Abstracts

Rationale: Functional connectivity (FC) measures can differentiate epileptogenic zones (EZs) from Non-EZs in patients with medically refractory epilepsy during interictal resting state.¹ However, there is a lack of research formally evaluating the stability of stereo-EEG (SEEG) FC measures and the influence of anti-epileptic drugs (AEDs) on FC measures. Therefore, our study aims to answer the following critical questions: 1) are measures influenced by AEDs, 2) are measures stable over time, 3) does patient collection state impact FC measures?

Methods: In 32 patients with medically refractory focal epilepsy, 2 minutes of SEEG pseudo-rest data (awake, eyes open, non-standard) was collected for days 1-7 post-implantation and 2 minutes of SEEG rest data (awake, eyes closed, standard). FC measures were evaluated using non-directed measures, imaginary coherence (ImCoh) between and within,² and directed measures, partial directed coherence (PDC) inward and outward.³ Standard clinical interpretation of ictal SEEG data was used to classify brain regions as EZs and Non-EZs using Desikan Killiany atlas. AED dosages were recorded for days 1-7 post-implantation and to standardize AED dosages across patients, a composite drug load score (DLS) per day was calculated. Linear mixed effects models were used to evaluate the relationship between time and DLS with FC measures.

Results: EZs consistently exhibited stronger FC than Non-EZs when using ImCoh between (p=0.0018-0.023), ImCoh within (p=0.025-0.036), and PDC inward (p=0.0089-0.016) for days 1-7 post-implantation. PDC outward did not distinguish between EZ and Non-EZs. When values were reordered from rank of highest to lowest DLS (Fig 1A), a linear mixed effects model demonstrated a significant negative relationship between the difference in ImCoh between FC and rank, indicating the difference between EZ and Non-EZ was smaller when patients were on lower AED dosages (p=0.033), whereas ImCoh within and PDC inward was not affected by DLS (Fig 1B&C). FC measures of EZ and Non-EZ regions were compared between resting state versus pseudo-rest for each measure to determine differences between collecting condition. This demonstrated ImCoh between and within FC of EZ and Non-EZs was highly correlated between collection conditions (both p< 0.001, Pearson Correlation), whereas PDC inward and outward were not (both p >0.05), with PDC inward better distinguishing between EZ and Non-EZ with resting state data.

Conclusions: Some FC measures are not impacted by AED dose, such as PDC inward, while other FC measures, such as ImCoh between, perform better on higher AED doses. However, ImCoh between and within were highly correlated between rest and pseudo-rest collecting conditions, whereas PDC inward and outward were not, with resting state better differentiating between EZ and Non-EZs. PDC inward is less affected by AED, but ImCoh between measure is less affected by condition, and all measures were stable over time, so selection of FC measure should be influenced by AED dose/variability and condition of the study.

Funding: Please list any funding that was received in support of this abstract.: R01 4047380141.