Article
Refine
Year of publication
Document Type
- Article (459) (remove)
Has Fulltext
- yes (459)
Is part of the Bibliography
- no (459)
Keywords
- Heavy Ion Experiments (20)
- Hadron-Hadron Scattering (14)
- Hadron-Hadron scattering (experiments) (10)
- LHC (9)
- Heavy-ion collision (7)
- Jets (5)
- Collective Flow (4)
- Heavy Quark Production (4)
- Petri net (4)
- Quark-Gluon Plasma (4)
Institute
- Informatik (459) (remove)
Co-design of a trustworthy AI system in healthcare: deep learning based skin lesion classifier
(2021)
This paper documents how an ethically aligned co-design methodology ensures trustworthiness in the early design phase of an artificial intelligence (AI) system component for healthcare. The system explains decisions made by deep learning networks analyzing images of skin lesions. The co-design of trustworthy AI developed here used a holistic approach rather than a static ethical checklist and required a multidisciplinary team of experts working with the AI designers and their managers. Ethical, legal, and technical issues potentially arising from the future use of the AI system were investigated. This paper is a first report on co-designing in the early design phase. Our results can also serve as guidance for other early-phase AI-similar tool developments.
Human lymph nodes play a central part of immune defense against infection agents and tumor cells. Lymphoid follicles are compartments of the lymph node which are spherical, mainly filled with B cells. B cells are cellular components of the adaptive immune systems. In the course of a specific immune response, lymphoid follicles pass different morphological differentiation stages. The morphology and the spatial distribution of lymphoid follicles can be sometimes associated to a particular causative agent and development stage of a disease. We report our new approach for the automatic detection of follicular regions in histological whole slide images of tissue sections immuno-stained with actin. The method is divided in two phases: (1) shock filter-based detection of transition points and (2) segmentation of follicular regions. Follicular regions in 10 whole slide images were manually annotated by visual inspection, and sample surveys were conducted by an expert pathologist. The results of our method were validated by comparing with the manual annotation. On average, we could achieve a Zijbendos similarity index of 0.71, with a standard deviation of 0.07.
Network graphs have become a popular tool to represent complex systems composed of many interacting subunits; especially in neuroscience, network graphs are increasingly used to represent and analyze functional interactions between multiple neural sources. Interactions are often reconstructed using pairwise bivariate analyses, overlooking the multivariate nature of interactions: it is neglected that investigating the effect of one source on a target necessitates to take all other sources as potential nuisance variables into account; also combinations of sources may act jointly on a given target. Bivariate analyses produce networks that may contain spurious interactions, which reduce the interpretability of the network and its graph metrics. A truly multivariate reconstruction, however, is computationally intractable because of the combinatorial explosion in the number of potential interactions. Thus, we have to resort to approximative methods to handle the intractability of multivariate interaction reconstruction, and thereby enable the use of networks in neuroscience. Here, we suggest such an approximative approach in the form of an algorithm that extends fast bivariate interaction reconstruction by identifying potentially spurious interactions post-hoc: the algorithm uses interaction delays reconstructed for directed bivariate interactions to tag potentially spurious edges on the basis of their timing signatures in the context of the surrounding network. Such tagged interactions may then be pruned, which produces a statistically conservative network approximation that is guaranteed to contain non-spurious interactions only. We describe the algorithm and present a reference implementation in MATLAB to test the algorithm’s performance on simulated networks as well as networks derived from magnetoencephalographic data. We discuss the algorithm in relation to other approximative multivariate methods and highlight suitable application scenarios. Our approach is a tractable and data-efficient way of reconstructing approximative networks of multivariate interactions. It is preferable if available data are limited or if fully multivariate approaches are computationally infeasible.
The amyloid precursor protein (APP) was discovered in the 1980s as the precursor protein of the amyloid A4 peptide. The amyloid A4 peptide, also known as A-beta (Aβ), is the main constituent of senile plaques implicated in Alzheimer’s disease (AD). In association with the amyloid deposits, increasing impairments in learning and memory as well as the degeneration of neurons especially in the hippocampus formation are hallmarks of the pathogenesis of AD. Within the last decades much effort has been expended into understanding the pathogenesis of AD. However, little is known about the physiological role of APP within the central nervous system (CNS). Allocating APP to the proteome of the highly dynamic presynaptic active zone (PAZ) identified APP as a novel player within this neuronal communication and signaling network. The analysis of the hippocampal PAZ proteome derived from APP-mutant mice demonstrates that APP is tightly embedded in the underlying protein network. Strikingly, APP deletion accounts for major dysregulation within the PAZ proteome network. Ca2+-homeostasis, neurotransmitter release and mitochondrial function are affected and resemble the outcome during the pathogenesis of AD. The observed changes in protein abundance that occur in the absence of APP as well as in AD suggest that APP is a structural and functional regulator within the hippocampal PAZ proteome. Within this review article, we intend to introduce APP as an important player within the hippocampal PAZ proteome and to outline the impact of APP deletion on individual PAZ proteome subcommunities.
