New methodology might scale back the time required to localize epileptogenic zones

The study discussed in this summary was published as a preprint on medRxiv.org and has not yet been peer-reviewed.

The central theses

• There is a disruption of the normal arousal / inhibitory balance in the brain of patients with epilepsy, and this study provided evidence that, in interictal rest, there is greater excitability in non-seizure-onset (non-SCO) zones compared to seizure-onset (SCO) zones .

• The flow of information was mainly from non-SCO to SCO, which may inhibit seizure activity.

• These results, combined with a random forest data model, enabled the localization of SOZs with an accuracy of 85% compared to conventional methods of interpreting SEEG (stereotactic electroencephalography) with seizure detection.

Why is that important?

• A third of patients with epilepsy become resistant to pharmacological treatments.

• Patients with refractory epilepsy may be candidates for surgical resection and undergo an SEEG to locate epileptogenic zones (EZ), a procedure that requires the temporary implantation of electrodes in the cortex. Seizure monitoring for localization can take days to weeks.

• The idle flow detection method allows EZS to be identified without waiting for seizures to occur, which has the potential to save monitoring time and costs.

Study design

• This study was a data analysis with random 10-minute periods from interictal recordings from long-term SEEG.

• Forty-three patients with drug-resistant focal epilepsy completed a preoperative SEEG between 2014 and 2019 at the University of Pittsburgh Medical Center.

• Directed connectivity was estimated between brain areas based on directed transfer function and cross-frequency directionality. Synaptic arousal-inhibition ratios were estimated using the power law exponent.

• A balanced Random Forest machine learning technique was used to predict SOZ at the individual electrode level.

Main results

• The flow of information was weighted on the movement from non-SCO to SCO during interictal idle states.

• The excitability of non-SCOs was increased compared to SCO.

• A larger discrepancy in information flow predicted favorable seizure outcomes.

• SOZ were located with an accuracy of 85% and the outcome of seizures was predicted with an accuracy of 77%.

limitations

• The use of the clinical SOZ to determine the EZ can be limiting as the SOZ can be a subset of the EZ.

• The 10 minute data period can be limiting as studies have shown that seizure activity fluctuates over time.

• More research is needed to examine how behavior, including sleep / wake states, affects outcomes.

Disclosure

• The authors are inventors of a US provisional patent application covering the analytical techniques filed by Carnegie Mellon University.

This is a summary of a preprint research study entitled “Interictal SEEG Resting-State Connectivity Localizes Seizure Onset Zone and Predicts Seizure Outcome” by Hiateng Jiang and colleagues at Carnegie Mellon University, University of Pittsburgh, and Massachusetts General Hospital on MedRxiv, provided to you by Medscape. This study has not yet been reviewed. The full text of the study can be found on medRxiv.org.