Micro-electrode array (MEA) assay using human iPSC-derived neurons are expected to one of in vitro assays to predict the toxicity and predict the mechanism of action of drugs. MEA subteam of NeuTox Committee in Health and Environmental Science Institute (HESI) have started the pilot study for the prediction of seizure liability of drugs. In this study, we aimed to develop an analytical method enabling the evaluation of toxicity of convulsants using deep learning. Human iPSC-derived cortical neurons and astrocytes were cultured on 24-wells MEA plate for extracellular recording using MED64 Presto. HESI twelve compounds were tested at 5 concentrations for each compound (n>6). We firstly had artificial intelligence (AI) learned the data of convulsants and the data of non-convulsants. Next this AI predicted the Toxicity of the data not used for learning. The toxicity probability of unlearned sample data was 90% or more, and the toxicity probability of the unlearned convulsants was also 80% or more. In addition, the negative probability of non-convulsants was more than 80%. These results indicated that this AI analysis method is useful for predicting the convulsion toxicity using hiPSC-derived neurons.