Human functional brain connectivity can be temporally decomposed into states of high and low cofluctuation, defined as coactivation of brain regions over time. Rare states of particularly high cofluctuation have been shown to reflect fundamentals of intrinsic functional network architecture and to be highly subject-specific. However, it is unclear whether such network-defining states also contribute to individual variations in cognitive abilities – which strongly rely on the interactions among distributed brain regions. By introducing CMEP, a new eigenvector-based prediction framework, we show that as few as 16 temporally separated time frames (< 1.5% of 10min resting-state fMRI) can significantly predict individual differences in intelligence (N = 263, p < .001). Against previous expectations, individual’s network-defining time frames of particularly high cofluctuation do not predict intelligence. Multiple functional brain networks contribute to the prediction, and all results replicate in an independent sample (N = 831). Our results suggest that although fundamentals of person-specific functional connectomes can be derived from few time frames of highest connectivity, temporally distributed information is necessary to extract information about cognitive abilities. This information is not restricted to specific connectivity states, like network-defining high-cofluctuation states, but rather reflected across the entire length of the brain connectivity time series.
Event-related potentials (ERPs) are widely used in basic neuroscience and in clinical diagnostic procedures. In contrast, neurophysiological insights from ERPs have been limited, as several different mechanisms lead to ERPs. Apart from stereotypically repeated responses (additive evoked responses), these mechanisms are asymmetric amplitude modulations and phase-resetting of ongoing oscillatory activity. Therefore, a method is needed that differentiates between these mechanisms and moreover quantifies the stability of a response. We propose a constrained subspace independent component analysis that exploits the multivariate information present in the all-to-all relationship of recordings over trials. Our method identifies additive evoked activity and quantifies its stability over trials. We evaluate identification performance for biologically plausible simulation data and two neurophysiological test cases: Local field potential (LFP) recordings from a visuo-motor-integration task in the awake behaving macaque and magnetoencephalography (MEG) recordings of steady-state visual evoked fields (SSVEFs). In the LFPs we find additive evoked response contributions in visual areas V2/4 but not in primary motor cortex A4, although visually triggered ERPs were also observed in area A4. MEG-SSVEFs were mainly created by additive evoked response contributions. Our results demonstrate that the identification of additive evoked response contributions is possible both in invasive and in non-invasive electrophysiological recordings.
Heterologously expressed genes require adaptation to the host organism to ensure adequate levels of protein synthesis, which is typically approached by replacing codons by the target organism’s preferred codons. In view of frequently encountered suboptimal outcomes we introduce the codon-specific elongation model (COSEM) as an alternative concept. COSEM simulates ribosome dynamics during mRNA translation and informs about protein synthesis rates per mRNA in an organism- and context-dependent way. Protein synthesis rates from COSEM are integrated with further relevant covariates such as translation accuracy into a protein expression score that we use for codon optimization. The scoring algorithm further enables fine-tuning of protein expression including deoptimization and is implemented in the software OCTOPOS. The protein expression score produces competitive predictions on proteomic data from prokaryotic, eukaryotic, and human expression systems. In addition, we optimized and tested heterologous expression of manA and ova genes in Salmonella enterica serovar Typhimurium. Superiority over standard methodology was demonstrated by a threefold increase in protein yield compared to wildtype and commercially optimized sequences.
Correction to: Scientifc Reports https://doi.org/10.1038/s41598-019-43857-5, published online 17 May 2019. In the original version of this Article, Jan-Hendrik Trösemeier was incorrectly affiliated with ‘Division of Allergology, Paul Ehrlich Institut, Langen, Germany’. Te correct afliations are listed below...
Background: Although mortality after cardiac surgery has significantly decreased in the last decade, patients still experience clinically relevant postoperative complications. Among others, atrial fibrillation (AF) is a common consequence of cardiac surgery, which is associated with prolonged hospitalization and increased mortality.
Methods: We retrospectively analyzed data from patients who underwent coronary artery bypass grafting, valve surgery or a combination of both at the University Hospital Muenster between April 2014 and July 2015. We evaluated the incidence of new onset and intermittent/permanent AF (patients with pre- and postoperative AF). Furthermore, we investigated the impact of postoperative AF on clinical outcomes and evaluated potential risk factors.
