Prior exposure to a sensorimotor game in virtual reality does not enhance stress reactivity toward the OpenTSST VR.

Compared to the in-person Trier Social Stress Test (TSST), virtual reality (VR) variants reduce resource-intensity and improve standardization but induce stress with smaller effect sizes. However, higher cortisol reactivity is given for more immersive TSST-VRs. Immersivity depends on the VR-system, but perceived immersion may be targeted by exposure to, or interaction with the VR. We investigated whether stress reactivity towards the openly accessible OpenTSST VR can be enhanced by prior exposure to a sensorimotor game completed in VR as mediated by increased immersion. Therefore, N = 58 healthy participants underwent the OpenTSST VR or its inbuilt control condition (placebo TSST-VR, pTSST-VR). Beforehand, participants completed a sensorimotor game either in VR or in real life. Stress was measured by means of self-reports, salivary cortisol concentrations, and salivary alpha-amylase (sAA) activity. Perceived immersion was assessed with the Igroup Presence Questionnaire (IPQ). The TSST-VR-group showed higher subjective stress than the pTSST-VR-group. Even though area under the curve measures indicated significant differences in cortisol levels between TSST-VR and pTSST-VR, this effect was not replicated in omnibus-analyses. Likewise, sAA was not responsive to stress. Our data suggests the OpenTSST VR does not reliably trigger physiological stress reactivity. Likewise, participants playing the VR-game before exposure to the TSST-VR did not show enhanced stress reactivity. Importantly, playing the VR-game did not lead to increased immersion (indicated by the IPQ), either. The key question resulting from our study is which manipulation may be fruitful to obtain a comparable stress response toward the TSST-VR compared to the in-person TSST.

A Model of Semantic Completion in Generative Episodic Memory.

Many studies have suggested that episodic memory is a generative process, but most computational models adopt a storage view. In this article, we present a model of the generative aspects of episodic memory. It is based on the central hypothesis that the hippocampus stores and retrieves selected aspects of an episode as a memory trace, which is necessarily incomplete. At recall, the neocortex reasonably fills in the missing parts based on general semantic information in a process we call semantic completion. The model combines two neural network architectures known from machine learning, the vector-quantized variational autoencoder (VQ-VAE) and the pixel convolutional neural network (PixelCNN). As episodes, we use images of digits and fashion items (MNIST) augmented by different backgrounds representing context. The model is able to complete missing parts of a memory trace in a semantically plausible way up to the point where it can generate plausible images from scratch, and it generalizes well to images not trained on. Compression as well as semantic completion contribute to a strong reduction in memory requirements and robustness to noise. Finally, we also model an episodic memory experiment and can reproduce that semantically congruent contexts are always recalled better than incongruent ones, high attention levels improve memory accuracy in both cases, and contexts that are not remembered correctly are more often remembered semantically congruently than completely wrong. This model contributes to a deeper understanding of the interplay between episodic memory and semantic information in the generative process of recalling the past.