Ge, S., Luo, S., Yan, S. and Shen, X. (2022)

Why this paper is useful

This paper is less directly connected to autism or anxiety than the previous one, as it studies biofeedback within a VR performance rather than therapy. However, it gives me a useful interaction-design example of how real physiological information can be translated into visual and auditory changes inside a virtual environment.

The researchers combine heart-rate data from a performer with EEG data from the audience to investigate whether real-time biofeedback affects perception, concentration, immersion and understanding. The paper sits specifically within Human-Computer Interaction (HCI), human-centred computing and interaction design, which makes it particularly relevant to the design side of my project.

Turning invisible physiological data into part of an experience

One of the most useful ideas is that physiological information does not have to remain as numbers or medical data. The paper discusses transforming biosignals into visual metaphors, sound and adaptive feedback, effectively making an invisible bodily response visible within the virtual environment.

It identifies two approaches to representing biosignals in VR. They can either appear as separate visual or auditory symbols, such as abstract shapes or heartbeat sounds, or the biosignal can be mapped directly into the virtual environment itself. Previous examples include breathing controlling the growth of a virtual object and physiological signals altering elements within an experience.

Heart rate as real-time visual feedback

The researchers used a Polar H10 heart-rate monitor, while EEG brain activity was collected using a Muse 2 headband. The information was streamed into TouchDesigner and incorporated into a Unity-based virtual environment.

The heart-rate data was represented visually through changes within the scene. For example, the colour of visual effects changed depending on the performer's heart rate. The system also used auditory feedback: as heart rate increased, the heartbeat became faster and louder.

This gives me a practical precedent for: physiological change → data captured → virtual environment responds in real time.

Biofeedback does not automatically improve an experience

The researchers found that visual heart-rate feedback increased participants' concentration, and EEG Beta biofeedback improved narrative understanding. However, adding more sensory feedback did not necessarily make the VR experience better.

Continuous heartbeat audio actually distracted some participants. One participant reported that the heartbeat sound was annoying and affected their concentration, while another became distracted trying to understand where the sound was coming from. The study also found no significant improvement in immersion from one of the visual biofeedback conditions.

This feels particularly important for my autism project. A biofeedback system intended to regulate overload should not simply keep adding sound, movement and information. An intervention could become another source of sensory demand.

Self-sensory monitoring

The paper uses the term self-sensory monitoring, which is interesting for my research. It describes research in which physiological information is used to help represent or respond to a person's internal state. This connects to my interest in reducing reliance on somebody having to continuously recognise and verbally explain their own level of anxiety or overload. I still need autism-specific research to establish this connection properly, but this provides useful HCI terminology and precedent for designing around physiological information.

What I take from this paper

This paper does not prove that biofeedback VR will reduce anxiety or sensory overload in autistic adults. Its value is more as a design and technical precedent. It demonstrates that physiological signals can become active components of a VR experience, that heart rate and EEG can influence virtual visuals and sound in real time, and most importantly that the design of the feedback matters. More feedback is not automatically better. The more interesting question is how physiological information could be used to make an immersive experience respond appropriately to the individual without adding further sensory demand.

Key Terminology

Biosignals - Physiological information produced by the body, including heart rate, EEG and EDA.

Real-time biofeedback - Physiological information being captured and responded to while the experience is happening.

Sonification - Translating data into sound. In this study, heart rate was represented through heartbeat audio.

Affective visualisation - Visual representations linked to emotional or physiological states.

Human-Computer Interaction (HCI) - Research concerned with how people interact with technological systems; highly relevant to the interaction-design side of my project.