Abstract: PURPOSE: Detection and recognition of biological motion play an important role in evolutionary survival of human. The biological motion, which was termed by Johansson, was produced by a few lights attached to the head and major joints of an actor. Even without contours linking the joints, observers can still quickly identify the biological motion as actions. Despite of extensive studies suggesting the superior temporal sulcus (STS) a key to processing biological motion information, the neuronal mechanisms underlying the processing of biological motion features remain unknown. In the dorsal pathway of the visual system, several areas were specialized for the analysis of complex motions, including the medial temporal area (MT), the middle temporal area (MST), the fundus of superior temporal area (FST) and the superior temporal polysensory area (STP) in the STS. Our lab’s previous study has revealed that MST can encode biological motion information by the way of dynamic. In this study, we aimed to explore the neuronal mechanism of biological motion information processing in FST, especially focusing on form, walking direction and body orientation. And then, we sought to compared the encoding abilities of FST with those of MST to uncover potential functional distinctions between these two areas in processing biological motion.
METHODS: We examined the neuronal properties in FST by training macaque monkeys to perform passively viewing tasks. Once a cell was isolated, we used the receptive field (RF) test to determine the spatial location and size of the cell’s RF. Then, different optic flow stimuli were presented in the RF to investigate the motion direction tunning and optic flow selectivity, including four non-linear patterns (expansion, contraction, clockwise rotation and counter-clockwise rotation) and eight linear stimuli (dots translating at 8 directions with 45° apart). Subsequently, we randomly presented eight biological motion stimuli in the form of point-light in the RF. These stimuli varied in terms of form (intact VS scrambled), walking direction (right VS left), body orientation (upright VS inverted). Form scrambled animations were created by randomly placing the initial position of all dots. Right-walking and inverted biological motion was obtained by horizontally and vertically mirroring the left-walking and upright point-light walkers, respectively.
RESULTS: The responses of FST neurons revealed that some cells preferred intact biological motion, a minority of cells exhibited a preference for scrambled animations. Additionally, some cells preferred right or left walking walker, while some cells showed a preference for upright or inverted biological motion. To further understand how FST encode biological motion features at the population level, we employed the modified F1/F0 technique, same as the analysis approach used in MST in our previous study. The technique is to evaluate the neuronal modulation based on the dynamic structure of spike trains over the time course of biological motion. The results demonstrated that FST is capable of extracting form from biological motion, distinguish vertical (body orientations) and horizontal spatial transformations (walking directions), but does not exhibit a preference for upright biological motion, which was observed in MST.
CONCLUSIONS: FST can encode form, body orientation, and walking direction of biological motion. However, unlike MST, FST does not show a preference for upright walkers, indicating the absence of inversion effect in FST. These findings imply that MST and FST might have distinct role in processing biological motion information.

Key words: monkey, the fundus of superior temporal area, biological motion, form extracting, walking direction.