Patients with intractable epilepsy can benefit from surgical resection of the tissue generating the seizures. However, difficulties in precisely localising seizure foci have limited the number of patients undergoing surgery and lowered its efficacy, with seizure recurrence in the 10 years post surgery in over half of patients¹. Stereoencephalography (SEEG) is the current gold standard for presurgical monitoring. However, it may suffer from imprecise localisation of the epileptogenic zone1, as it is insensitive to sources which are more than about 1cm distant from or tangentially oriented with respect to the nearest electrode².
Fast neural Electrical Impedance Tomography (fnEIT) is a method for imaging electrical impedance altered by brain activity. In the brain, multiple transfer impedances are measured by injecting safe, insensible electrical currents into the brain of epilepsy patients undergoing presurgical monitoring through scalp and also depth electrodes previously implanted for SEEG recordings. Images of fast electrical circuit activity every millisecond are reconstructed from the measured voltages using benchtop hardware similar to an EEG machine. In rat cortex, this has provided images with a resolution of <200um and 2 msec³. This would translate to <5mm in human epilepsy anywhere in the brain. It can also operate in ”slow” mode where different frequencies are injected simultaneously through 32 pairs of electrodes, and real-time images are reconstructed, similar to those that would be obtained while the patient was in an fMRI, but can be recorded continuously at any time during presurgical video-telemetry. In pigs, EIT enabled accurate reconstruction of seizure onset with chemically-induced (benzylpenicillin, BPN) epilepsy which were 9±1.5 mm from the BPN cannula and 7.5±1.1 mm from the closest SEEG contact (p<0.05, n =37 focal seizures in 3 pigs)⁴. Modelling, using co-ordinates from 5 implanted depth electrodes in 3 human subjects, indicates much improved spatial coverage and localisation accuracy compared to SEEG².
Work in progress is to record EIT images in human patients during SEEG telemetry. Using the improved spatial information from EIT in addition to SEEG, we aim to build a model of the epileptogenic network and study the spatiotemporal dynamics that drive seizure initiation and propagation.  This will facilitate the ultimate goal to transition to intelligent EIT-aided closed-loop reversible neurostimulation techniques for the treatment of refractory epilepsy.