Interactive epilepsy real-time prediction method based on neurodynamics

The invention discloses an interactive epilepsy real-time prediction method based on neurodynamics, and the method sequentially comprises the following steps: A, carrying out the collection and preprocessing of scalp electroencephalogram data, and obtaining a high-dimensional space scalp electroencephalogram sequence; b, establishing a spatio-temporal information conversion model, and optimizing the spatio-temporal information conversion model in real time; c, acquiring real-time scalp electroencephalogram data of the patient to be monitored, and acquiring a critical unpredictable early warning signal and a critical fluctuation early warning signal; and D, performing epileptic seizure prediction and epileptic seizure early warning notification according to the obtained critical unpredictability early warning signal and the critical fluctuation early warning signal. The method can quickly adapt to data change and effectively cope with epilepsy condition diversity and individual difference of electroencephalogr