Excitation-inhibition imbalance in Alzheimer’s disease using multiscale neural model inversion of resting-state fMRI

Abstract

AbstractAlzheimer’s disease (AD) is a serious neurodegenerative disorder without a clear understanding of the etiology and pathophysiology. Recent experimental data has suggested excitation-inhibition (E-I) imbalance as an essential element and critical regulator of AD pathology, but E-I imbalance has not been systematically mapped out in both local and large-scale neuronal circuits in AD. Using a multiscale neural model inversion framework, we identified disrupted E-I balance as well as impaired excitatory and inhibitory connections in a large network during AD progression based on resting-state functional MRI data from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database. We observed that E-I balance is progressively disrupted from mild cognitive impairment (MCI) to AD and alteration of E-I balance is bidirectional varying from region to region. Also, we found that inhibitory connections are more significantly impaired than excitatory connections and the strength of the majority of excitatory and inhibitory connections reduces in MCI and AD, leading to gradual decoupling of neural populations. Moreover, we revealed a core AD network comprising mainly of limbic and cingulate regions including the hippocampus, pallidum, putamen, nucleus accumbens, inferior temporal cortex and caudal anterior cingulate cortex. These brain regions exhibit consistent and stable E-I alteration across MCI and AD, which may represent a stable AD biomarker and an important therapeutic target. Overall, our study constitutes the first attempt to delineate E-I imbalance in large-scale neuronal circuits during AD progression, which facilitates the development of new treatment paradigms to restore pathological E-I balance in AD.