Shared attention in virtual immersive reality enhances electrophysiological correlates of implicit sensory learning

Abstract

Shared attention effects on learning and memory demonstrate that experiences are amplified when we are not alone. Virtual reality poses new challenges to the study of co-presence. Above all, is coattending together with someone else's avatar in an immersive VR setting comparable with shared experiences at a neural processing level? In the present study we investigate shared attention effects in VR for the first time. We recorded mismatch negativities (MMN) during an auditory roving paradigm, a well-known index of implicit perceptual learning. EEG responses to deviant and standard sounds were registered while subjects were alone (Solo condition) or together (Other condition) with a virtual avatar (Virtual scenario) or physically present confederate (Physical scenario). We found an overall main effect of co-presence on MMN revealed by a point-by-point 2 × 2 ANOVA, thereby replicating previous studies on physical co-presence. Additionally, we found no significant interaction between the scenario (Physical vs. Virtual) and co-presence (Solo vs. Other). Our results indicate that virtual immersive co-presence mimics physical co-presence.

Keywords: Co-presence; Mismatch negativity; Shared attention; Virtual reality.

Figure 1

MMN waveforms. The graph represents MMN differential waveforms obtained by subtracting average responses to standard sounds from response elicited by deviant sounds at Fz. Coherently with previous studies the negative peak of MMN waveforms is frontally distributed and occurs at around 170 ms post-onset. Blue and red lines correspond to the Physical and Virtual scenario, respectively. Solid and dashed lines represent MMN corresponding to the Other and Solo conditions, respectively. The grey shaded area at 152–182 ms, corresponding to the MMN peak latency, represents the significant cluster surviving cluster correction in the point-by-point ANOVA and evidencing a significant main effect of experimental conditions Other vs. Solo. The grey line at the bottom depicts the corresponding F-values, showing a peak effect of condition corresponding to the MMN negative peak latency. Scalpmaps at the top represent the MMN amplitude distribution across the scalp for each single experimental condition. The scalpmap at the bottom shows cluster-corrected p-values corresponding to the main effect of co-presence (Solo vs. Other) resulting from the point-by-point ANOVA. The resulting significant cluster (152–182 ms) is localized over frontal electrodes overlapping the MMN scalp distribution.

Figure 2

Experimental conditions and MMN task. In roving auditory paradigms high-pitch and low-pitch intervals can represent both Deviant and Standard stimuli as shown in the top panel. MMN responses are elicited by sounds deviating from a sequence of repeated sounds. Human figures at the bottom represent the 4 experimental conditions. The order of the physical and virtual sessions was randomized across participants.