Improving the pipeline for the data processing of fNIRS

Abstract

This thesis strives to meet the pressing demand within the ADvanced hEalth intelligence and brain-insPired Technologies (ADEPT) Lab for Biomedical Engineering researchers at Oslo Metropolitan University, for a more adaptable and customizable pipeline for pre-processing and post-processing of functional Near-Infrared Spectroscopy (fNIRS) data to examine cognitive activities. While existing tools such as nirsLAB, Homer and Satori offer some utility to the existing pipeline, their inflexibility and lack of customization restrict researchers from fully exploring the diverse and evolving landscape of cognitive neuroscience.

This thesis conducts a comprehensive review of neuroimaging techniques, tools, and data analysis methodologies, laying the foundation for designing and implementing a pipeline. It outlines the process of defining the problem, structuring the pipeline, and integrating theoretical and statistical aspects. The phased implementation demonstrates the pipeline's effectiveness in various data analysis stages, including GLM analysis, waveform averaging, functional connectivity analysis, and effective connectivity analysis. Furthermore, it discusses the implications of the findings, highlighting strengths, limitations, and potentials. Notably, it explores the potential for more efficient effective connectivity analysis and the hybridization of fNIRS-EEG data, opening new avenues for research unlike existing pre-processing software tools such as Satori, Homer, or nirsLAB.

To conclude, it outlines pathways for future research and development for the examining of fNIRS data, underscoring the pipeline's pivotal role in advancing the comprehension of cognitive activities and addressing health-related issues impacting individuals globally.