Neuroinformatics Assembly 2023 - Sessions
FAIR data: The role of journals
 Chair: Jan Bjaalie
Most neuroscience journals request authors to make their data publicly available in appropriate repositories. The requirements and policies put forward by journals vary, and the services provided for different types of data also differ considerably across repositories. Researchers trying to navigate to a suitable repository are challenged with several questions and a need to compromise parameters such as storage space available, time and efforts needed for data sharing, data governance considerations et cetera, all influencing the extent of FAIRness their data may achieve. Many journals provide data repository guidance, but with many options and priorities available, data tend to be dispersed across many repositories. In this session we discuss how journal policies and recommendations contribute to open and FAIR neuroscience and influence the findability and interoperability of public data. Topics:
|
AI for Neurodegenerative Diseases
Chair: Mah Parsa About two percent of the world's population suffers from various types of mental health disorders. Furthermore, more than 50 million people worldwide are living with different types of neurodegenerative dementias including Alzheimer’s disease (AD). Early detection of mental health disorders and dementia would help patients seek out different intervention programs, as well as clinical interventions so that they can maintain their quality of life at the normal level. The size of the above problems and their global effect requires international and multidisciplinary collaborations that address the needs of underserved patient populations without access to costly care in the hands of sparsely available specialists such as psychiatrists. To fill these gaps in a cost-effective and scalable manner, it is critical to foster research partnerships that result in the development of Artificial Intelligence (AI)-driven solutions to measure, diagnose, and treat mental and neurodegenerative disorders. Goals:We aim to discuss ongoing projects for:
By providing a comprehensive discussion on the above topics, the workshop will allow mental health professionals such as psychologists, psychiatrists and mental health counselors and researchers in the field of mental disorders and neurodegenerative diseases expand their knowledge regarding ongoing AI research projects and innovative AI applications. It also offers breathtaking advantages for the AI community to improve their apps by receiving valuable real-world feedback from clinical audiences. Expected outcomes:We expect to strengthen the collaboration between scientists and researchers from the computer science domain, in particular ML researchers in the field of speech, natural language and image processing and neurologists and mental health professionals. We also intend to motivate industry labs to financially and technically support research projects related to mental disorders and neurodegenerative diseases. We assume holding the workshop can encourage principal investigators at universities to currently employ services for the early diagnosis, treatment, and management of mental disorders and neurodegenerative diseases Additionally, the workshop will:
International neurologists and mental health professionals, neuroscientists, psychiatrists, and computer scientists (in particular ML scientists and data scientists) are the main groups of the workshop's audience. Attendance of the workshop will provide them a unique chance to start multidisciplinary, international cutting-edge research collaborations that could lead to original and innovative AI-based solutions for mental disorders and neurodegenerative diseases. Students and trainees from different disciplines can learn about advanced tools by attending this workshop. Schedule:We will endeavor to organize a one-day interactive workshop that will be started by keynote speakers to present: Challenges in diagnosing and/or quantitatively detecting neurodegenerative diseases and Alzheimer’s. and ML-based diagnostic/measurement systems that can mitigate these challenges. The workshop will include lightning talks on the topic of analysis of neuroimaging datasets for neurodegenerative diseases and Alzheimer’s, as well as a description of newly developed AI systems and ML algorithms that could be useful to analyze neuroimaging datasets. One of the exciting parts of our workshop is group discussions to review how the state-of-the-art AI systems and ML algorithms can be utilized for:
We will have two panel discussions: 1) an academy panel discussion and 2) an industry panel discussion to discuss the following questions:
The two panel discussions will aim to fill the gap between neurologists and senior researchers in mental healthcare and industrial leaders in AI. Also, to discuss AI entrepreneurs can effectively collaborate with psychiatrists, neurologists to use AI to detect neurodegenerative diseases. We also aim for such panels to encourage students in psychiatry, psychology, neuroscience and computer scientists, and share their innovative ideas with entrepreneurs to develop amazing AI systems to detect neurodegenerative diseases such as AD. In addition to the sessions mentioned above, we will offer two poster sessions during lunch and coffee breaks. Finally, we will wrap up the workshop by announcing awards for the “best” lightning talk and poster presentation and the “best” reviewer.
