Brain AnalyzIR: A software platform for improving scientific rigor in functional NIRS statistical analysis

Abstract. Functional near-infrared spectroscopy (fNIRS) is a non-invasive neuroimaging technique that uses light to measure changes in cerebral blood oxygenation. Our Brain AnalyzIR toolbox was developed to specifically deal with the unique statistical and noise properties of fNIRS data. This tool was introduced based on the observation that without consideration of fNIRS-specific issues including correction for serially correlated errors, heavy-tailed noise distributions introduced by motion artifacts, and the highly heterogenous spatial noise typical of fNIRS reflecting probe placement and anatomical features, the fNIRS analysis methods popular at the time generated very high false-positive rates and uncontrolled type-I error. These prior methods which had been derived from existing tools from other fields (e.g., fMRI) were not sufficient to deal with these fNIRS-specific errors, which led to a low level of reproducibility and rigor in the fNIRS field. The Brain AnalyzIR toolbox is a MATLAB-based, fNIRS-specific, statistical program to correctly address these issues. In this grant renewal of the original R01 funding for the toolbox, we propose to continue to develop this tool and support the more than a hundred end-users that have now adopted these methods in their research. The aims of this project seek to improve the current infrastructure and to make this more accessible to users, as well as to add new features requested by users, and to innovate new methods based on future directions of the fNIRS field. The specific aims of this proposal are: Aim 1. Infrastructure of the AnalyzIR toolbox. Aim 2. Improve the community usability and generalizability of the toolbox Aim 3. Implementation and testing of currently needed fNIRS methods. Aim 4. Innovating for the future.