Refine
Has Fulltext
- yes (3) (remove)
Is part of the Bibliography
- yes (3)
Year of publication
- 2023 (3) (remove)
Document Type
- Journal article (3)
Language
- English (3) (remove)
Keywords
- PAG (1)
- amygdala (1)
- cardiology (1)
- cerebellum (1)
- defensive states (1)
- fear (1)
- heart rate (1)
- machine learning (1)
- microscopy (1)
- neurology (1)
Institute
EU-Project number / Contract (GA) number
- 956414 (2)
Despite fine tuning voluntary movement as the most prominently studied function of the cerebellum, early human studies suggested cerebellar involvement emotion regulation. Since, the cerebellum has been associated with various mood and anxiety-related conditions. Research in animals provided evidence for cerebellar contributions to fear memory formation and extinction. Fear and anxiety can broadly be referred to as defensive states triggered by threat and characterized by multimodal adaptations such as behavioral and cardiac responses integrated into an intricately orchestrated defense reaction. This is mediated by an evolutionary conserved, highly interconnected network of defense-related structures with functional connections to the cerebellum. Projections from the deep cerebellar nucleus interpositus to the central amygdala interfere with retention of fear memory. Several studies uncovered tight functional connections between cerebellar deep nuclei and pyramis and the midbrain periaqueductal grey. Specifically, the fastigial nucleus sends direct projections to the ventrolateral PAG to mediate fear-evoked innate and learned freezing behavior. The cerebellum also regulates cardiovascular responses such as blood pressure and heart rate-effects dependent on connections with medullary cardiac regulatory structures. Because of the integrated, multimodal nature of defensive states, their adaptive regulation has to be highly dynamic to enable responding to a moving threatening stimulus. In this, predicting threat occurrence are crucial functions of calculating adequate responses. Based on its role in prediction error generation, its connectivity to limbic regions, and previous results on a role in fear learning, this review presents the cerebellum as a regulator of integrated cardio-behavioral defensive states.
Bioimages frequently exhibit low signal-to-noise ratios due to experimental conditions, specimen characteristics, and imaging trade-offs. Reliable segmentation of such ambiguous images is difficult and laborious. Here we introduce deepflash2, a deep learning-enabled segmentation tool for bioimage analysis. The tool addresses typical challenges that may arise during the training, evaluation, and application of deep learning models on ambiguous data. The tool’s training and evaluation pipeline uses multiple expert annotations and deep model ensembles to achieve accurate results. The application pipeline supports various use-cases for expert annotations and includes a quality assurance mechanism in the form of uncertainty measures. Benchmarked against other tools, deepflash2 offers both high predictive accuracy and efficient computational resource usage. The tool is built upon established deep learning libraries and enables sharing of trained model ensembles with the research community. deepflash2 aims to simplify the integration of deep learning into bioimage analysis projects while improving accuracy and reliability.