N400 evoked potential lie detection method based on sample entropy

The invention provides an N400 evoked potential lie detection method based on sample entropy. The method comprises the following steps of extracting and synchronously amplifying electroencephalogram, collecting the electroencephalogram, receiving and synchronously storing the electroencephalogram, preprocessing the electroencephalogram, conducting superposition averaging on the electroencephalogram, extracting an electroencephalogram sample entropy feature value, conducting state judgment on whether a lie is told or not, measuring the complexity of the N400 induced electroencephalogram on the basis of the sample entropy, reflecting nonlinear features of the electroencephalogram, finding a feature threshold value for distinguishing whether a lie is told or not through statistical analysis of data, comparing the threshold value with an absolute value of the entropy gradient of probe stimulus and indifferent stimulus ERP electroencephalogram data samples, and judging whether a lie is told or not. The defects existing when lie testing judgment is performed by means of evoked potential waveform geometric properties in the prior art are overcome, so that the method has the advantage that the stable lie recognition accuracy can be effectively ensured.