Exploring indirected relations between topics in neuroscience literature using augmented reality to inform experimental design

Before conducting a costly experiment, neuroscientists need to identify potentially useful hypotheses. Given the vast amount of neuroscience literature, it is useful to gain a bird’s-eye view of the field to understand what information is established and what may inform future experiments. To achieve this, topic-based literature exploration in Augmented Reality (AR) has been used to investigate relations between neuroscience topics, such as brain regions and brain diseases. The DatAR team at Utrecht University has developed a 3D-based AR prototype that provides a visual representation of direct relations between brain regions and brain diseases. A co-occurrence of two topics, such as a brain region and a brain disease, in the same sentence of the title or abstract of a publication implies a direct relation between them. These direct relations can help neuroscientists intuitively understand which brain regions are affected by specific brain diseases. A brain region may also be connected indirectly to a brain disease through an intermediate topic, such as a gene or a mental process. We define an indirect relation between two topics when there is no direct relation between them, but each co-occurs with at least one other intermediate topic. Neuroscientists have proposed that identifying such indirect relations could provide additional insights for experimental design. These indirect relations may reflect weak evidence of a potential link between two topics that has not yet been confirmed by any single publication. This motivates Study 1 on exploring indirect relations between topics. I follow a user-centred design approach: defining functional requirements for exploring indirect relations, designing interactive AR visualisations for the specified functionalities, and engaging neuroscientists in evaluating the usefulness of exploring indirect relations. Six of the participating neuroscientists noted that the identified indirect relations primarily serve as inspiration for further literature review, rather than as definitive evidence for designing an experiment. They suggested that tracking trends in indirect relations over time and identifying when these indirect relations become established as direct relations in the literature could provide evidence for the usefulness of current indirect relations. This insight motivated Study 2 on exploring the evolution of indirect to direct relations using a timeline. One neuroscientist also suggested exploring the specific intermediate topic, such as a gene, responsible for the indirect relation. Identifying the specific intermediate topic could help neuroscientists understand the mechanisms underlying the indirect relation and better assess its usefulness. This suggestion motivated Study 3 on exploring indirect relations via a specific intermediate topic. My research provides an interactive 3D AR approach to assist neuroscientists in identifying indirect relations between neuroscience topics. It also offers additional information, such as the evolution of indirect to direct relations and the specific intermediate topics involved, helping researchers assess the usefulness of the indirect relations. By enabling the exploration of hundreds of thousands of publications simultaneously and supporting the identification of useful hypotheses, the DatAR prototype serves as a complementary step in the early phase of designing a potentially useful experiment.

• View at Publisher

Free Full Text ( Final Version , 32mb )