The effects of dynamic facial information on sensitivity to facial expressions.

点击翻译按钮获取中文摘要

Facial emotion recognition is viewed as one of the defining aspects of human visual perception. However, for some members of the population extracting emotion-related information from faces is a challenging task. Deficits in facial emotion recognition are particularly prominent in autism spectrum disorders (ASD), and in certain mood-disorders. Behavioural intervention programmes have been used to facilitate facial emotion learning by utilising interactive games with computer generated agents. Exposure therapy using VR has had positive effects on reducing social phobias and effects of post-traumatic stress disorder. The objective of the present study is pilot novel stimuli that reflect naturalistic facial emotion perception within VR, with a long-term view for using this as a model for behavioural intervention programmes within clinical populations. This has been of particular interest to Sinewave; software developers whom the primary mentor of the project is already in connection with. A notable drawback facing current behavioural paradigms is their use of synthetic and 3D animated characters whose qualities do not reflect the fluid nature of human expression as it emerges in real time. This presents a gap within literature and industry that requires a realistic set of real-life, dynamic human facial expressions. If piloted now, such a stimulus set would lend itself to an expanse of areas, from clinical research areas, to cutting-edge user software development. The second mentor on the project will offer their expertise in this particular area, and possess the research background and hardware required to develop and pilot such a face set using VR. For facial stimuli to be effective tools used for measuring and improving facial emotion recognition, they must be realistic enough to simulate natural social perception: these objectives bring together expertise from both Abigail Webb and Peter Scarfe. The project will be the first to address how real-life, dynamic facial information in videos of human faces contributes to observers' emotion recognition accuracy. This is yet to be tested using an immersive VR platform; a cost-effective and novel technique for assessing and promoting the practice of facial emotion recognition in clinical populations. Such dynamic, real-time facial stimuli may be more effective when measuring natural facial emotion recognition, an advantage that is bolstered by using immersive VR. Our pilot study uses a non-clinical population, and will serve as a preliminary investigation of the efficacy of our model for immersive facial recognition. This is a particular source of interest to Sinewave. Using specialised 3D facial scanning techniques at the University of Reading (Peter Scarfe), it will be possible to create real human facial stimuli that are both 3D and dynamic, changing expression in real time. This will be investigated both outside of and within immersive VR, to measure the relative effects of VR context on facial emotion recognition. Software companies for service users, including Facebook Reality Labs (another pre-existing relationship with mentors) are now emphasising the importance of such research input, in order to develop the hardware and software necessary for relaying the nuances of such socially complex interactions. These applications are vast, and lend themselves equally well to clinical research, empirical areas of cognitive psychology, sociology, industry, and business. This area of research now taking off in areas outside of academia. We see this is the case with Facebook Reality Labs -another pre-existing relationship that exists with the primary mentor of the project- where the construction of cutting-edge immersive software is becoming increasingly informed by expert researchers within face perception. Together, this project unifies experts across psychophysics and face perception (AW), VR (LvD), creation of real-life human facial stimuli (PS).