Digital Scholarship and Open Science in Psychology and the Behavioral Sciences
( 19. Jul – 24. Jul, 2015 )
- Alexander Garcia Castro (Technical University of Madrid, ES)
- Janna Hastings (European Bioinformatics Institute - Cambridge, GB)
- Robert Stevens (University of Manchester, GB)
- Erich Weichselgartner (ZPID - Trier, DE)
- Dagmar Glaser (für administrative Fragen)
- Digital Scholarship and Open Science in Psychology and the Behavioral Sciences (Dagstuhl Perspectives Workshop 15302). Alexander Garcia Castro, Janna Hastings, Robert Stevens, and Erich Weichselgartner. In Dagstuhl Reports, Volume 5, Issue 7, pp. 42-68, Schloss Dagstuhl - Leibniz-Zentrum für Informatik (2016)
It is widely acknowledged that data and documents are of the most value when they are interconnected rather than isolated. Understanding mental health disorders requires correlating information from diverse sources — e.g. cross-referencing clinical, psychological, and genotypic sources. This interoperability layer has several components; many are related to data standards and ontologies. Data standards and ontologies are also central for Open Science and Digital Scholarship. Open Science, the movement to make scientific research, data and dissemination accessible to all members of an inquiring society - amateur or professional - is emerging under the influence of new technologies. Open Science has the potential to deliver disruptive technology that will impact digital scholarship across all disciplines. Interoperability is central to Open Science; data should be open and self-describing so that an intelligent, machine readable, interoperable layer may emerge. Thus, the Web becomes a platform supporting Open Science and Digital Scholarship. Critical issues in scholarly communication such as reproducibility, reporting structures for experimental results, and data annotation, can be solved only with widespread deployment and adoption of data standards and ontologies.
Information in the biomedical domain is highly interconnected so it is expected that Brain Mapping Data (BAM) will generate new insights by complementing existing datasets across biomedical science – e.g. genetic clues for Post-Traumatic Stress Disorder (PTSD) or molecular markers for cognitive disorders in relation to behavioral traits. Understanding and developing treatment breakthroughs of disorders such as schizophrenia, Alzheimer’s, suicide, PTSD and others will require a much more sophisticated infrastructure than currently exists. Genotypic, phenotypic, environmental, and psychological information should effortlessly come together in support of scientific inquiries. Although studies in psychology usually involve a large number of variables as well as an accurate description of the population, a shortage of data standards makes it difficult to share and exchange data. Experience from the biomedical domain clearly shows that an interdisciplinary and technologically innovative approach is needed. Psychology and the behavioral sciences pose an interesting challenge and opportunity to computer scientists as well as to those working on open science and digital scholarship. Both these domains are rapidly becoming data-driven research areas with limited experience in data and knowledge management. Our ability to make continued progress in understanding the mind and brain depends on finding new ways to organize and synthesize an ever-expanding body of information. The challenges discussed during this Dagstuhl Perspectives workshop will contribute to solidifying this interaction by opening new paths for collaboration.
Topics of the workshop
- Open Science
- Cyber infrastructures and e-science
- Data standards and ontologies in Psychology and Behavioral Sciences
- How data standards, cyber infrastructures, and open science relate to each other
- Replicability, reproducibility, and reporting structures
- Security and privacy implications of open data in psychology
- Experiences from other domains and applicability in psychology and behavioral sciences
Goals of the workshop
- To foster and initiate the discussion about open science and digital scholarship in psychology and the behavioral sciences, addressing specific issues such as data standards, interoperability, knowledge representation, ontologies, and linked data.
- To identify useful experiences from other domains, including requirements and issues to be addressed. More importantly, we seek to define a common vision, a road map for this community to build cyber infrastructures in support of open science and digital scholarship.
Researchers across many domains have invested significant resources to improve transparency, reproducibility, discoverability and, in general, the ability to share and empower the community. Digital Scholarship and Open Science are umbrella terms for the movement to make scientific research, its tools and data and dissemination accessible to all members of an inquiring society, amateur or professional. Digital infrastructures are an essential prerequisite for such open science and digital scholarship; the biomedical domain illustrates this culture. An impressive digital infrastructure has been built; this allows us to correlate information from genomes to diseases, and, by doing so, to support movements such as panomic studies and personalized medicine. A high degree of interdisciplinary work was necessary in building this infrastructure; the large quantities of data being produced, the high degree of interrelatedness, and, most of all, the need for this mingling of many types of data in a variety of forms forged this collaboration across the community and beyond.
The Behavioral Sciences, comprising psychology but also psychobiology, criminology, cognitive science and neuroscience, are also producing data at significant rates; by the same token, understanding mental health disorders requires correlating information from diverse sources - e.g. cross-referencing clinical, psychological, and genotypic sources. For example, flagship projects such as the Brain Activity Map (BAM, also known as the BRAIN initiative) are generating massive amounts of data with potential benefit to mental health and psychology; conversely, projects like BAM could benefit from information currently being generated by psychologists. Our ability to make continued progress in understanding the mind and brain depends on finding new ways to organize and synthesize an ever-expanding body of information.
