Abstract
A balanced excitation-inhibition ratio (E/I ratio) is critical for healthy brain function, but challenging to measure in-vivo in humans. Here we present an approach to non-invasively estimate whole-cortex E/I ratio by fitting a large-scale biophysically plausible model to resting-state functional MRI (fMRI) data. We first confirm our model generates realistic brain dynamics in the Human Connectome Project. Next, we show that the estimated E/I ratio is sensitive to the GABA-agonist benzodiazepine alprazolam during fMRI. Furthermore, drug-induced E/I changes are spatially consistent with positron emission tomography measurement of benzodiazepine receptor density. We then leverage the model to investigate the hypothesis that the E/I ratio is normatively refined during development and supports improvement in cognitive ability. We find that the E/I ratio declines heterogeneously across the cerebral cortex during youth, with the greatest reduction occurring in sensorimotor systems relative to association systems. Importantly, among children with the same chronological age, a lower E/I ratio (especially in association cortex) is linked to better cognitive performance. This result is replicated across North American (8.2 to 23.0 years old) and Asian (7.2 to 7.9 years old) cohorts, suggesting that a more mature E/I ratio indexes improved cognition during normative development. Overall, our approach opens the door to studying cortex-wide E/I ratio changes across the lifespan and in neuropsychiatric disorders.
Introduction: The E/I Balance Challenge
Healthy brain function relies on a delicate "tug-of-war" between excitatory (E) and inhibitory (I) signals. When this balance is disrupted, it is linked to various neuropsychiatric conditions like schizophrenia, autism, and depression. However, measuring the E/I ratio in living humans is notoriously difficult. Traditional methods are either invasive (animal models), limited to small brain regions (MR spectroscopy), or require radioactive tracers (PET). This research addresses the lack of a non-invasive, whole-brain "map" of E/I balance, which is essential for understanding how the human brain matures and how individual differences in brain chemistry relate to intelligence.
Methods: The pFIC Modeling Framework
The authors developed a parametric Feedback Inhibition Control (pFIC) model. Unlike standard fMRI analysis which looks at correlations between regions, this approach uses a set of nonlinear differential equations to simulate actual neural populations (excitatory and inhibitory) within 68 different cortical regions.
To account for biological heterogeneity, the team parameterized local synaptic properties using two biological "gradients": an anatomical map (myelin density via T1w/T2w ratios) and a functional map (resting-state connectivity gradients). They utilized a "black-box" optimization algorithm (CMA-ES) to tune model parameters until simulated brain activity matched real-world fMRI data from over 1,000 participants in the Human Connectome Project, achieving high accuracy (r = 0.71).
Validation: Sensitivity to Pharmacological Modulation
To prove the model could detect real chemical changes, the researchers analyzed a pharmacological fMRI dataset where adults were given alprazolam (a GABA agonist that increases inhibition). The model successfully detected a significant reduction in the E/I ratio that spatially overlapped with known benzodiazepine receptor maps obtained via PET (r = 0.52), providing strong biological validation for the framework.
Results: E/I Ratio as a Marker of Cortical Maturation and Cognition
Applying the model to developmental cohorts—the Philadelphia Neurodevelopmental Cohort (PNC) and the GUSTO cohort—revealed that the E/I ratio is a dynamic marker of maturation. The mean cortical E/I ratio showed a strong negative correlation with age (r = -0.51), suggesting that E/I decline represents the "closing" of critical periods of plasticity as the brain moves toward a more stable, efficient state.
This decline is not uniform across the cortex. Sensory and motor regions mature and lower their E/I ratio early, while "association" regions responsible for complex reasoning show a much slower, more protracted decline. Perhaps most strikingly, among children of the same chronological age, those with lower E/I ratios in their association cortex performed better on cognitive tests. This suggests E/I balance is a robust indicator of "functional brain age" and cognitive potential.
Conclusion
This work provides a biologically interpretable lens for fMRI, moving the field closer to using brain scans as a diagnostic tool for the chemical imbalances underlying mental health disorders. By enabling non-invasive estimation of the E/I ratio across the whole cortex, it opens new avenues for studying brain development across the lifespan.
Summary updated