A method for estimating functional connectivity corrected for the effects of the local hemodynamic response and its
Riku Tanabe1, Takuto Niinuma1, Masato Sugino1
1Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa-shi, Chiba 277-8563, Japan.
Abstract:
Functional connectivity, an index that quantifies temporal correlations between activities in different brain regions, has attracted attention for its potential applications in the diagnosis and pathological understanding of psychiatric disorders. Functional connectivity is estimated from fMRI blood-oxygen-level-dependent (BOLD) signals, which provide high spatial resolution. However, because the BOLD signal is an indirect measure of neural activity reflecting hemodynamic response, local variations in the hemodynamic response can hinder the accurate estimation of functional connectivity. Here, we proposed a method that corrects for the influence of the hemodynamic response in functional connectivity estimation by assuming that time lags observed in BOLD signal components originate from local variations in the hemodynamic response. By correcting the hemodynamic response while accounting for these time lags, which have a critical impact on functional connectivity that represents temporal correlations of signals, we achieved improved consistency across different sessions within the same subject (within-subject consistency). In addition, functional connectivity estimated by the proposed method outperformed conventional approaches in detecting patients with schizophrenia. These findings suggest that the proposed method enables reliable connectivity estimation and holds promise as a useful biomarker for psychiatric disorders.
