Henrik Bellhäuser

Diagnostic utility of different MRI and MR angiography measures in Fabry disease

Background: Neurologic hallmarks of Fabry disease (FD) include small fiber neuropathy as well as cerebral micro- and macroangiopathy with premature stroke. Cranial MRI shows progressive white matter lesions (WML) at an early age, increased signal intensity in the pulvinar, and tortuosity and dilatation of the larger vessels. To unravel the most promising imaging tool for the detection of CNS involvement in FD we compared the diagnostic utility of the different MR imaging findings. Methods: Twenty-five clinically affected patients with FD (age 36.5 ± 11.0) and 20 age-matched controls were investigated by structural MRI, MR angiography, and diffusion tensor imaging (DTI). Individual WML volumes, global mean diffusivity (MD), and mean cerebral artery diameters were determined. Results: Using receiver operating characteristic analyses, enlarged diameters of the following cerebral arteries significantly separated patients with FD from controls: middle cerebral artery: area under curve (AUC) = 0.75, p = 0.005; posterior cerebral artery: AUC = 0.69, p = 0.041; carotid artery: 0.69, p = 0.041; basilar artery: AUC = 0.96, p < 0.0005. A total of 87% of the individuals were correctly classified by basilar artery diameters (sensitivity 95%, specificity 83%). WML volumes and global MD values did not significantly separate patients from controls. Conclusions: With an accuracy of 87%, basilar artery diameters were superior to all other MR measures for separating patients with Fabry disease (FD) from controls. Future studies should adopt basilar artery measurements for early detection and monitoring of brain involvement in FD. Moreover, further investigations should reveal if the dilated vasculopathy in FD could be a screening marker to detect FD in a cohort of other cerebrovascular diseases, especially in cryptogenic stroke.

PeerLA - Assistant for Individual Learning Goals and Self-Regulation Competency Improvement in Online Learning Scenarios

While online learning is already a part of university education and didactics, not all students have the necessary self-regulation competency to really learn on their own efficiently and effectively. In classroom a teacher can take over a moderating part, set intermediate goals and give feedback to ones progress, but participants of online learning courses (e.g. in blended scenarios or Massive Open Online Courses (MOOCs)) face a higher demand of self-regulation competency. This paper presents a course and content independent assistant, PeerLA, which assists in improving self-regulation competency. PeerLA allows setting of long-term goals, breakdown into intermediate goals and keeps track of knowledge increase or time needed. A graphical feedback allows comparison of existing and aimed level of knowledge or time investments. PeerLA adds peer comparison to the visualization charts for social frame of reference. This comparison is course-wide or only with similar learners (close in goals and knowledge levels). PeerLA is implemented as a Learning Management System (LMS) plugin to support learning progress in mixed formal and informal learning scenarios. PeerLA was evaluated with 83 students in an online mathematics preparation course over four weeks. Results indicate the benefits of such a self-regulation assistance, especially for university freshmen.

MoodlePeers: Factors Relevant in Learning Group Formation for Improved Learning Outcomes, Satisfaction and Commitment in E-Learning Scenarios Using GroupAL

High-scale and pure online learning scenarios (like MOOCs) as well as blended-learning scenarios offer great possibilities to optimize the composition of learning groups working together on the assigned (or selected) tasks. While the benefits and importance of peer learning for deep learning and improvement of e.g. problem-solving competency and social skills are indisputable, little evidences exist about the relevant factors for group formation and their combination to optimize the learning outcome for all participants (in all groups). Based on the GroupAL algorithm, MoodlePeers proposes an plugin solution for Moodle. Evaluated in a four-week online university mathematics preparation course MoodlePeers proved significant differences in submission rate of homework, quality of homework, keeping up, and satisfaction with group work compared to randomly created groups. The significant factors from personality traits, motivation and team orientation are discussed as well as the algorithmic key functionality behind.

Who is the Perfect Match

Abstract Using digital tools for teaching allows to unburden teachers from organizational load and even provides qualitative improvements that are not achieved in traditional teaching. Algorithmically supported learning group formation aims at optimizing group composition so that each learner can achieve his or her maximum learning gain and learning groups stay stable and productive. Selecting and weighting relevant criteria for learning group formation is an interdisciplinary challenge. This contribution presents the status quo of algorithmic approaches and respective criteria for learning group formation. Based on this theoretical foundation, we describe an empirical study that investigated the influence of distributing two personality traits (conscientiousness and extraversion) either homogeneously or heterogeneously on subjective and objective measures of productivity, time investment, satisfaction, and performance. Results are compared to an earlier study that also included motivation and prior knowledge as criteria. We find both personality traits to enhance group satisfaction and performance when distributed heterogeneously.