The medial temporal lobe (MTL) is important for mnemonic processes, navigation and social cognition. At the basis of all these cognitive processes lies the integration of a multitude of stimuli and their classification into conceptual groups. Neurons in human MTL indeed show concept-specific activity profiles in that they fire specifically for pictures presenting concepts of animals, landscapes, faces, concept relations and their grade of novelty. How such representations are organized during moving scenes, where different concepts are presented simultaneously, has so far received little attention. Here, we analysed data from seven epilepsy patients with intracranial electrodes implanted in the amygdala, anterior and posterior hippocampus and entorhinal cortex. Patients were watching a movie in which periods with emotionally charged facial expressions alternated with periods of restful landscape. Next, we segmented the movie to separate other factors such as degree of movement or visibility of the most prominent features. From 385 neurons isolated from different MTL regions we identified 60 that changed their firing rate when fearful faces were presented. These responsive neurons were distributed across all regions, with the amygdala exhibiting the highest (25%) and the posterior hippocampus displaying the lowest (9%) proportion of responsive neurons. Furthermore, we observed that 30% of the neurons were responsive to video sequences characterized by intense motion. We isolated various features of the presented videos and incorporated them into a linear model. We thereby distinguish the unique response characteristics of neurons within different regions of the human MTL to dynamic stimuli.