Feed

Functional magnetic resonance imaging (fMRI) is the most commonly used technique for non-invasive exploration of human cognition and neuroscience. Conventional T2*-weighted single-echo resting-state fMRI (SE-rfMRI) is limited by signal loss in regions prone to susceptibility artifacts and inseparable confounding from non-neuronal baseline (i.e., initial signal intensity S0) fluctuations. While multi-echo optimally combined rfMRI (ME-OC) effectively mitigates signal loss, confounding from S0 fluctuations remains unaddressed. Here, we leverage quantitative T2* fluctuation from ME-rfMRI to derive intrinsic functional metrics that are biophysically cleaner and more robust. Across both cortical and subcortical brain regions, we systematically evaluated the test-retest reliability and functional validity of two key local metrics, amplitude of low-frequency fluctuation (ALFF) and regional homogeneity (ReHo), derived from our proposed method against those from conventional SE-rfMRI and ME-OC. In subcortical regions, T2*-derived metrics showed the highest mean ICCs among the three signal representations for both ALFF (ICC = 0.67, 0.66, and 0.54) and ReHo (ICC = 0.73, 0.67, and 0.63) for T2* fluctuation, ME-OC, and SE-rfMRI, respectively. A similar regionally selective pattern was observed in cortical areas susceptible to magnetic field inhomogeneities, including the orbitofrontal cortex (ALFF ICC = 0.87, 0.75, and 0.82), insula (0.83, 0.61, and 0.71), and temporal pole (0.80, 0.73, and 0.79) for T2* fluctuation, ME-OC, and SE-rfMRI, respectively. In terms of functional validity, ALFF from T2* fluctuation uniquely revealed significant condition-related differences in the right putamen, inferior frontal gyrus, and suborbital sulcus, which both ME-OC and SE-rfMRI failed to detect. Together, our findings suggest that quantitative T2*-derived metrics may provide a reliable and more biophysically specific characterization of intrinsic fMRI signal fluctuations, particularly in subcortical and susceptibility-prone cortical regions, and may complement conventional approaches in future research applications.