM) to learn a low dimensional posture space which is fitted to random-walker generated pre-segmentations. We evaluate our collision resolution algorithm using three publicly available datasets and compare it with the current state-of-the-art methods. In addition, we introduce a refined dataset enabling a segmentation evaluation on the level of pixel accuracy. The results demonstrate that our approach outperforms the state-of-the-art approaches in both accuracy and computational time. We will incorporate this algorithm into our widely used tracking program to improve the statistical strength of the behavioral quantification and allow marker-free studies of interacting Drosophila larvae.

The Drosophila NCAM homolog Fas2 signals independently of adhesion.

The development of tissues and organs requires close interaction of cells. To achieve this, cells express adhesion proteins such as the neural cell adhesion molecule (NCAM) or its Drosophila ortholog Fasciclin 2 (Fas2). Both are members of the Ig-domain superfamily of proteins that mediate homophilic adhesion. These proteins are expressed as isoforms differing in their membrane anchorage and their cytoplasmic domains. To study the function of single isoforms, we have conducted a comprehensive genetic analysis of Fas2 We reveal the expression pattern of all major Fas2 isoforms, two of which are GPI anchored. The remaining five isoforms carry transmembrane domains with variable cytoplasmic tails. We generated Fas2 mutants expressing only single isoforms. In contrast to the null mutation, which causes embryonic lethality, these mutants are viable, indicating redundancy among the different isoforms. Cell type-specific rescue experiments showed that glial-secreted Fas2 can rescue the Fas2 mutant phenotype to viability. This demonstrates that cytoplasmic Fas2 domains have no apparent essential functions and indicate that Fas2 has function(s) other than homophilic adhesion. In conclusion, our data suggest novel mechanistic aspects of a long-studied adhesion protein.

Towards image-based animal tracking in natural environments using a freely moving camera.

BACKGROUND: Image-based tracking of individual animals can provide rich data to underpin breakthroughs in biological and medical research, but few if any existing methods extend to tracking unconstrained natural behaviour in the field. NEW METHOD: We have developed a visual tracking system for animals filmed with a freely moving hand-held or drone-operated camera in their natural environment. This exploits a global inference method for detecting motion of an animal against a cluttered background. Trajectories are then generated by a novel video key-frame selection scheme in combination with a geometrically constrained image stitching algorithm, resulting in a two-dimensional panorama image of the environment on which the dense animal path is displayed. RESULTS: By introducing a minimal and plausible set of constraints regarding the camera orientation and movement, we demonstrate that both per-frame animal positions and overall trajectories can be extracted with reasonable accuracy, for a range of different animals, environments and imaging modalities. COMPARISON: Our method requires only a single uncalibrated camera, does not require marking or training data to detect the animal, and makes no prior assumptions about appearance of the target or background. In particular it can detect targets occupying fewer than 20 pixels in the image, and deal with poor contrast, highly dynamic lighting and frequent occlusion. CONCLUSION: Our algorithm produces highly informative qualitative trajectories embedded in a panorama of the environment. The results are still subject to rotational drift and additional scaling routines would be needed to obtain absolute real-world coordinates. It nevertheless provides a flexible and easy-to-use system to obtain rich data on natural animal behaviour in the field.

From skylight input to behavioural output: A computational model of the insect polarised light compass.

Many insects navigate by integrating the distances and directions travelled on an outward path, allowing direct return to the starting point. Fundamental to the reliability of this process is the use of a neural compass based on external celestial cues. Here we examine how such compass information could be reliably computed by the insect brain, given realistic constraints on the sky polarisation pattern and the insect eye sensor array. By processing the degree of polarisation in different directions for different parts of the sky, our model can directly estimate the solar azimuth and also infer the confidence of the estimate. We introduce a method to correct for tilting of the sensor array, as might be caused by travel over uneven terrain. We also show that the confidence can be used to approximate the change in sun position over time, allowing the compass to remain fixed with respect to 'true north' during long excursions. We demonstrate that the compass is robust to disturbances and can be effectively used as input to an existing neural model of insect path integration. We discuss the plausibility of our model to be mapped to known neural circuits, and to be implemented for robot navigation.

The sulfite oxidase Shopper controls neuronal activity by regulating glutamate homeostasis in Drosophila ensheathing glia.

Specialized glial subtypes provide support to developing and functioning neural networks. Astrocytes modulate information processing by neurotransmitter recycling and release of neuromodulatory substances, whereas ensheathing glial cells have not been associated with neuromodulatory functions yet. To decipher a possible role of ensheathing glia in neuronal information processing, we screened for glial genes required in the Drosophila central nervous system for normal locomotor behavior. Shopper encodes a mitochondrial sulfite oxidase that is specifically required in ensheathing glia to regulate head bending and peristalsis. shopper mutants show elevated sulfite levels affecting the glutamate homeostasis which then act on neuronal network function. Interestingly, human patients lacking the Shopper homolog SUOX develop neurological symptoms, including seizures. Given an enhanced expression of SUOX by oligodendrocytes, our findings might indicate that in both invertebrates and vertebrates more than one glial cell type may be involved in modulating neuronal activity.

Automatic non-invasive heartbeat quantification of Drosophila pupae.

The importance of studying model organisms such as Drosophila melanogaster has significantly increased in recent biological research. Amongst others, Drosophila can be used to study heart development and heartbeat related diseases. Here we propose a method for automatic in vivo heartbeat detection of Drosophila melanogaster pupae based on morphological structures which are recorded without any dissection using FIM imaging. Our approach is easy-to-use, has low computational costs, and enables high-throughput experiments. After automatically segmenting the heart region of the pupa in an image sequence, the heartbeat is indirectly determined based on intensity variation analysis. We have evaluated our method using 47,631 manually annotated frames from 29 image sequences recorded with different temporal and spatial resolutions which are made publicly available. We show that our algorithm is both precise since it detects more than 95% of the heartbeats correctly as well as robust since the same standardized set of parameters can be used for all sequences. The combination of FIM imaging and our algorithm enables a reliable heartbeat detection of multiple Drosophila pupae while simultaneously avoiding any time consuming preparation of the animals.