Loading…
CNS*2023 Leipzig has ended
32nd Annual Computational Neuroscience Meeting
CNS*2023, Leipzig, Germany
Monday, July 17 • 16:40 - 18:40
P128: BDBRA: Database for extracting descriptions of neural projections and brain regions from neuroscience literature with dialogue prompts

Log in to save this to your schedule, view media, leave feedback and see who's attending!


Brain Reference Architecture (BRA), an approach to developing brain-based software, uses Brain Information Flow (BIF), which extracts mesoscopic-level information about brain anatomy, to design a hypothetical component diagram (HCD)[1]. In BRA-driven development, it is expected that computational functions corresponding to anatomical structures and neural activities can be constructed as functional hypotheses, which can be used as software development in a wide range of brain regions by linking the functions of various brain regions in the whole brain to the computational processes.  The present BIF relies on the WholeBIF database, which is compiled by experts reading neuroscience articles. As it is challenging to encompass the extensive range of anatomical findings in a single human-constructed database, it is possible that an accurate BIF may not be available when referring to BIF at the time of design,  depending on the Region of Interest (ROI) of the HCD designer. Therefore, we describe the Bibliographic Database for BRA (BDBRA), a database that automatically extracts the anatomical connections to construct BIF from the neuroscience literature. To extract anatomical neural projections from neuroscience literature, we have previously employed an approach of extracting brain region-related terms from images in the literature and retrieving the corresponding paragraphs for data processing. However,  advancements in large-scale language models, such as ChatGPT, have paved the way for a novel approach to extracting anatomical neural projections from neuroscience literature. Specifically, by feeding literature into these models and engaging in interactive dialogue, including both questioning and discussion, it has become possible to discern brain region names and corresponding functional descriptions within the text. This new technique represents a significant advancement to extract key information from scientific literature, and has the potential to yield significant breakthroughs in our understanding of neural systems. Furthermore, it is possible to extract information regarding the transmission and reception of neural projections between brain regions, as well as the functions of these projections, through the use of questioning in the extraction process. This represents a promising avenue for further exploration and understanding of neural systems and their interconnectivity. This presentation will summarize information about dialogue prompts that would appropriately extract knowledge about brain regions from the neuroscience literature and provide suggestions about prompts that are useful for extracting neuroscience knowledge.


Acknowledgement This work has been supported by the Mohammed bin Salman Center for Future Science and Technology for Saudi-Japan Vision 2030 at The University of Tokyo (MbSC2030).


References
1. Yamakawa, H.  The whole brain architecture approach: Accelerating the development of artificial general intelligence by referring to the brain. Neural Netw. 2021, 144, 478-495.



Monday July 17, 2023 16:40 - 18:40 CEST
Kongresshalle

Attendees (6)