中央研究院 資訊科學研究所

The essence of social interaction is the exchange of information between individuals. During this reciprocal process, the response timing and the response predictability play an important role. However, as social stimulus is the behavior of another animal, these parameters of the stimulus are difficult to manipulate. To enable a systematic control of the social stimulus, we are developing 3D models of adult zebrafish with controllable behaviors. The virtual fish, which is displayed on a computer monitor, performs naturalistic swimming patterns and adjusts its locomotion according to the behavior of a real animal swimming in an adjacent tank. The virtual and real animals thus establish a close-loop interaction in which experimenters have a full access to manipulate the response of the virtual animal through computer programing. To enable measurements of brain activity during behavior, we are also developing a VR system using a 6-axis force/torque sensor that in principal enables a head-restrained zebrafish to navigate in a 3D virtual space. This VR system is integrated into a two-photon microscope for non-invasive measurements of activity throughout the dorsal telencephalon. In addition, the respiration and the eye angle of the adult animal are simultaneously monitored, which enables us to correlate neural activities to behaviors driven by emotions and visual attention. Overall, this ongoing project demonstrates the potential of using virtual reality techniques to investigate the role of reciprocal interaction during social behavior.

I am a neuroscientist with experiences in developing new technologies. My past research experience has covered areas in cellular and systems neuroscience, in vivo brain imaging, and development of complex virtual reality (VR) systems. As a PhD student at Harvard University, I worked in Dr. Florian Engert’s lab investigating the role of hindbrain spinal projection neurons in controlling the swimming behavior of zebrafish (Huang et al., 2013; Ahrens et al., 2013), as well as the ontogeny of associative learning behaviors in zebrafish (Valente et al., 2012). As a postdoc in Dr. Rainer Friedrich’s lab at FMI, I developed the first VR system for adult zebrafish that enables the investigation of prediction error signals in the forebrain using a two-photon microscope (Huang et al., 2020). The novelty and impact of this study is recognized by the Chiquet Originality Prize issued by FMI in 2022. Since 2020 my lab at Academia Sinica has been investigating how social information is perceived, computed and integrated into learning circuits. In particular, we focus our research on social interactions with virtual partners in VR. We have developed realistic and interactive virtual animals whose behaviors can be fully controlled. We have also developed an advanced VR system which enables an animal to virtually navigate in a 3D virtual space while the brain activities across multiple regions can be monitored simultaneously at a single-cell resolution. We are currently investigating how the innate social information is delivered from the ventral forebrain to the dorsal forebrain to enable social learning. I believe that our approach of using VR to investigate social interaction is not only novel but also relevant as social media and other virtual platforms are intensively used to build interpersonal relationship in modern life.