Prospective Study of Effects of Sleep-Wake Dynamics on Seizure Predictability
Patients with refractory focal epilepsy admitted for subdural intracranial evaluation for resective surgery will be given the option of participating in the study. after an informed consent is obtained, monitoring will proceed as per the existing protocol except for the addition of a few scalp electrodes to allow scoring of sleep. Based on the american association of Sleep Medicine (aaSM) guidelines, scalp eeG will be acquired from the following locations: F3/F4, C3/C4, o1/o2, and a1/a2 (or M1/M2), along with two electro-oculogram (eoG) channels to assess eye movements and submental eMG to assess muscle tone. eKG is already a standard measurement in the intracranial monitoring protocol. Respiratory parameters will not be recorded as our objective is sleep staging rather than diagnostic evaluation of sleep apnea. Given the possibility that the bandages from the craniotomy interfere with placement of scalp electrodes ipsilateral to the craniotomy, our plan is to place the scalp electrodes on the contralateral side, i.e., only one of each of the above scalp eeG electrode pairs will be utilized. These additional scalp channels will be recorded synchronously with intracranial eeG and saved in digital format. in the event of electrode detachment during the monitoring period, the clinical staff will reattach the electrodes and make appropriate annotations.
Selected interictal, preictal, ictal and postictal data segments will be scored according to aaSM guidelines by Dr. Modur and his staff. Patient data will then be stripped of protected health information (PHi) to the extent allowed by the vendor software without corrupting the files. The data files will then be shared by the research team, including Dr. Sunderam at the university of Kentucky, who will perform further analysis related to seizure predictability in consultation with Dr. Modur. analysis will proceed in the following directions:
1. intracranial PSG. The possibility of identifying intracranial contacts that allow sleep-wake discrimination with accuracy comparable to scalp will be explored. First, a few intracranial grid contacts roughly corresponding to the F3/F4, C3/C4 and o1/o2 sites will be selected for comparison with scalp. Then a supervised sleep-wake state classifier will be constructed with the help of PSG-scored data and tested on out-of-sample recordings. The ability to score sleep using ieeG will permit retrospective analysis of sleep-seizure interactions and their effect on seizure predictability in existing ieeG databases that do not include scalp eeG or eoG/eMG measurements.
2. analysis of sleep-seizure interactions. Correlations between sleep stage and seizure likelihood will be computed. one important hypothesis to be tested is that seizures are not merely facilitated by certain states, but occur during transitions between states such as arousal from sleep, brief arousals, and transition back into deep sleep.
3. Seizure predictability analysis. Linear and nonlinear dynamical measures touted for their seizure predictive ability  will be computed for these recordings, and their performance as seizure predictors will be tested retrospectively on these recordings. Further, the effect of sleep-wake dynamics on SPa performance will be quantified by correlating SPa output with sleep-wake state and the possibility of proposing state-dependent corrections for improved performance will be investigated.
1) Individuals with epilepsy (focal epilepsy included) who are subdural intracranial evaluation for resective surgery.
2) Willing to undergo invasive Monitoring for treatment of refractory epilepsy.
3) Placement of subdural grid for intracranial monitoring
4) Ability to provide consent