The ability of anterior thalamic signals to predict seizures in temporal lobe epilepsy in kainate-treated rats.

ABSTRACT

OBJECTIVE: To analyze the local field potential (LFP) of the anterior nucleus of the thalamus (ANT) of epileptic rats using the Generic Osorio-Frei algorithm (GOFA), and to determine the ability of the ANT LFP to predict clinical seizures in temporal lobe epilepsy. METHODS: GOFA is an advanced real-time technique used to detect and predict seizures. In this article, GOFA was utilized to process the electrical signals of ANT and the motor cortex recorded in 12 rat models of temporal lobe epilepsy (TLE) induced via the injection of kainic acid into the unilateral hippocampus. The electroencephalography (EEG) data included (1) 161 clinical seizures (each contained a 10-min segment) involving the ANT and cortical regions and (2) one hundred three 10-min segments of randomly selected interictal (no seizure) data. RESULTS: Minimal false-positives (0.51 ± 0.36/h) and no false-negatives were detected based on the ANT LFP data processed using GOFA. In ANT LFP, the delay from electrographic onset (EO) to automated onset (AO) was 1.24 ± 0.47 s, and the delay from AO to clinical onset (CO) was 7.73 ± 3.23 s. The AO time occurred significantly earlier in the ANT than in the cortex (p = 0.001). In 75.2% of the clinical onsets predicted by ANT LFP, it was 1.37 ± 0.82 s ahead of the prediction of cortical potentials (CPs), and the remainder were 0.84 ± 0.31 s slower than the prediction of CPs. SIGNIFICANCE: ANT LFP appears to be an optimal option for the prediction of seizures in temporal lobe epilepsy. It was possible to upgrade the responsive neurostimulation system to emit electrical stimulation in response to the prediction of epileptic seizures based on the changes in the ANT LFP.