Rationale:
Aggressive tumors such as glioblastoma (GBM) significantly disrupt the brain microenvironment, often leading to neuronal hyperexcitability and seizures.1 Tumor-related epilepsy (TRE) not only degrades quality of life but may also accelerate GBM progression and worsen clinical outcomes, however the causal relationship between TRE and GBM prognosis is contentious, and a better mechanistic understanding is needed2. Our prior work demonstrated that epileptiform activity in the peritumoral cortex modulates tumor evolution and is associated with localized transcriptomic changes.3 However, studying these neural dynamics typically requires direct, invasive, technically demanding cortical imaging, a method unsuitable for real-time monitoring in human patients.
Methods:
Recent studies have identified the aperiodic 1/f slope of the EEG power spectrum as a promising non-invasive biomarker of cortical excitation/inhibition (E/I) balance.4. While this technique has been validated in non-tumor models, its application to TRE and longitudinal GBM progression has been limited. Here, we recorded and analyzed EEG from mouse models of GBM and controls under different behavioral (actively whisking/running vs resting) and anesthetic states (Isoflurane).
Results:
Data from GBM-bearing mice reveal a 5% decrease in 1/f aperiodic slope (i.e. relative increase in excitatory activity) during resting versus active (running+whisking) periods (N=3, SD=3.8%) and a 6% decrease over tumor progression (N=3, SD=3%). In contrast, this compares to control (WT) mice where the 1/f slope increased by 3% (i.e. a relative increase in inhibitory drive) during inactive vs active periods (N=2, SD=2.9%). Further, in another cohort of non-tumor control mice, we observed a 14% increase in the 1/f slope under isoflurane compared to the pre-anesthetic baseline periods (N=2, SD=7.4%). Our adapted EEG processing code allowed for 1/f slope calculation in < 10% of the recorded time span.
Conclusions:
Our findings suggest that aperiodic 1/f slope captures both tumor-induced shifts in E/I balance and anesthetic modulation of neural activity. WT control mice display an expected increase in inhibitory drive during inactive states vs periods of running/whisking behavior, while tumor-bearing brains instead seem to fail to establish this inhibitory restraint during inactivity. This is consistent with a lower seizure threshold, and appears to be mediated by relatively small differences in 1/f slope, consistent with the idea that the brain’s E/I balance is tightly constrained to operate within normal parameters. The increased slope under anesthesia supports its inhibition of cortical activity. Our results validate the potential of 1/f slope analysis as a non-invasive biomarker for monitoring epileptiform activity, GBM progression, and pharmacological effects in real time, further offering a clinically relevant tool for tracking neural dysfunction in patients with TRE. 1) Meyer, J., et al. Nat Commun 15, 4503 (2024), 2) Stella, M., et. al. Biomedicines 11, 582 (2023), 3) Soeung, V. et al. Cell Reports Medicine, Vol 5/8, 101691 (2024), 4) Donoghue, T.,
et al. Nat Neurosci 23, 1655–1665 (2020)
Funding:
NIH R01-CA263628 (to JM)