TLDR: This research paper introduces a vision for constructing a knowledge graph from the Motor Research (MO|RE) data repository in sports science. The approach utilizes a Basic Formal Ontology (BFO)-based ontology to formally represent the interrelation of plan specifications, specific processes, and related measurements. The goal is to standardize and make motor performance data machine-understandable, addressing challenges with traditional data processing and enabling advanced AI analysis. The paper also highlights the importance of privacy-aware extensions for sensitive sports science data and demonstrates the modeling with an example of the Handgrip test, showcasing its potential for interdisciplinary collaboration and improved data reusability.
In the dynamic field of sports science, understanding and comparing human physical and cognitive capabilities is paramount. This often involves extensive testing of motor performance, generating vast amounts of data. However, traditional methods of handling this data often fall short, especially with the rise of advanced artificial intelligence and large language models that demand machine-understandable and interoperable information.
A recent research paper, titled “Ontologies in Motion: A BFO-Based Approach to Knowledge Graph Construction for Motor Performance Research Data in Sports Science,” addresses this challenge head-on. Authored by Sarah Rebecca Ondraszek, Jörg Waitelonis, Katja Keller, Claudia Niessner, Anna M. Jacyszyn, and Harald Sack, the paper outlines a visionary approach to transform how motor performance data is modeled and shared.
The Challenge with Current Data
The Motor Research (MO|RE) data repository, developed at the Karlsruhe Institute of Technology, serves as a crucial infrastructure for publishing and archiving sports science research data, particularly in motor performance. While MO|RE offers valuable data, its current organization, based on traditional relational databases, struggles to formally represent complex temporal relationships and the intricate connections between predefined test items, their executions, and the resulting measurements from real-life tests. This limitation hinders sophisticated automated analysis and interdisciplinary integration.
A Semantic Solution: Ontologies and Knowledge Graphs
The researchers propose a groundbreaking solution: creating a knowledge graph from MO|RE data using an ontology rooted in the Basic Formal Ontology (BFO). Ontologies and knowledge graphs are powerful tools that organize scientific knowledge in a way that enables machines to understand, integrate, and even discover new insights across different fields. The goal is to make motor performance data standardized and machine-understandable, fostering greater collaboration and evidence-based practices.
The Basic Formal Ontology (BFO) Foundation
At the heart of this approach is the BFO, an upper-level ontology that provides a foundational framework for representing entities and their relationships. BFO distinguishes between ‘continuants’ (entities that persist through time, like a person) and ‘occurrents’ (processes that unfold over time, like a test). This distinction is crucial for capturing the temporal complexity inherent in motor performance research. BFO also defines ‘qualities’ that inhere in their bearers, allowing for the formal representation of measurable features such as height, weight, or fitness evaluation results. By building upon BFO, the MO|RE ontology ensures a high degree of interoperability with other knowledge systems, making cross-disciplinary applications possible.
The MO|RE ontology also integrates concepts from other established ontologies, such as the Information Artifact Ontology (IAO) for representing information content and measurement data, and the Ontology for Biomedical Investigations (OBI) for experimental procedures.
Addressing Privacy and Data Protection
A critical aspect of sports science data is its sensitive nature, often including personal information like anthropometric values and health indicators. The paper highlights the need for robust privacy mechanisms within the ontology and knowledge graph to comply with regulations like the European Union’s General Data Protection Regulation (GDPR). The proposed solution involves annotating properties and classes that represent identifying or health-related data with metadata, allowing for automated distinction between sensitive and non-sensitive entities. Future work will focus on implementing role-based access control directly within the knowledge graph, ensuring different user groups have appropriate levels of access.
An Illustrative Example: The Handgrip Test
To demonstrate their modeling approach, the researchers use the Handgrip test as an example. In the MO|RE ontology, a research study is modeled as a ‘plan specification,’ an information artifact outlining a sequence of actions. Similarly, a ‘test item’ like the Handgrip test is also an ‘action specification’ that defines what participants are expected to do. The actual Handgrip test process of a participant is then modeled as a ‘test process,’ which realizes the specified plan. Participants are represented as ‘material entities’ with measurable ‘qualities’ and ‘dispositions’ that are realized through the test processes. The measurement results, such as the maximum force exerted, are captured as ‘scalar measurement data’ with specific values.
This detailed modeling allows for a clear distinction between the planned test, its concrete execution by a participant, and the resulting values. The approach is scalable and can be extended to other test items, such as a shuttle run, by adapting the connected entities and dispositions measured.
Also Read:
- Dynamic Knowledge: How Context Transforms Concept Graphs
- Unlocking Multi-Dimensional Data: A New Approach with LLM Agents and Knowledge Graphs
Impact and Future Directions
The development of the MO|RE ontology and its corresponding knowledge graph holds significant promise for the future of sports science. It enables the comparison of longitudinal motor test data, even when protocols vary, which is invaluable for researchers studying trends like changes in motor performance in children over time. This semantic infrastructure will facilitate interdisciplinary collaboration with fields like health, education, and psychology, unlocking new possibilities for research questions that transcend single-domain perspectives, such as the relationship between motor competence and academic achievement.
By structuring knowledge explicitly and making connections machine-readable, sports science can benefit from tools already well-established in bioinformatics and clinical research. This strategic investment in shared ontological frameworks promotes transparency, reproducibility, and reusability of data—core principles of open science. For more details, you can refer to the full research paper here.
