Phase-Cone Biomarkers Derived from Microgrid ECoG and Its Temporal Derivatives During Seizure Evolution and Reading

Abstract EEG and micro-electrocorticographic (µECoG) recordings exhibit transient spatiotemporal phase-cone structures that are thought to reflect cortical phase transitions underlying cognitive processing and epileptic activity. In this study, we investigated phase-cone dynamics in human micro-electrocorticographic (µECoG) recordings obtained from an 8 × 8 cortical microgrid implanted near a seizure onset zone. Phase-cone rates (PCRs) were computed from the original signal (PCR1), its first temporal derivative (PCR2), and its second temporal derivative (PCR3) in theta (4–8 Hz), alpha (8–12 Hz), beta (12–30 Hz), and low-gamma (30–50 Hz) frequency bands during an evolving focal seizure and a normal reading task. All three PCR measures exhibited a robust spectral hierarchy, increasing progressively from theta to low-gamma frequencies. During a reading task, PCR activity was relatively stable and showed modest temporal fluctuations consistent with ongoing cortical state transitions. In contrast, seizure activity produced substantially larger PCR magnitudes and greater temporal variability, particularly in the beta and low-gamma bands, indicating enhanced phase-transition dynamics during ictal network evolution. Analysis of temporally differentiated signals demonstrated that organized phase-cone structures remain detectable after first- and second-order differentiation, suggesting that cortical state transitions are reflected not only in instantaneous phase patterns but also in their temporal rates of change. These findings indicate that phase-cone dynamics are preserved across multiple dynamical representations of cortical activity and support the use of derivative-based PCR measures as complementary biomarkers of mesoscopic cortical organization. The proposed framework may provide new insights into normal cognitive processing, seizure dynamics, and other conditions involving large-scale cortical reorganization.