Object recognition requires the visual system to maintain stable category assignment while flexibly handling changes in viewpoint. Although the encoding mechanisms of object category and viewpoint have been well characterized in the primate ventral visual pathway, it is still unknown how the avian brain, which lacks a layered cortical structure, implements similarly complex visual computations. Here, we systematically examined how neurons in the pigeon entopallium (ENTO), the terminal stage of the tectofugal visual pathway, encode object identity and viewpoint. Large-scale electrophysiological recordings revealed that ENTO neurons exhibited category selectivity and viewpoint tuning, with a subset showing strong joint sensitivity to both dimensions. At the population level, neural activity formed high-dimensional representational manifolds that simultaneously supported categorical separation and viewpoint continuity. Furthermore, ENTO neurons were highly sensitive to color, and subpopulation analyses revealed a progressive integration of low-level visual features (e.g., color, shape) with high-level semantic representations. Collectively, our findings provide neurophysiological support for category representations in the ENTO and further lend support to a cross-species, geometry-based encoding framework.
Juncai et al. (Thu,) studied this question.