Biology subjects

Kist, A. M.

Publications and source records attributed to Kist, A. M..

9 recordsLinked to original sources

bioRxiv2026

POME: Graph-based embeddings for partially observed mixed-type data

Partially observed mixed-type (POM) data, as often encountered in clinical, epidemiological, and phenotypic datasets, are very common in biomedical research. Yet, advanced data analysis and machine learning based on POM data is complicated by their heterogeneous nature and often substantial fractions of missing values. While one promising way to overcome these issues is to compute vector-valued embeddings for POM datasets which can then be used for downstream analyses, existing embedding methods are mostly not designed for POM data. To address this gap, we developed POME (partially observed mixed-type data embeddings), a self-supervised model that yields low-dimensional representations of both samples and variables, using shared concept learning and a bipartite graph representation of the underlying POM data. We validated POME through extensive experiments on three real-world biomedical datasets, with diverse downstream tasks and objectives: POME achieves state-of-the-art imputation performance and produces high-quality patient representations that support not only unsupervised discovery of well-separated and clinically meaningful patients subgroups but also supervised predictive modeling and zero-shot representation space mining for use cases such as adjuvant therapy modality recommendation. POME is available as a Python package on GitHub (https://github.com/bionetslab/POME) and PyPI (https://pypi.org/project/pome-py).

bioinformatics↗

bioRxiv2024

Predicting semantic segmentation quality in laryngeal endoscopy images

Endoscopy is a major tool for assessing the physiology of inner organs. Contemporary artificial intelligence methods are used to fully automatically label medical important classes on a pixel-by-pixel level. This so-called semantic segmentation is for example used to detect cancer tissue or to assess laryngeal physiology. However, due to the diversity of patients presenting, it is necessary to judge the segmentation quality. In this study, we present a fully automatic system to evaluate the segmentation performance in laryngeal endoscopy images. We showcase on glottal area segmentation that the predicted segmentation quality represented by the intersection over union metric is on par with human raters. Using a traffic light system, we are able to identify problematic segmentation frames to allow human-in-the-loop improvements, important for the clinical adaptation of automatic analysis procedures.

bioengineering↗

bioRxiv2023

DeepD3, an Open Framework for Automated Quantification of Dendritic Spines

Dendritic spines are the seat of most excitatory synapses in the brain, and a cellular structure considered central to learning, memory, and activity-dependent plasticity. The quantification of dendritic spines from light microscopy data is usually performed by humans in a painstaking and error-prone process. We found that human-to-human variability is substantial (inter-rater reliability 82.2{+/-}6.4%), raising concerns about the reproducibility of experiments and the validity of using human-annotated ground truth as an evaluation method for computational approaches of spine identification. To address this, we present DeepD3, an open deep learning-based framework to robustly quantify dendritic spines in microscopy data in a fully automated fashion. DeepD3s neural networks have been trained on data from different sources and experimental conditions, annotated and segmented by multiple experts and they offer precise quantification of dendrites and dendritic spines. Importantly, these networks were validated in a number of datasets on varying acquisition modalities, species, anatomical locations and fluorescent indicators. The entire DeepD3 open framework, including the fully segmented training data, a benchmark that multiple experts have annotated, and the DeepD3 model zoo is fully available, addressing the lack of openly available datasets of dendritic spines while offering a ready-to-use, flexible, transparent, and reproducible spine quantification method.

neuroscience↗

bioRxiv2021

Efficient and Explainable Deep Neural Networks for Airway Symptom Detection in Support of Wearable Health Technology

Mobile health wearables are often embedded with small processors for signal acquisition and analysis. These embedded wearable systems are, however, limited with low available memory and computational power. Advances in machine learning, especially deep neural networks (DNNs), have been adopted for efficient and intelligent applications to overcome constrained computational environments. In this study, evolutionary optimized DNNs were analyzed to classify three common airway-related symptoms, namely coughs, throat clears and dry swallows. As opposed to typical microphone-acoustic signals, mechanoacoustic data signals, which did not contain identifiable speech information for better privacy protection, were acquired from laboratory-generated and publicly available datasets. The optimized DNNs had a low footprint of less than 150 kB and predicted airway symptoms of interests with 83.7% accuracy on unseen data. By performing explainable AI techniques, namely occlusion experiments and class activation maps, mel-frequency bands up to 8,000 Hz were found as the most important feature for the classification. We further found that DNN decisions were consistently relying on these specific features, fostering trust and transparency of proposed DNNs. Our proposed efficient and explainable DNN is expected to support edge computing on mechano-acoustic sensing wearables for remote, longterm monitoring of airway symptoms.

bioengineering↗

bioRxiv2021

A single latent channel is sufficient for biomedical image segmentation

Glottis segmentation is a crucial step to quantify endoscopic footage in laryngeal high-speed videoendoscopy. Recent advances in using deep neural networks for glottis segmentation allow a fully automatic workflow. However, exact knowledge of integral parts of these segmentation deep neural networks remains unknown. Here, we show using systematic ablations that a single latent channel as bottleneck layer is sufficient for glottal area segmentation. We further show that the latent space is an abstraction of the glottal area segmentation relying on three spatially defined pixel subtypes. We provide evidence that the latent space is highly correlated with the glottal area waveform, can be encoded with four bits, and decoded using lean decoders while maintaining a high reconstruction accuracy. Our findings suggest that glottis segmentation is a task that can be highly optimized to gain very efficient and clinical applicable deep neural networks. In future, we believe that online deep learning-assisted monitoring is a game changer in laryngeal examinations.