|
Practical guide to overcome the reproducibility crisis in small animal neuroimaging: workflows, tools, and repositories
Co-Chairs: Markus Aswendt, Joanes Grandjean Small animal MRI has undergone rapid development over the past decade in terms of improved hardware and acquisition protocols. Similar to human MRI, it is now possible to acquire whole brain functional and structural connectomes in a variety of disease models. Recent large-scale multicenter studies have shown that there is a non-negligible variability in MRI setups, animal protocols, and post-processing methods. Further, only a minority of labs actively apply standardized quality control, analysis tools, and FAIR data sharing. Audience:PhD students, Postdocs, (PIs). Expertise in programming not necessary. Prior experience with rodent MRI data acquisition and processing is a plus. Methods:The workshop will include interactive seminars given by selected experts in the field covering all aspects of (FAIR) small animal MRI data acquisition, analysis, and sharing. The seminars will be followed by hands-on training where participants will perform use case scenarios using software established by the organizers. This will include an introduction to the basics of using command line interfaces, Python installation, working with Docker/Singularity containers, Datalad/Git, and BIDS. Schedule:
|
Neuroscience data integration through use of digital brain atlasesCREDIT COURSE - Level: PhD, Credits: 2 ECTS
Trygve Brauns Leergaard, Ingvild Elise Bjerke - University of OsloUnderstanding how the brain works is one of the grand challenges in science and requires the integration of huge amounts of heterogeneous and complex data. Numerous research publications present experimental data at various levels of granularity and describe a wide range of structural and functional aspects of the brain. The management of this deluge of data represents a bottleneck for progress, with a main challenge being that the multiple data categories are difficult to compare. In this context, reference atlases of the brain are important tools for assigning anatomical location and (semi-)automatically analyzing data captured with the many methods and instruments used to study the brain. Reference atlases for the brain rank among the most frequently used and highest cited publications in neuroscience. But integration of data through the use of conventional reference atlases has been difficult to achieve. With a new generation of three-dimensional digital reference atlases, new solutions for integrating and disseminating brain data are being developed. In many ways, future digital reference atlases and the data systems that will be built around them will be similar to current geographical atlases, such as Google Maps and Google Earth, which provide interactive access to huge amounts of high resolution image data, together with additional information (annotations, practical information, photographs) and more detailed visualizations (e.g. "street view") for specific areas. Digital brain atlases play an important role in several large international projects, including the European Union ICT Future Emerging Technologies Flagship project, the Human Brain Project.This course contains an introduction to currently available reference atlases for mouse and rat brain. It will demonstrate how the 3D brain templates for the reference atlases are acquired, how they are used as a basis for delineating the structures of the brain, how they can be enriched by other data modalities, and how they can be used as a basis for assigning location (coordinate based or semantic) to a wide range of structural and functional data collected from the brain. The course will also outline examples of data system employed to organize neuroscience data collections in the context of reference atlases as well as analytical workflows applied to the data, with opportunities for hands-on exploration of selected tools. Read more about how digital brain atlases are used in the EBRAINS infrastructure at https://ebrains.eu/services/atlases Registration
Schedule Practical information about attendance via Zoom, program and syllabus / reading list will be published on this page. Examination will be in the form of a home assignment, to be completed within two days after the course (estimate work time = 8 hours). |
Streamlining Cross-Platform Data Integration: Processes and Solutions for Rapidly Developing an Integrated Workflow Across Independent Systems for the US BRAIN Initiative Cell Census
Chairs: Bing-Xing Huo, Patrick Ray, Maryann Martone
Advances in neuroscience increasingly require researchers and practitioners to work with data from multiple sources to gain insights into brain organization, function and disease. Many neuroscience data platforms have made these large, complex datasets available to the community. Integrating data from these diverse sources has been a challenging task as each platform is developed independently. At the same time, the infrastructure developed through large investments by the national-level brain projects is expected to be leveraged by new projects which may have custom workflows. Developing cross-platform data integration processes and solutions is essential therefore both for enabling integrated workflow from independent systems in a short time frame and gaining insights from their data. This workshop is designed as a forum for neuroscientists, data scientists, and software engineers to come together and discuss the challenges, lessons, strategies and best practices of cross-platform neuroscience data integration. Participants will also share and communicate on software tools and infrastructure solutions that support efficient data integration, such as data management systems, data integration platforms, cloud computing, and APIs. The workshop will cover topics at ecosystem level such as standardization, governance, and success metrics drawing from experiences in the US BRAIN Initiative Cell Census Network (BICCN) and its next phase, the BRAIN Initiative Cell Atlas Network (BICAN), with a focus on streamlining integrated workflows across independent systems. The challenge in both these projects was to quickly develop an end-to-end system to support the specific requirements of these consortia from systems that were developed independently. Projects like these present the ultimate use cases for FAIR, as FAIR practices should