Open Science
The reporting period was marked by a further consolidation and institutionalization of Open Science at the MPIB. A central milestone was the adoption of Institute-wide guidelines on Open Science and research data management (see https://os-rdm.mpib.berlin/guidelines/) by the Institute's Board of Directors in fall 2025. The guidelines were developed in a participatory process involving researchers from different disciplines, research support service units, external experts, and the MPG Administrative Headquarters (Blunk, 2025). They articulate the MPIB's strategic commitment to Open Science in line with (inter)national recommendations while explicitly accommodating the methodological and disciplinary diversity of the Institute's empirically oriented research. The guidelines define shared standards and workflows and provide orientation for Institute members. To ensure adaptability, annual revisions are planned, supported by "guideline ambassadors" embedded in the research centers.
The twice-awarded Open Science Innovation Award further highlights the MPIB's commitment to Open Science. In 2024, Stefan Appelhoff was honored for adding EEG as a modality to the Brain Imaging Data Structure (BIDS) and developing a dedicated Python software package MNE-BIDS, which simplifies the conversion of EEG data to BIDS format, and reads and analyzes entire BIDS datasets. In 2025, Aaron Peikert, Hannes Diemerling, Andreas M. Brandmaier, and Maximilian S. Ernst received the award for designing and implementing a workshop on "Reproducible Research in R," including openly reusable teaching materials, software, and publications.
Beyond these institutional measures, MPIB researchers made substantial contributions to Open Science. These include conceptual considerations regarding, e.g., preregistration as a means to reduce inferential uncertainty (Peikert et al., in press), and the extension of FAIR principles to scientific theories as versioned, reusable research objects (Van Lissa et al., 2026). In terms of Open Science practices, robustness analyses were conducted (e.g., Pit, 2026), and open data sources were compiled for reuse (in addition to data accompanying a published manuscript, e.g., Yang et al., 2025). Open software tools were developed that facilitate, e.g., automated testing of computational reproducibility embedded in R Markdown workflows (Brandmaier & Peikert, 2025), or the integration of large language models into online experiments via a chat interface (Bermudez Schettino et al., 2025). Open educational resources were published mostly as online tutorials. Examples include a practical overview of open-source large language models for behavioral science using the Hugging Face ecosystem (Hussain et al., 2024) or estimating statistical power for structural equation models in developmental cognitive science (Buchberger et al., 2024).
Open Access (OA) publishing remains a strong pillar of the Institute's Open Science profile. Of the 754 journal articles (including special issues/sections) published between 2023 and 2025 (as of June 30, 2026), about 91% were available in OA (distributed as follows: 41% in hybrid journals, 36% in genuine OA journals, 14% via green OA, plus 9% closed access). In addition, seven OA books were published during the reporting period, three of which were co-financed by the MPIB's OA publication fund, which also supported nine OA journal articles where no central MPG or third-party funding was available.
Since 2022, systematic output monitoring has been expanded beyond publications to include published research data and software code, now also documented on the MPIB website.
Finally, MPIB researchers actively contribute to Open Science networks through lectures, workshops, and engagement in MPG-internal and external initiatives, including the MPG Open Science Ambassadors network and the German Reproducibility Network.
Selected Publications