bioengineering↗

bioRxiv2020

Rethinking glottal midline detection

A healthy voice is crucial for verbal communication and hence in daily as well as professional life. The basis for a healthy voice are the sound producing vocal folds in the larynx. A hallmark of healthy vocal fold oscillation is the symmetric motion of the left and right vocal fold. Clinically, videoendoscopy is applied to assess the symmetry of the oscillation and evaluated subjectively. High-speed videoendoscopy, an emerging method that allows quantification of the vocal fold oscillation, is more commonly employed in research due to the amount of data and the complex, semi-automatic analysis. In this study, we provide a comprehensive evaluation of methods that detect fully automatically the glottal midline. We use a biophysical model to simulate different vocal fold oscillations, extended the openly available BAGLS dataset using manual annotations, utilized both, simulations and annotated endoscopic images, to train deep neural networks at different stages of the analysis workflow, and compared these to established computer vision algorithms. We found that classical computer vision perform well on detecting the glottal midline in glottis segmentation data, but are outper-formed by deep neural networks on this task. We further suggest GlottisNet, a multi-task neural architecture featuring the simultaneous prediction of both, the opening between the vocal folds and the symmetry axis, leading to a huge step forward towards clinical applicability of quantitative, deep learning-assisted laryngeal endoscopy, by fully automating segmentation and midline detection.

bioengineering↗

bioRxiv2020

The cerebellum recalibrates a feedback controller involved in motor control

Animals must adapt their behavior to survive in a changing environment. Behavioral adaptations can be evoked by two mechanisms: feedback control and internal-model-based control. Feedback controllers can maintain the sensory state of the animal at a desired level under different environmental conditions. In turn, internal models learn the relationship between behavior and resulting sensory consequences in order to modify the behavior when this relationship changes. Here, we present multiple perturbations in visual feedback to larval zebrafish performing the optomotor response and show that they react to these perturbations through a feedback control mechanism. In contrast, if a perturbation is long-lasting, fish adapt their behavior by updating a cerebellum-dependent internal model. We use modelling and functional imaging to show that neuronal requirements for these mechanisms are met in the larval zebrafish brain. Our results illustrate the role of the cerebellum in encoding internal models and how these can calibrate neuronal circuits involved in reactive behaviors depending on the interactions between animal and environment. HighlightsO_LIBehavioral reactions to unexpected changes in visual feedback are implemented by a feedback control mechanism C_LIO_LIA long-lasting change in visual feedback updates the state of the neuronal controller C_LIO_LIThe cerebellar internal model mediates this recalibration C_LI

neuroscience↗

bioRxiv2018

Stytra: an open-source, integrated system for stimulation, tracking and closed-loop behavioral experiments

We present Stytra, a flexible, open-source software package, written in Python, that we designed to cover all the general requirements involved in larval zebrafish behavioral experiments. It allows timed stimulus presentation, interfacing with external devices and simultaneous real-time tracking of position, tail and eye motion in both freely-swimming and head-restrained preparations. It logs in standardized formats all recorded quantities, metadata, and code version to allow full provenance tracking, from data acquisition through analysis to publication. The package is modular and expandable for different experimental protocols and setups. Current releases can be found at https://github.com/portugueslab/stytra. We also provide complete documentation with examples for extending the package to new stimuli and hardware, as well as a schema and parts list for behavioral setups. We showcase Stytra by reproducing two previously behavioral protocols, one with head-restrained and the other with freely-swimming larvae, as well as an experiment in which the software is used in the context of a calcium imaging experiment, where it can trigger and communicate with other acquisition devices. Our aims are to help laboratories with little or no experience in the field to start implementing behavioral experiments and to provide a platform for sharing stimulus protocols to enable easy reproduction of experiments and straightforward validation. In addition, Stytra can easily serve as a base platform to design behavioral experiments in other model organisms.

neuroscience↗

bioRxiv2018

Motor context dominates output from Purkinje cell functional regions during reflexive visuomotor behaviors

The cerebellum integrates sensory stimuli and motor actions to enable smooth coordination and motor learning. Here we harness the innate behavioral repertoire of the larval zebrafish to characterize the spatiotemporal dynamics of feature coding across the entire Purkinje cell population during visual stimuli and the reflexive behaviors that they elicit. Population imaging reveals three spatially-clustered regions of Purkinje cell activity along the rostrocaudal axis. Complementary single-cell electrophysiological recordings assign these Purkinje cells to one of three functional phenotypes that encode a specific visual, and not motor, signal via complex spikes. In contrast, simple spike output of most Purkinje cells is strongly driven by motor-related tail and eye signals. Interactions between complex and simple spikes show heterogeneous modulation patterns across different Purkinje cells, which become temporally restricted during swimming episodes. Our findings reveal how sensorimotor information is encoded by individual Purkinje cells and organized into behavioral modules across the entire cerebellum.

neuroscience↗