Oscillatory correlates of threat imminence during virtual navigation

Abstract

The Predator Imminence Theory proposes that defensive behaviors depend on the proximity of a threat. While the neural mechanisms underlying this proposal have been well studied in animal models, it remains poorly understood in humans. To address this issue, we recorded EEG from twenty-four (15 female) young adults engaged in a first-person virtual reality Risk-Reward Interaction task. On each trial, participants were placed in a virtual room and then presented with either a threat or reward conditioned stimulus (CS) in the same room location (proximal) or different room location (distal). At a behavioral level, all participants learned to avoid the threat-CS, with most using the optimal behavior to actively avoid the proximal threat-CS (88% accuracy) and passively avoid the distal threat-CS (69% accuracy). By contrast, participants learned to actively approach the distal reward-CS (82% accuracy) and to remain still (passive) to the proximal reward-CS (72% accuracy). At an electrophysiological level, we observed a general increase in theta power (4-8 Hz) over right posterior channel P8 across all CS conditions, with the proximal threat-CS evoking the largest theta response. By contrast, distal CS cues induced two bursts of gamma (30-60 Hz) power over midline-parietal channel Pz (approx. 200 msec post-cue) and right frontal channel Fp2 (approx. 300 msec post-cue). Interestingly, while both bursts were sensitive to distal-CS cues, the first burst of gamma power was sensitive to the distal threat-CS requiring a passive response, and the second burst at channel Fp2 was sensitive to the distal reward-CS requiring an active response. Together, these findings demonstrate that oscillatory processes differentiate between the spatial proximity information during threat and reward encoding, likely optimizing the selection of the appropriate behavioral response.