In the evolutionary mental state transition model, language, vision, and acoustic features are first extracted from the data and encoded, retaining their chronological order.