OPEN FAU Online-Publikationssystem der Friedrich-Alexander-Universität Erlangen-Nürnberg
Not a member yet
23478 research outputs found
Sort by
Telemedicine networks for acute stroke: An analysis of global coverage, gaps, and opportunities
No full textBackground: Despite the proven efficacy of telestroke in improving clinical outcomes by providing access to specialized expertise and allowing rapid expert hyperacute stroke management and decision-making, detailed operational evidence is scarce, especially for less developed or lower income regions. Aim: We aimed to map the global telestroke landscape and characterize existing networks. Methods: We employed a four-tiered approach to comprehensively identify telestroke networks, primarily involving engagement with national stroke experts, stroke societies, and international stroke authorities. A carefully designed questionnaire was then distributed to the leaders of all identified networks to assess these networks’ structures, processes, and outcomes. Results: We identified 254 telestroke networks distributed across 67 countries. High-income countries (HICs) concentrated 175 (69%) of the networks. No evidence of telestroke services was found in 58 (30%) countries. From the identified networks, 88 (34%) completed the survey, being 61 (71%) located in HICs. Network setup was highly heterogeneous, ranging from 17 (22%) networks with more than 20 affiliated hospitals, providing thousands of annual consultations using purpose-built highly specialized technology, to 11 (13%) networks with fewer than 120 consultations annually using generic videoconferencing equipment. Real-time video and image transfer was employed in 64 (75%) networks, while 62 (74%) conducting quality monitoring. Most networks established in the past 3 years were located in low- and middle-income countries (LMICs). Conclusion: This comprehensive global survey of telestroke networks found significant variation in network coverage, setup, and technology use. Most services are in HICs, and a few services are in LMICs, although an emerging trend of new networks in these regions marks a pivotal moment in global telestroke care. The wide variation in quality monitoring practices across networks, with many failing to report key performance metrics, underscores the urgent need for standardized, resource-appropriate, quality assurance measures that can be adapted to diverse settings.World Stroke Organizatio
Measuring Low Viscosities and High Shear Rates with a Rotational Rheometer in a Thin-gap Parallel-disk Configuration
Get PDFWe modify a commercial rheometer so that the disks are aligned perpendicular to the axis of rotation with a precision in parallelism of about 1 μm independent of the rheometer reading. This leads to decrease the zero-gap error by a factor of 25 and more. It enables samples to be studied at gap widths well below the absolute error of commercial rheometers. At gap widths of 20 μm, the modification allows the measurement range to be extended to shear rates up to 105 s-1 enabling to measure low viscosities such as that of solvents or water and of dilute polymer solutions. The measurements are restricted mainly by the torque resolution at low shear rates and by inertia at high shear rates
A Multisensory Deficit in the Perception of Pleasantness in Parkinson’s Disease
No full textBackground: There is growing interest in non-motor symptoms in Parkinson’s disease (PD), due to the impact on quality of life. Anhedonia, the inability to experience joy and lust, has a prevalence of up to 46% in PD. The perception of pleasantness of an odor is reduced in anhedonia without PD. We previously showed a reduced hedonic olfactory perception in PD, i.e., patients evaluated odors as less pleasant or unpleasant compared to controls. This deficit correlated with anhedonia. Objective: We aimed to confirm these findings. Moreover, we hypothesized that the perception of pleasantness in PD is affected on a multisensory level and correlates with anhedonia. Therefore, we assessed olfactory, visual and acoustic evaluation of pleasantness in PD and healthy individuals. Methods: Participants had to rate the pleasantness of 22 odors, pictures, and sounds on a nine-point Likert scale. Depression, anhedonia, and apathy were assessed by means of questionnaires. Results of the pleasantness-rating were compared between groups and correlated to scores of the questionnaires. Results: In particular pleasant and unpleasant stimuli across all three modalities are perceived less intense in PD, suggesting that a reduced range of perception of pleasantness is a multisensory phenomenon. However, only a reduction of visual hedonic perception correlated with anhedonia in PD. A correlation of reduced perception of pleasantness with apathy or depression was not present. Conclusion: We provide evidence for a multisensory deficit in the perception of pleasantness. Further studies should delineate the underlying neural circuity and the diagnostic value to detect neuropsychiatric symptoms in PD
Dateneffizientes Deep Learning Vision Anwendungen für die Segmentierung von Mikroskopiebildern und die Klassifizierung natürlicher Bilder
No full textDeep neural networks have made a considerable breakthrough in automating many
computer vision tasks. However, their effectiveness relies heavily on having a large
amount of annotated data, which can be expensive and impractical due to the associated
training and annotation costs. Prior work has proposed learning approaches, such as
few-shot and semi-supervised learning, to relax the annotation burden and facilitate
data-efficient learning. In line with this goal, this dissertation addresses, for the first
time, learning with limited annotated data in the context of few-shot microscopy image
cell segmentation to relax the laborious efforts and enable ubiquitous automation with
deep neural networks. Moreover, selection approaches are studied to optimize the
annotation efforts to determine label-worthy data samples from the task of interest in
microscopy image cell segmentation and natural image recognition to improve few-shot
fine-tuning performance.
