 This paper proposes a novel approach for detecting abnormal traces in cardiodecography, CTG, signals. It uses three different machine learning algorithms, one-dimensional convolutional neural networks, one DCNN, long short-term memory, LSTM networks, and two-dimensional convolutional neural networks, two DCNN. The one DCNN and LSTM networks were trained on data from 51,449 births at term while the two DCNN was trained on data from 20 minutes of FHR recordings. The one DCNN LSTM model achieved the best results with a partial area under the curve PUC of 0.20 and a sensitivity of 20 percent. Future work will involve expanding the dataset, analyzing longer FHR traces, and incorporating clinical risk factors. This article was authored by Daniela's fall, Ivan Jordanov, Lawrence Impey, and others. We are article.tv, links in the description below.