脳が情報を処理する方法はどう違うのか

タイトル: Comparison of the cortical hierarchy between macaque monkeys and mice based on cell-type specific microcircuits

概要: The primate neocortex contains a hierarchy of cortical areas, with feedforward connections running from lower to higher levels, and feedback connections running in the opposite direction. The relative hierarchical position of cortical areas has been well established by retrograde tracing studies that allow to determine whether type of output from different source areas is predominantly feedforward or feedback. This method can not determine whether the cortico-cortical input within target areas is predominantly feedforward or feedback. We here make use of cell-type specific microcircuits to provide an intrinsic measure of the strength of feedforward versus feedback processing within cortical areas in macaque monkeys and mice. This allows a more complete map of the cortical hierarchy of different species, and therefore a more direct cross-species comparison. Parvalbumin-expressing interneurons were used as a marker of feedforward processing and calretinin-expressing interneurons as a marker of feedback processing. We found steep gradients in the distributions of these two interneuron types across macaque monkey cortical areas, indicating a deep cortical hierarchy where early visual areas are dominated by feedforward neural circuits and higher cortical areas become dominated by top-recurrent circuits. In contrast, the gradient in interneuron distribution across mouse cortical areas was limited, indicating a shallow cortical hierarchy and remaining dominated by feedforward neural circuits. This implies that the mouse cortex is comparable to early visual areas in primates, remaining dominated by bottom-up input. While primates are unique in having a deep cortical hierarchy that allows neural processing in higher cortical areas to become predominantly internal. Significance statementThe cortex in primates is known to contain a hierarchy of cortical areas, where bottom-up processing allows sensory information to get more and more compressed towards higher cortical areas, and where top-down processing allows lower areas to remain informed about the internal goals of the animal. However, it has been challenging to directly compare cortical hierarchies between animal species. Here, we developed an intrinsic measure of the strength of bottom-up and top-down processing within cortical areas, based on the density of different types of interneurons within areas. Or results indicate that the mouse cortical hierarchy is shallow and is driven by bottom-up processing, while the primate cortical hierarchy is deep and slowly becomes dominated by internal processing towards the top.