This is a cumulative dissertation founded on four publications. The first two
publications [1, 2] focus on learning with limited annotated data paradigms with the
goal of adapting the trained models using few annotated samples from the dataset
of interest. Notably, the first publication showcases the first-ever study of few-shot
microscopy image cell segmentation, where a generic segmentation model is trained using
a collection of different microscopy datasets. The second publication presents a semi-
supervised few-shot microscopy image cell segmentation, which trains a segmentation
model using a partially annotated training dataset. In both studies, the trained models
are adapted to the microscopy dataset of interest using a few annotated samples. The
third and fourth publications [3, 4] propose bootstrapping the few-shot fine-tuning
performance of microscopy image cell segmentation and natural image recognition by
optimizing the selection of the few annotated samples from the dataset of interest.
In summary, this dissertation presents a foundational study of microscopy image
cell segmentation under the paradigms of few-shot and semi-supervised learning.
Furthermore, it emphasizes automating support set selection to enhance few-shot
fine-tuning performance for microscopy image cell segmentation and natural image
recognition.Tiefe neuronale Netze haben einen beträchtlichen Durchbruch bei der Automatisierung vieler Bildverarbeitungsaufgaben erzielt. Ihre Effektivität hängt jedoch stark von einer großen Menge an kommentierten Daten ab, was aufgrund der damit verbundenen Schulungs- und Kommentierungskosten teuer und unpraktisch sein kann. In früheren Arbeiten wurden Lernansätze, wie zum beispiel few-shot und semi-supervised learning, vorgeschlagen, um den Annotationsaufwand zu verringern und dateneffizientes Lernen zu ermöglichen. Im Einklang mit diesem Ziel befasst sich diese Arbeit zum ersten Mal mit dem Lernen mit begrenzten annotierten Daten im Zusammenhang mit der Zellsegmentierung von Mikroskopiebildern mit wenigen Aufnahmen, um die mühsamen Anstrengungen zu verringern und eine allgegenwärtige Automatisierung mit tiefen neuronalen Netzen zu ermöglichen. Darüber hinaus werden Auswahlansätze untersucht, um die Annotation Bemühungen zu optimieren, um Label-würdigen Datenproben aus der Aufgabe von Interesse in der Mikroskopie Bild Zellsegmentierung und natürliche Bilderkennung zu bestimmen, um die wenigen Schuss Feinabstimmung Leistung zu verbessern.
Dieses Manuskript ist eine kumulative Dissertation, die auf vier Veröffentlichungen beruht. Die ersten beiden Publikationen konzentrieren sich auf das Lernen mit begrenzten annotierten Daten mit dem Ziel, die trainierten Modelle mit wenigen annotierten Proben aus dem Datensatz von Interesse anzupassen. In der ersten Publikation wird die allererste Studie zur Zellsegmentierung von Mikroskopiebildern mit wenigen Aufnahmen vorgestellt, bei der ein generisches Segmentierungsmodell mit einer Sammlung verschiedener Mikroskopiedatensätze trainiert und dann mit wenigen beschrifteten Proben an den Mikroskopiedatensatz von Interesse angepasst wird. In der zweiten Veröffentlichung wird eine semi-supervised few-shot microscopy image cell segmentation vorgestellt, bei der ein Segmentierungsmodell anhand eines teilweise annotierten Trainingsdatensatzes trainiert und dann anhand einiger annotierter Proben an den interessierenden Mikroskopiedatensatz angepasst wird. In der dritten und vierten Veröffentlichung wird vorgeschlagen, die Feinabstimmungsleistung der Mikroskopie Bildzellsegmentierung und der natürlichen Bilderkennung durch Optimierung der Auswahl der wenigen kommentierten Proben aus dem interessierenden Datensatz zu testen.
