Abstracts

Rationale: Epilepsy is a chronic neurological disorder characterized by recurrent spontaneous seizures. Abortive interventions based on scalp or intracranial EEG (iEEG) have shown limited efficacy, partially due to the involvement of large populations of neurons in the epileptic discharge at the time of seizure detection. Few animal studies and a handful of human studies have demonstrated changes in single neuron activity preceding the seizure onset. We aimed to systematically study the activity and interactions among single neurons before the onset of epileptic seizures in humans. Methods: We included patients with drug-resistant focal epilepsy who undergo iEEG using depth electrodes [Behnke-Fried intracerebral depth electrodes, Ad-Tech, Racine, WI, USA] as part of their clinical pre-surgical work-up. 6-12 depth electrodes were implanted stereotactically, and patients were monitored in the epilepsy monitoring unit, until enough seizures were recorded allowing precise localization of the epileptic focus. The epileptic focus was confirmed by electro-clinical and neuroimaging data. Each iEEG electrode has 8 macro-contacts [1.3-1.6 mm2 surface, 5 mm contact spacing] and a set of 9 platinum?"iridium micro-wires [40-m diameter] protruding from its distal tip. All patients had micro-wires located in the epileptic focus. For recording we used a 256-channel acquisition system [NeuroPort monitoring system, Blackrock Microsystems, Salt Lake City, UT, USA]. The macro-contacts record iEEG sampled at 2kHz, and micro-wires record local field potentials (LFPs) and single neuron activity sampled at 30kHz. iEEG data were reviewed by an epileptologist and the times of seizure onset and termination were marked, as well as the location of the focus. Data from the micro-wire recordings were pre-processed and single neuron activity extracted using supramagnetic clustering and wavelet analysis. Neuronal firing rates (FR), inter-spike intervals (ISI), and cross-correlations were analyzed before, during and after the seizure and their waveforms assessed for cell-type specification (putative excitatory or inhibitory neurons, example in Figure 1). Results: We recorded neuronal activity of 82 neurons before, during and after epileptic seizures (total of 6 seizures from 3 patients). 26 neurons were found in the focus (25 excitatory, 1 inhibitory), and 56 neurons outside the focus (55 excitatory, 1 inhibitory). In the focus, 10 excitatory neurons changed their FR prior to seizure onset (7 increased, 3 decreased) and 16 neurons did not change their FR prior to seizure. Outside the focus, 2 excitatory neurons decreased FR and 54 neurons did not change their FR prior to seizure. Inhibitory neurons did not change their FR prior to seizure. Prior to seizure onset, more neurons changed their FR In the focus than outside the focus (p < 0.01, X2 test). We found an increase in FR in the focus starting as early as 280 seconds before seizure onset [Figure 2, Patient 1]. At the time of FR change, no visible changes were in the EEG (not shown). Conclusions: Excitatory neurons change their FR prior to epileptic seizures almost exclusively in the epileptic focus, and not outside the focus. The change in FR is heterogeneous, but more neurons increase than decrease their FR. FR changes can occur more than 280 seconds prior to first visible ictal changes on raw iEEG. Our data support the use of single neuron activity for early seizure detection and seizure prediction in humans. Funding: This work was supported by TLVMC BRAIN grant.