Relationship between aperiodic dynamics and transcriptomic alterations and a neural signature of glioma-induced excitation-inhibition dysregulation.

Abstract

2062 Background: Diffuse gliomas disrupt neuronal dynamics, leading to excitation-inhibition (E/I) imbalance and associated functional impairments. The aperiodic component of the power spectral density (1/f slope) has emerged as a proxy for estimating E/I balance, offering a novel framework for understanding glioma-induced neural dysregulation. This study is the first to validate the relationship between 1/f slope and E/I dysregulation in glioma by integrating electrophysiological, genomic, and behavioral data. Methods: Resting-state intraoperative subdural electrocorticography (ECoG) data were recorded from 13 glioma patients. Power spectral analysis at a frequency of 70–150Hz (high-gamma) computed 1/f slopes, and electrodes were classified as glioma-infiltrated or normal-appearing based on preoperative MRI T2-FLAIR. Linear mixed-effects models assessed E/I balance across tissue and glioma subtypes. Single-nucleus RNA sequencing (snRNA-seq) was performed on 14 spatially annotated glioma tissue samples from regions classified as inhibitory or excitatory by 1/f slope. Behavioral analysis of language tasks examined functional correlates of E/I imbalance. Results: The cohort included 23.0% WHO grade 2 IDH-mutant oligodendrogliomas, 38.5% WHO grade 2-3 IDH-mutant astrocytoma, and 38.5% glioblastoma (GBM). Glioma-infiltrated electrodes (n=142) exhibited significantly lower 1/f slopes than normal-appearing electrodes (n=518;p < 0.0001), reflecting an excitation-dominant state. Subtype analysis revealed hierarchical E/I imbalance, with GBM showing the steepest reductions in 1/f slope compared to astrocytoma and oligodendroglioma (glioma-infiltrated: p<0.0001; normal-appearing: GBM vs. oligodendroglioma, p=0.012; GBM vs. astrocytoma, p=0.019). SnRNA-seq revealed elevated excitatory and reduced inhibitory signaling gene expression in glutamatergic and GABAergic neuronal populations across glioma-infiltrated (n=12; n=4 per subtype) and normal cortex (n=2) samples. Excitatory module scores were significantly higher in excitatory 1/f samples compared to inhibitory 1/f samples, validating the 1/f slope as a genomic correlate of E/I imbalance in human cortical tissue. Behavioral analysis of language tasks demonstrated error-related reductions in 1/f slope (e > i), emphasizing the functional impact of glioma-induced dysregulation. Conclusions: Diffuse gliomas are associated with a profound shift toward excitation dominance in both glioma-infiltrated and normal-appearing cortex. For the first time, the 1/f slope is validated as a robust measure of E/I imbalance through electrophysiological, genomic, and behavioral analyses. These findings position the 1/f slope as a physiologically relevant biomarker of glioma-induced neural dysregulation, offering significant potential to inform therapeutic strategies.