Zusammenfassend stellt diese Arbeit eine grundlegende Studie der Mikroskopie Bild Zellsegmentierung unter den Paradigmen der wenigen Schuss und semi-supervised Lernen. Darüber hinaus betont sie die Automatisierung der Support-Set-Auswahl zur Verbesserung der few-shot Feinabstimmung Leistung für Mikroskopie Bild Zellsegmentierung und natürliche Bilderkennung
TFAP2E is implicated in central nervous system, orofacial and maxillofacial anomalies
Get PDFBackground Previous studies in mouse, Xenopus and zebrafish embryos show strong tfap2e expression in progenitor cells of neuronal and neural crest tissues suggesting its involvement in neural crest specification. However, the role of human transcription factor activator protein 2 ( TFAP2E) in human embryonic central nervous system (CNS), orofacial and maxillofacial development is unknown. Methods Through a collaborative work, exome survey was performed in families with congenital CNS, orofacial and maxillofacial anomalies. Exome variant prioritisation prompted TFAP2E gene for functional analysis in zebrafish embryos. Embryonic morphology and development were assessed after antisense morpholino (MO) knockdown (KD), CRISPR/Cas9 knockout and overexpression of tfap2e in fluorescent zebrafish reporter lines using in vivo microscopy. Computational structural protein modelling of the identified human variants was performed. Results In total, exome survey identified novel or ultra-rare heterozygous missense variants in TFAP2E in seven individuals from five independent families with predominantly CNS, orofacial and maxillofacial anomalies. One variant was found de novo and another variant segregated in an affected multiplex family. Protein modelling of the identified variants indicated potential distortion of TFAP2E in the transactivation or dimerisation domain. MO KD and CRISPR/Cas9 knockout of tfap2e in zebrafish revealed hydrocephalus and a significant reduction of brain volume, consistent with a microencephaly phenotype. Furthermore, mRNA overexpression of TFAP2E indicates dosage-sensitive phenotype expression. In addition, zebrafish showed orofacial and maxillofacial anomalies following tfap2e KD, recapitulating the human phenotype. Conclusion Our human genetic data and analysis of Tfap2e manipulation in zebrafish indicate a potential role of TFAP2E in human CNS, orofacial and maxillofacial anomalies.Herbert-Reeck foundationNational Human Genome Research Institute (NHGRI)Deutsche Forschungsgemeinschaft, DFGElse Kröner-Fresenius Foundation Graduate SchoolBONFORthe National Institutes of Health, National Institute of Neurologic Disease and Strokehttp://dx.doi.org/10.13039/501100001659Deutsche ForschungsgemeinschaftNational Heart, Lung, and Blood Institute (NHLBI)http://dx.doi.org/10.13039/501100007952Eva Luise und Horst Köhler Stiftunghttp://dx.doi.org/10.13039/100005202Muscular Dystrophy Associatio
Newbornscreening SMA – From Pilot Project to Nationwide Screening in Germany
No full textNow that targeted therapies for spinal muscular atrophy are available, attempts are being made worldwide to include screening for spinal muscular atrophy in general newborn screening. In Germany, after pilot projects from 2018–2021, it was included in the general newborn screening from October 2021. To ensure a smooth transition, criteria for follow-up were developed together with key stakeholders. At the beginning of the transition to nationwide screening, false positive findings were reported in 3 patients. After optimization of the screening method in the laboratories concerned, all findings have been subsequently confirmed. On average, the first presentation to a neuromuscular center occurred on day 12 of life, and in patients with 2 or 3 SMN2 copies, therapy started on day 26 of life. Compared with the pilot project, there was no significant delay in timing
Establishment Survival in East and West Germany: A Comparative Analysis
Get PDFUsing a large administrative dataset, this paper compares the development of new establishments" survival chances in East and West Germany for the period 1994 – 2008. A central question is whether convergence with respect to survival rates between East and West Germany can be observed. Using methods of survival analysis, I find that new establishments" survival chances do not differ strongly between East and West Germany at the beginning of the observation period. In 1998 and 1999 the exit hazard increases strongly in East but not in West Germany, which is likely to be due to a change in the subsidy policy affecting East Germany. Since the turn of the millennium, the difference in establishments" exit hazard between East and West Germany becomes smaller, indicating that there is convergence with respect to establishments" survival chances.Anhand umfangreicher administrativer Daten vergleicht diese Studie die Entwicklung der Überlebenschancen neu gegründeter Betriebe in West- und Ostdeutschland für die Jahre 1994 bis 2008. Eine zentrale Frage lautet dabei, ob eine Angleichung der Überlebensraten zwischen West- und Ostdeutschland zu beobachten ist. Anhand von Methoden der Verweildaueranalyse kommt die Studie zu dem Ergebnis, dass sich die Überlebenschancen neu gegründeter Betriebe zu Beginn des Beobachtungszeitraums nicht stark zwischen West- und Ostdeutschland unterscheiden. In den Jahren 1998 und 1999 steigt die Schließungswahrscheinlichkeit in Ostdeutschland stark an, in Westdeutschland jedoch nicht, was vermutlich auf eine Änderung der Subventionspolitik für Betriebe in Ostdeutschland zurückzuführen ist. Seit der Jahrtausendwende nimmt der Unterschied in den Schließungswahrscheinlichkeiten zwischen West- und Ostdeutschland ab, was auf eine Angleichung der Überlebenschancen hindeutet
