Automation of fNIRS research by ontological analysis

Efforts have been made to create guidelines [1] and standards [2] to improve the quality of the fNIRS experimentation and of the fNIRS research in general. Yet despite all the careful efforts during the experiment design and data collection, and despite the sophisticated analysis pipelines available nowadays including those with AI and ML, such experience-based efforts cannot however guarantee optimality. A neuroimaging experiment can be affected by factors of very different nature; from data collection flaws, to instrumental failures or the presence of unwanted physiological responses. Something as apparently simple as the optimal coupling between the experiment design and the subsequent analysis pipeline, which are known from basic mathematical principles to be tightly coupled, defies regular practice. In other words, best practices and standards are insufficient to optimize scientific practice in fNIRS or afford automation of science; a paradigm which is already achievable in other branches of science [3]. Underpinning such feat is an ontology of experimental design and analysis.

This project intends to advance OntoNIRS [4] from its current embryonic state into a fully operational ontology capable of supporting reasoning regarding experimental designs and analysis for fNIRS neuroimaging. In order to do so, the different stages of ontology engineering processes should be completed and the outcome verified and validated against existing literature as well as formal methods. To realize these goals, contributions in computational neuroimaging, knowledge representation (inc ontologies and logic) and natural language processing are needed in a truly multidisciplinary project requiring knowledge from fields as diverse as optics, mathematics, neurosciences, and computer sciences.

Applicants should have a strong background in mathematics and physics, and being enthusiastic about computer sciences. You should have a commitment to research in mathematical modelling and computer simulations and acquiring fNIRS neuroimaging data with applications in neuroscience. You must have exceptional programming and communication skills. You will be a team-player capable of independent learning.

Expected starting date: September 2025

About the Computational Neuroimaging Lab

Part of the School of Computer Science, the Computational neuroimaging lab develops models and analysis tools to understand the neural system. This involves multidisciplinary research from computing, mathematics and statistics, and a bit of physics and neuroscience. We have a key research area on optical neuroimaging modalities such as fNIRS and DCS.