The identification of a conspecific is an ecologically relevant task. In primates, face selective neurons are found in the inferior temporal cortex, a region along the ventral visual pathway that displays complex feature coding. However, studies that characterized the avian equivalent to the ventral visual pathway (tectofugal pathway), using images as stimuli, found no evidence of complex feature coding nor conspecific face selectivity. In this study, we used a set of naturalistic videos with controlled lower-level features to assess if the pigeon’s tectofugal pathway encode categorical representations of pigeons performing ecologically relevant tasks. The videos depicted pigeons performing four types of behaviors. Each one had a corresponding control, in which the motion of the behavior was preserved but the pigeon was substituted by an abstract shape. Six head-fixated pigeons watched the videos while single-cell activity was recorded from the mesopallium ventrolaterale (MVL), a visual associative area, and the nidopallium caudolaterale (NCL), the avian analogue to the mammalian prefrontal cortex. We used generalized linear mixed models to assess neuronal selectivity. All the analyzed MVL units were selective for at least one type of behavior, and 40% of them modulate their firing rate conditionally to the subject of the video (pigeon or control). In NCL, 80% of analyzed cells show selectivity for at least one of the behaviors, and 37.5% of them display a significant interaction between subject and behavior. So far, the results indicate that MVL encodes sufficient information to form categorical distinctions that are further specified in NCL.