greatly reduce the barrier to integration and increase the usability of the data. Objective:Through a combination of presentations, panel discussions, and interactive sessions, participants will engage in an in-depth discussion on the principles of cross-platform data integration in neuroscience and learn how to apply these principles to their own platforms and research projects. Participants will also learn about open-source software tools and infrastructure solutions that support open data and open science. By the end of the workshop, participants will have increased awareness and alignment on the strategies and challenges of cross-platform data integration processes, better communication channels across independent systems, and also the skills and knowledge needed to develop efficient and effective cross-platform data integration. Target audience:The target audience for this workshop would likely be professionals working at the intersection of neuroscience and data science, including researchers, IT personnel, and bioinformaticians. The workshop would be beneficial for those who are involved in or responsible for the integration of data from independent systems and are interested in learning about processes and solutions for streamlining this process. Attendees may be at different stages of their careers, and may come from academic, industry, or government backgrounds. |
Research workflows for collaborative neuroscience
Co-Chair: Dimitri Yatsenko To tackle the challenging questions of the brain, scientists are turning to innovative strategies to automate and organize their research workflows. Automated research workflows integrate computing infrastructure, automation, instrument integration, advanced data operations, and machine learning to enable researchers to process vast amounts of data and analyze complex patterns. This enhances the speed, accuracy, and reliability of collaborative research activities and enables scientists to generate new insights and knowledge that were previously unattainable. The implementation of automated research workflows is not just about technical solutions but also about creating novel team structures that foster collaboration and innovation. With these tools, researchers can work more efficiently, share knowledge more easily, and focus on the intellectual challenges of their research. In this workshop, general approaches to workflow management, team organizations, software tools, and online platforms and resources will be reviewed. Across various projects, from neurophysiology to genomics, these workflows are transforming the way that scientists conduct research, opening up new avenues for discovery and innovation. |
Infrastructure for Sensitive Data
Chair: Petra Ritter
Regulation of personal data - such as medical brain data, has been radically transformed in recent years with the introduction of several new legislation, including the European GDPR (General Data Protection Regulation). This means an upheaval of prior work routines with the regard to collecting, processing and storing, as well as sharing personal data. These new regulations seek to establish a balance between safeguarding the data protection rights of citizens, while enabling innovative uses of research tools at the same time. We present the new EBRAINS privacy-compliant and scalable services for brain research to store, share and analyze complex multi-modal FAIR sensitive neuroscience data originating from human subjects: Health Data Cloud (HDC). The workshop provides participants both lecture and hands-on training session. Audience:The target audience for this workshop would likely be professionals working at the intersection of neuroscience and data science, including researchers, IT personnel, and bioinformaticians. The workshop would be beneficial for those who are involved in or responsible for the integration of data from independent systems and are interested in learning about processes and solutions for streamlining this process. Attendees may be at different stages of their careers, and may come from academic, industry, or government backgrounds. Topics:
9-10 Legal frameworks for sharing sensitive data at different legislations (Petra Ritter) |
Event annotation in neuroimaging using HED: from experiment to analysis
Chair: Scott Makeig Hierarchical Event Descriptors (HED) fill a major gap in the neuroinformatics standards toolkit, namely the specification of the nature(s) of events and time-limited conditions recorded as having occurred during time series recordings (EEG, MEG, iEEG, fMRI, etc.). Without standardized annotation of dataset events and conditions, neuroimaging researchers cannot re- use shared datasets to perform most standard or advanced analyses. Relating recorded brain dynamics to stimuli, experience and behavior requires extensive manual coding based on dataset- specific information gleaned from data authors’ notes or publications. Standardized annotation of temporal structure and events is therefore fundamental for new analyses exploiting new contrasts and commonalities across datasets and study designs. HED provides an easy-to-use machine-actionable framework for providing this annotation within BIDS-formatted datasets. In the last two years, the HED Working Group has made major advances in the structure, syntax, and tools for HED annotation of time series data saved in the BIDS format. These advances have made HED tagging of BIDS compatible datasets ready for widespread use and are documented at https://www.hed-resources.org. We, the HED Working Group, propose a half-day online INCF workshop on the need for, structure of, tools for, and use of HED annotation to prepare neuroimaging time series data for storing, sharing, and advanced analysis. Audience:PhD students, Postdocs, (PIs). Expertise in programming not necessary. Prior experience with rodent MRI data acquisition and processing is a plus. Methods:The workshop will include interactive seminars given by selected experts in the field covering all aspects of (FAIR) small animal MRI data acquisition, analysis, and sharing. The seminars will be followed by hands-on training where participants will perform use case scenarios using software established by the organizers. This will include an introduction to the basics of using command line interfaces, Python installation, working with Docker/Singularity containers, Datalad/Git, and BIDS. Schedule:
|