Measurement Prediction and Power Analysis for fNIRS and DOT

Functional near-infrared spectroscopy (fNIRS) and diffuse optical tomography (DOT) are valuable neuroscience tools, but their feasibility for some research questions can be uncertain due to interacting anatomical, measurement, and design limitations. Here, we present an experimentally informed computational framework for a priori power analysis that combines within- and between-subject variability estimates with forward-model–based estimates of task-evoked absorption changes to predict measurements. Cluster-based permutation tests applied to these measurements estimate statistical power for signal and contrast detection across sample sizes and stimulus block counts. We demonstrate how channel density and cortical source features—depth, size, spatial separation, and location—influence statistical power. Our approach reframes the physical constraints of fNIRS and DOT into interpretable design guidance and can be adopted to support principled power planning across diverse experimental designs.