Results: In total, 999 patients were included in the analysis. New onset AF occurred in 24.9% of the patients and the incidence of intermittent/permanent AF was 59.5%. Both types of postoperative AF were associated with prolonged ICU length of stay (median increase approx. 2 days) and duration of mechanical ventilation (median increase 1 h). Additionally, new onset AF patients had a higher rate of dialysis and hospital mortality and more positive fluid balance on the day of surgery and postoperative days 1 and 2. In a multiple logistic regression model, advanced age (odds ratio (OR) = 1.448 per decade increase, p < 0.0001), a combination of CABG and valve surgery (OR = 1.711, p = 0.047), higher C-reactive protein (OR = 1.06 per unit increase, p < 0.0001) and creatinine plasma concentration (OR = 1.287 per unit increase, p = 0.032) significantly predicted new onset AF. Higher Horowitz index values were associated with a reduced risk (OR = 0.996 per unit increase, p = 0.012). In a separate model, higher plasma creatinine concentration (OR = 2.125 per unit increase, p = 0.022) was a significant risk factor for intermittent/permanent AF whereas higher plasma phosphate concentration (OR = 0.522 per unit increase, p = 0.003) indicated reduced occurrence of this arrhythmia.
Conclusions: New onset and intermittent/permanent AF are associated with adverse clinical outcomes of elective cardiac surgery patients. Different risk factors implicated in postoperative AF suggest different mechanisms might be involved in its pathogenesis. Customized clinical management protocols seem to be warranted for a higher success rate of prevention and treatment of postoperative AF.
Die Erwartungen von Studieninteressierten weichen häufig beträchtlich von den tatsächlichen Studieninhalten und Anforderungen ab. Ein Grund dafür ist, dass viele sich nicht genügend Klarheit verschaffen, welche eigenen Stärken und Schwächen für den Erfolg in Studium und Beruf »tatsächlich« relevant sind. So könnte zum Beispiel ein Abiturient mit guten Noten in Mathematik und Physik und mäßigen Zensuren in Deutsch und Englisch noch schlussfolgern, dass ihm »das Naturwissenschaftliche mehr liegt«. Ob das naturwissenschaftliche Verständnis für ein erfolgreiches Studium der Informatik jedoch gut genug ausgeprägt ist, lässt sich nicht so leicht erschließen. Noch schwieriger ist es für Studieninteressierte einzuschätzen, wie ihre »Soft Skills« ausgeprägt sind – also die Persönlichkeitsmerkmale, die in der Schule nicht systematisch beurteilt werden, jedoch hochgradig aussagekräftig für langfristigen Erfolg in Studium und Beruf sind. Ein Wechsel des Studienfaches zu Beginn des Studiums führt häufig zu einer Verlängerung der Studiendauer. Auch wenn eine derartige »Orientierungsphase« oftmals als normal und wichtig eingeschätzt wird, zeigt die praktische Erfahrung, dass Studierende mit kurzer Studiendauer jenen, die länger studiert haben, bei der Stellenvergabe tendenziell vorgezogen werden. Eine längere Studiendauer wird von Arbeitgebern häufig als Zeichen mangelnder Zielstrebigkeit oder fehlender Berufsmotivation interpretiert und kann sich so Chancen mindernd für Berufseinsteiger auswirken. Ebenso ist es im Interesse der Universitäten, die Zahl der Studienfachwechsel und -abbrüche so gering wie möglich zu halten – nicht zuletzt aus wirtschaftlichen Gründen. Deshalb bietet die Universität Frankfurt Studieninteressierten – zunächst in den Fächern Informatik und Psychologie – mit dem Self-Assessment konkrete Entscheidungshilfen an. Der verfolgte Ansatz zielt darauf ab, Abiturientinnen und Abiturienten möglichst frühzeitig und mit vertretbarem Aufwand die Möglichkeit zu bieten, selbst zu überprüfen, inwieweit ihre Erwartungen an einen Studiengang mit den tatsächlichen Inhalten und Anforderungen übereinstimmen. Das Konzept zur Erstellung eines Self-Assessments, das hier beispielhaft für den Studiengang Informatik vorgestellt wird, entstand nicht umsonst in enger Kooperation mit dem Institut für Psychologie (Prof. Dr. Helfried Moosbrugger, Dr. Siegbert Reiß, Ewa Jonkisz). Denn neben der fachlichen Qualifikation entscheiden über den Studienerfolg auch persönliche Eigenschaften wie Leistungsbereitschaft und Hartnäckigkeit. Die Auswertung des anonym durchgeführten Self-Assessments deckt außerdem Wissenslücken bei den Studieninteressierten auf, so dass eine gezielte Vorbereitung auf das Studium möglich wird. Zum Beispiel bietet der Fachbereich Mathematik und Informatik gezielte Vorbereitungskurse für Studienanfänger an, und zwar in Programmierung und Mathematik. Auch werden in den Semesterferien Repetitorien und Vorbereitungskurse angeboten – alles aus Studienbeiträgen finanziert. Auf diese Weise kann es zu einem homogeneren Kenntnisstand speziell bei den Studierenden im ersten Semester kommen. Ziel ist es, dadurch auch den »Erstsemesterschock « zu mildern. Das Online-Beratungsangebot trägt damit zu einer direkten Verbesserung der Lern- und Lehrsituation bei.