Abstracts

Rationale: The implantation of invasive electrodes is associated with acute trauma and inflammation, potentially impacting the integrity of intracranial electroencephalographic (iEEG) recordings obtained during this period. The behavior of continuous human recordings during this period compared to longer time scales has potential implications for clinical interpretation and algorithm development but is not yet well described. We aim to characterize the temporal and spatial variability of chronic EEG recordings across 300 days following implantation. Methods: Fourteen patients with drug-refractory epilepsy originally implanted with 16 subdural platinum iridium macroelectrodes and continuously monitored for an average of 18 months were included in this study. Time and spectral domain features were calculated for each channel in one-minute non-overlapping windows and averaged over 24 hours. Within and between subject means, standard deviations, and coefficients of variation were calculated for each feature. A linear mixed model with autoregressive moving average errors was used to model each feature and to characterize transient group-level changes in feature values. Results: A significant change in mean line length, energy, and half-wave as well as mean power in delta, alpha, theta, beta, and gamma bands was observed in the majority of patients over the course of 100 days following implantation. In addition, the coefficient of variation across electrodes followed the same trend. All power bands exhibited a similar degree of change (~30%), stabilizing after 100 days. Conclusions: A better understanding of how changes occurring at the brain-electrode interface affect iEEG will improve our ability to properly interpret associated recordings. Our findings show increased variability both temporally and spatially in the initial periods of recording, suggesting that extended monitoring may be required to properly assess the underlying electrophysiological network. This also has implications for the development of neurodevice algorithms, which must account for this initial variability. Funding: This study was funded by National Institutes of Health (NIH) and the Mirowski Family Foundation grants through the University of Pennsylvania: NIH U01NS073557, T32 NS091006. The International Epilepsy Electrophysiology Portal is funded by the NIH (U24NS063930).