Research Roundup (#31)
.png/:/rs=w:1440,h:1440)
Welcome...
Welcome to the thirty-first-ever Research Roundup! A fortnightly catch-up on the latest developments in the field of XR research.
It’s been a busy couple of weeks, so let's not hang around.
The Fortnight in 3 (Sentences)
New research on avatar design explored whether appearance and facial expression affect user perceptions of co-presence and trustworthiness, whilst it has been investigated how conversational fillers in VR-based generative AI could mitigate perceived latency issues.
In terms of therapeutic applications, researchers examined whether social VR environments could facilitate student counselling compared to video conferencing, and whether VR exposure therapy demonstrated superiority over traditional 2D video methods for treating fear of flying (it didn’t).
And finally, exploratory studies have investigated whether VR could induce stronger and more durable nostalgic feelings than traditional reflection tasks, and how VR visualisation of complex spectral data might enhance human detection of outliers beyond algorithmic analysis.
The Fortnight in 300 (words)
As conversations with avatars become more commonplace, researchers in France and Ireland have circled back to how they should look. They compared a conversation with a stylised (cartoony) avatar, a semi-stylised avatar, and a more realistic looking one. They also varied the intensity of the facial expressions. After playing a game of ‘Heads Up!’ with the avatar (controlled by an actress) participants rated the interactions on various dimensions. The intensity of the facial expressions did not affect those measures but, as you might expect, participants felt more social presence with the cartoon-like avatar as well as rating her less eerie and the most trustworthy. More evidence, if it was needed, that lower fidelity avatars have something going for them.
Exposure therapy (a treatment for fears and phobias) is frequently cited as an established use case for VR. Explored since the mid-90’s, it has established one of the most consistent and largest evidence bases. However, there have always been tiny niggles around some of the methods behind the evidence. These came to the forefront this week with the publication of data suggesting VR was no more effective for fear of flying treatment than using the same 360 videos on a 2D monitor. Whilst this new paper will itself go through a round of scrutiny, and methodological nit-picking, it nevertheless sends a clear message. It is increasingly important to have comparison groups that are ‘the same but 2D’ if we are going to narrow things down to the use cases for which VR truly excels. In a world of cost-effectiveness, a similar effect but without a headset will have a certain appeal (despite the new caveats which will go with this data).
And finally, in a world grappling with AI it was refreshing to read how VR could be used to exploit our natural capabilities and do something that an algorithim couldn’t! Apparently in the field of chemoanalytics algorithims have problems identifying outliers because of ‘masking’ and ‘swamping’ (basically a set of conditions that produce false negatives and false positive). Innovative new research has therefore explored the innate perceptual abilities of humans in combination with 3D/VR displays of the data as a potential solution. The researchers argue that putting the human back into the machine resulted in a deeper cognitive analysis than current algorithims are capable of. Take that you pesky machines…