Figure 5.

The ibogaine-induced brain state is considered a REM-like state by an automatic sleep scoring algorithm. (A) Schematic representation of the neural network employed to classify the states of wakefulness and REM sleep. The network contains one input layer which receives the gamma features (power and coherence of the OB, M1, and S1 electrodes), and 10 hidden layers (only three of them are shown in the picture). The network has one output layer with two nodes (wake and REM). (B) Confusion matrix for an individual animal. (C) Bar plots showing how ibogaine wakefulness epochs were classified. Each animal is shown separately and was tested with its own control network.