Quantum-enhanced interferometer using Kerr squeezing
No full textOne of the prime applications of squeezed light is enhancing the sensitivity of an interferometer below the quantum shot-noise limit, but so far, no such experimental demonstration was reported when using the optical Kerr effect. In prior setups involving Kerr-squeezed light, the role of the interferometer was merely to characterize the noise pattern. The lack of such a demonstration was largely due to the cumbersome tilting of the squeezed ellipse in phase space. Here, we present the first experimental observation of phase-sensitivity enhancement in an interferometer using Kerr squeezing.Ministerio de Ciencia e InnovaciónFoundation for the Advancement of Theoretical Physics and MathematicsRussian Foundation for Basic ResearchMinistry of Science and Higher Education of the Russian Federatio
Acoustic positioning for deep sea neutrino telescopes with a system of piezo sensors integrated into glass spheres
No full textPosition calibration in the deep sea is typically done by means of acoustic multilateration using three or more acoustic emitters installed at known positions. Rather than using hydrophones as receivers that are exposed to the ambient pressure, the sound signals can be coupled to piezo ceramics glued to the inside of existing containers for electronics or measuring instruments of a deep sea infrastructure. The ANTARES neutrino telescope operated from 2006 until 2022 in the Mediterranean Sea at a depth exceeding 2000 m . It comprised nearly 900 glass spheres with 432 mm diameter and 15 mm thickness, equipped with photomultiplier tubes to detect Cherenkov light from tracks of charged elementary particles. In an experimental setup within ANTARES, piezo sensors have been glued to the inside of such – otherwise empty – glass spheres. These sensors recorded signals from acoustic emitters with frequencies from 46545 to 60235 Hz . Two waves propagating through the glass sphere are found as a result of the excitation by the waves in the water. These can be qualitatively associated with symmetric and asymmetric Lamb-like waves of zeroth order: a fast (early) one with ve≈5mm/μsand a slow (late) one with vℓ≈2mm/μs. Taking these findings into account improves the accuracy of the position calibration. The results can be transferred to the KM3NeT neutrino telescope, currently under construction at multiple sites in the Mediterranean Sea, for which the concept of piezo sensors glued to the inside of glass spheres has been adapted for monitoring the positions of the photomultiplier tubes.Open Access funding enabled and organized by Projekt DEAL.Friedrich-Alexander-Universität Erlangen-Nürnberg (1041
Predictors for long-term outcome of pulmonary valve perforation and balloon valvuloplasty in neonates with critical pulmonary valve stenosis or pulmonary valve atresia with intact ventricular septum
No full textIntroduction Percutaneous balloon valvuloplasty is the treatment of choice for critical pulmonary valve stenosis (CPS) and pulmonary valve atresia with intact ventricular septum (PA/IVS) if the ventricle has a suitable size. This study aimed to evaluate the long-term outcomes and predictors for surgical intervention and pulmonary valve replacement in patients with CPS or with PA/IVS after PBV, considering different morphological and hemodynamic parameters. Methods Neonates with PA/IVS or CPS who were admitted to the University Medicine Mainz and University Hospital Erlangen between November 1994 and March 2013 and underwent successful PBV as an initial procedure, with a follow-up of at least 5 years (median 13.1 years), were included. The Z -scores of pulmonary valve diameter, balloon/annulus ratio, number of cusps, and persisting stenosis were analyzed. The endpoint was the need for surgical procedures or valve replacement. Results A total of 62 neonates (median age at intervention 5 days) were included. Among them, 15 patients (24.2%) reached the endpoint. The mean time of freedom from surgery differed according to the number of cusps ( P 1.2). Conclusion Pulmonary valve perforation and PBV achieved favorable long-term outcomes in neonates with PA/IVS and CPS. A small pulmonary valve diameter, reduced number of cusps, and persisting gradient of >40 mmHg increased the risk for reduced time of freedom from surgical intervention and/or pulmonary valve replacement