The 'Digital Scholarship and Open Science in Psychology and the Behavioral Sciences' Dagstuhl Perspectives Workshop was conceived with one problem in mind: that of facilitating the construction of an integrative infrastructure in Psychology and Behavioral Sciences. The motivation for this workshop was to `foster the discussion around the problem of understanding the Web as an integrative platform, and how e-science can help us to do better research.' With these points in mind, we gathered an interdisciplinary group of experts, including computer scientists, psychologists and behavioral scientists. In their research, they are addressing issues in data standards, e-science, ontologies, knowledge management, text mining, scholarly communication, semantic web, cognitive sciences, neurosciences, and psychology. Throughout the Workshop, this group worked on devising a roadmap for building such an interoperability layer.
The seminar started with a number of keynote sessions from well-known authorities in each area to introduce the necessary background and form a common baseline for later discussions. A core theme that emerged was the cross-domain challenge in establishing a common language. We jointly undertook the effort to define an integrative scenario illustrating how digital infrastructures could help psychologists and behavioral scientists to do research that takes advantage of the new digital research landscape. In order to achieve this, the computational scientists needed to better understand the current working practices of the psychologists. For instance, the nature and structure of their data and experiments; moreover, computer scientists needed to understand the flow of information, from the conception of an idea, through defining a study plan, executing it and finally having the investigation published. They learned that the work of psychologists and behavioral scientists strongly relies on questionnaires and experiments as ways of collecting data, and on statistics as a tool for analyzing data, and that the replicability of experiments is a key concern. In a similar vein, psychologists and behavioral scientists needed concrete examples illustrating how computer science enables FAIR (= findable, accessible, interoperable and reusable) infrastructures that allow researchers to discover and share knowledge - bearing in mind data protection issues.
Two break-out groups were organized. The purpose was to have a full picture of digital scholarship in action when applied to psychology and behavioral investigations, most importantly e-science assisting researchers in sharing, discovering, planning and running investigations. The full research life cycle had to be considered. Both groups worked up their respective scenarios independently. The visions were then exchanged in an inter-group meeting. Interestingly, various issues arose when discussing the specifics from each vision for digital scholarship; for instance, the importance of understanding scholarly communication beyond the simple act of getting one's results published. Furthermore, the need to integrate tools into platforms where researchers could openly register their projects and plan and manage their workflows, data, code and research objects, was extensively discussed. Within this framework, the need for controlled vocabularies, standards for publishing and documenting data and metadata, persistent identifiers for datasets, research objects, documents, organizations, concepts and people, open APIs to existing services and instruments, and reporting structures were understood; these elements were articulated in the examples where the researchers and research were at the center of the system. Discussions also addressed fears in the community and thus the need to open up the current research landscape in small steps.
The seminar proved to be a fertile discussion field for interdisciplinary collaborations and research projects across previously disparate fields with the potential of significant impact in both areas. The need for a digital infrastructure in psychology and behavioral sciences was accepted by all the attendants; communicating this message with a clear implementation vision to funding agencies, professional societies and the community in general was identified as a key priority. It was decided that we needed another meeting in 2016; during that follow-up, the emphasis should be on developing a research agenda. As this is a relatively new topic in psychology and behavioral sciences, it was also decided to contact publishers and professional organizations, e.g. the Sloan Foundation, the APA and the APS, and work with them in conveying the message about increasing openness. If we want to understand how cognition is related to the genome, proteome and the dynamics of the brain, then interoperability, data standards and digital scholarship have to become a common purpose for this community. Funding has to be made available, initially for an assessment of the uptake of existing key resources and infrastructures, and then for implementing further Digital Scholarship and Open Science infrastructures as well as for building the skills in a community that is not yet widely familiar with the relevant enabling technologies. Finally, once sufficient technical support is in place, sustainable incentives for sharing research objects should be put in practice.
- Xavier Aimé (LIMICS INSERM U 1142 - Paris, FR) [dblp]
- Dietrich Albert (TU Graz, AT) [dblp]
- George Alter (University of Michigan - Ann Arbor, US) [dblp]
- Björn Brembs (Universität Regensburg, DE) [dblp]
- Mike Conlon (University of Florida, US) [dblp]
- Oscar Corcho (Technical University of Madrid, ES) [dblp]
- Susann Fiedler (MPG - Bonn, DE) [dblp]
- Alexander Garcia Castro (Technical University of Madrid, ES) [dblp]
- Paul Groth (Elsevier Labs - Amsterdam, NL) [dblp]
- William Gunn (Mendeley Ltd. - London, GB) [dblp]
- Janna Hastings (European Bioinformatics Institute - Cambridge, GB) [dblp]
- Caroline Jay (University of Manchester, GB) [dblp]
- Iris-Tatjana Kolassa (Universität Ulm, DE)
- Silvia Koller (Federal University of Rio Grande do Sul, BR) [dblp]
- Christoph Lange (Universität Bonn, DE) [dblp]
- Maryann Martone (UC - San Diego, US) [dblp]
- Russell Poldrack (Stanford University, US) [dblp]
- Alec Smecher (Simon Fraser University - Burnaby, CA)
- Daniel Staemmler (Elsevier Publishing - Berlin, DE)
- Robert Stevens (University of Manchester, GB) [dblp]
- Gary VandenBos (American Psychological Association, US)
- Hal Warren (Vedatek Knowledge Systems, US) [dblp]
- Erich Weichselgartner (ZPID - Trier, DE) [dblp]
- data bases / information retrieval
- data structures / algorithms / complexity
- semantics / formal methods
- data standards
- semantic web
- open science