Refine
Year of publication
Document Type
- Article (25)
- Part of a Book (4)
- Preprint (3)
- Conference Proceeding (2)
- Book (1)
Has Fulltext
- no (35) (remove)
Keywords
Sleep scoring is a necessary and time-consuming task in sleep studies. In animal models (such as mice) or in humans, automating this tedious process promises to facilitate long-term studies and to promote sleep biology as a data-driven f ield. We introduce a deep neural network model that is able to predict different states of consciousness (Wake, Non-REM, REM) in mice from EEG and EMG recordings with excellent scoring results for out-of-sample data. Predictions are made on epochs of 4 seconds length, and epochs are classified as artifactfree or not. The model architecture draws on recent advances in deep learning and in convolutional neural networks research. In contrast to previous approaches towards automated sleep scoring, our model does not rely on manually defined features of the data but learns predictive features automatically. We expect deep learning models like ours to become widely applied in different fields, automating many repetitive cognitive tasks that were previously difficult to tackle.
Reliable methods for automatic readability assessment have the potential to impact a variety of fields, ranging from machine translation to self-informed learning. Recently, large language models for the German language (such as GBERT and GPT-2-Wechsel) have become available, allowing to develop Deep Learning based approaches that promise to further improve automatic readability assessment. In this contribution, we studied the ability of ensembles of fine-tuned GBERT and GPT-2-Wechsel models to reliably predict the readability of German sentences. We combined these models with linguistic features and investigated the dependence of prediction performance on ensemble size and composition. Mixed ensembles of GBERT and GPT-2-Wechsel performed better than ensembles of the same size consisting of only GBERT or GPT-2-Wechsel models. Our models were evaluated in the GermEval 2022 Shared Task on Text Complexity Assessment on data of German sentences. On out-of-sample data, our best ensemble achieved a root mean squared error of 0:435.
Data-driven prediction and prevention of extreme events in a spatially extended excitable system
(2015)
Detecting synchronization clusters in multivariate time series via coarse-graining of Markov chains
(2007)
Epilepsy
(2010)
We investigate the suitability of selected measures of complexity based on recurrence quantification analysis and recurrence networks for an identification of pre-seizure states in multi-day, multi-channel, invasive electroencephalographic recordings from five epilepsy patients. We employ several statistical techniques to avoid spurious findings due to various influencing factors and due to multiple comparisons and observe precursory structures in three patients. Our findings indicate a high congruence among measures in identifying seizure precursors and emphasize the current notion of seizure generation in large-scale epileptic networks. A final judgment of the suitability for field studies, however, requires evaluation on a larger database.
Network theory provides novel concepts that promise an improved characterization of interacting dynamical systems. Within this framework, evolving networks can be considered as being composed of nodes, representing systems, and of time-varying edges, representing interactions between these systems. This approach is highly attractive to further our understanding of the physiological and pathophysiological dynamics in human brain networks. Indeed, there is growing evidence that the epileptic process can be regarded as a large-scale network phenomenon. We here review methodologies for inferring networks from empirical time series and for a characterization of these evolving networks. We summarize recent findings derived from studies that investigate human epileptic brain networks evolving on timescales ranging from few seconds to weeks. We point to possible pitfalls and open issues, and discuss future perspectives.
We consider recent reports on small-world topologies of interaction networks derived from the dynamics of spatially extended systems that are investigated in diverse scientific fields such as neurosciences, geophysics, or meteorology. With numerical simulations that mimic typical experimental situations, we have identified an important constraint when characterizing such networks: indications of a small-world topology can be expected solely due to the spatial sampling of the system along with the commonly used time series analysis based approaches to network characterization.
The network approach towards the analysis of the dynamics of complex systems has been successfully applied in a multitude of studies in the neurosciences and has yielded fascinating insights. With this approach, a complex system is considered to be composed of different constituents which interact with each other. Interaction structures can be compactly represented in interaction networks. In this contribution, we present a brief overview about how interaction networks are derived from multivariate time series, about basic network characteristics, and about challenges associated with this analysis approach.