ConductScience VR
Context fear conditioning in virtual reality
People learn which of two virtual rooms brings a mild electric shock, as rats and mice learn to fear the chamber where they were shocked.
ConductScience Editorial · Published November 15, 2026 · 5 references
What virtual context conditioning tests
In context conditioning, the shock comes while the subject is in a particular place, and no discrete cue signals it.5 For a rat or mouse, that place is a conditioning chamber, and Kroes and colleagues ask what the human version of the chamber should be.5 The testing room itself makes a poor choice: human studies compare each person between two places, so a lab would need a second real room, one without shocks.5 Virtual reality supplies both rooms while keeping experimental control.5 With headsets now sold commercially, Kroes and colleagues write, the method is within reach of nearly any lab that already records human conditioning with physiological equipment.5
Baas and colleagues guided people through two virtual rooms joined by a street scene, a shock room and a safe room.2 In each room, a blue and a yellow panel on a wall served as cues, and one panel came with a shock in the shock room.2 No shock was given in the safe room.2 Startle probes, sudden loud noises, played with a panel in view measure fear of the cue, and probes played with no panel in view measure fear of the room.2
Context conditioning is a model of sustained, free-floating anxiety, which arises when nothing clear in the surroundings signals the threat.5,3 Baas and colleagues proposed virtual environments as a tool to study cued fear and contextual anxiety side by side in people.2
How a session runs
Kroes and colleagues built their version with an off-the-shelf headset and a cross-platform game engine, and made the paradigm publicly available.5 A virtual living room and a kitchen with a dining area are joined by a hallway.5 The living room has yellow walls, ceiling and floor, and the kitchen blue ones, so that fear learned in one room does not spread to the other.5
Explore the rooms first
People first explore the rooms freely for 2 minutes with a game controller, then rate how pleasant and how arousing each room feels.5 Animal research shows that exploring a place before any shock is critical for learning to fear it.5
Set the shock level and let the startle reflex habituate
The shock is a 200 millisecond pulse to the right wrist, raised until the participant finds it maximally uncomfortable but not painful.5 Before the task, people hear 9 startle probes, each a 50 millisecond burst of 100 dB white noise, so that the startle response can habituate, or shrink with repetition.5 Startle is measured as the size of the eye blink each probe sets off, recorded by electromyography.5
Guide people along a set path
During conditioning, people are guided along a fixed path through both rooms and do not steer.5 They can turn their heads, but they are asked to stay mostly still so that movement does not disturb the physiological recordings.5 The path runs from the hallway into a room, back to the hallway, into a room, and so on.5 Each room is entered 10 times, for 30 seconds each time, and each of the 20 hallway trips lasts 15 seconds.5 On 4 hallway trips the path turns back 180 degrees into the room just left, so that people cannot predict the next room with certainty.5 The whole task takes 15 minutes.5
Give shocks in one room only
Shocks come only in one room, never in the other room or the hallway, while startle probes play in all three places.5 A probe or shock can come 5 to 25 seconds after a person enters a room, with at least 5 seconds between events.5 Each person received 8 shocks in the shock room.5 People are not told which room brings the shocks, so they have to learn it from experience.5
Measure fear of each room
Startle blinks and skin conductance, a measure of sweating, are recorded in each room, both to the probes and on entering the room.5 After the task, people rate each room again and estimate how many shocks they received in each.5
Kroes and colleagues chose a guided path after pilot testing, in which steering through the whole task made a number of volunteers nauseous and kept some from attending to the shocks.5 The headset tracks head position, so people can turn and look in any direction, though they are asked to keep mostly still to protect the physiological recordings.5
What the test has shown in people
Fear of the room itself
After conditioning, startle was larger in the shock room than in the safe room, even with the shock-paired panel out of view.2 The shock-paired panel raised startle in both rooms, which Baas and colleagues read as fear of the cue spreading across contexts.2 In the headset task of Kroes and colleagues, people rated the room with shocks as more negative and more arousing than the other room.5 Late in conditioning, their startle blinks were larger in that room, and their skin conductance rose more on entering it.5 Afterward, people estimated more shocks in that room than in the other, and their estimates did not differ from the 8 shocks actually given there.5
Shocks that come without warning
Grillon and colleagues guided people through three virtual rooms, one with no shocks, one where shocks followed a colored light and one where shocks came at random.3 Startle between the lights was larger in the room with random shocks than in the other two rooms.3 Later, when people could choose two of the three rooms to enter and collect money, they strongly preferred the no-shock room and avoided the room with random shocks.3
From fear to avoidance
Glotzbach and colleagues gave mild electric shocks in an anxiety context and never in a safety context.1 People then chose two of three contexts to visit again, with a neutral context added.1 They avoided the anxiety context, but they did not prefer the safety context over the neutral one.1 People who showed strong conditioning in their ratings of valence (how pleasant a context felt), arousal and anxiety avoided the anxiety context.1
Brain activity
Alvarez and colleagues combined virtual contexts with functional MRI, pairing an unsignaled shock with one context and never giving one in the other.4 Skin conductance and anxiety ratings confirmed that context conditioning had taken place.4 Compared with the safe context, the shock context activated the right anterior hippocampus and the amygdala on both sides.4 Context conditioning was also linked with activity in the orbitofrontal cortex, thalamus, insula, anterior cingulate and parahippocampal cortex, among other areas.4 In 2017, Kroes and colleagues noted that commercially available headsets could not yet be used in an MRI scanner.5
Comfort in a headset
In the task of Kroes and colleagues, no participant dropped out or complained of motion sickness, and the authors judged the task well tolerated on average.5 Asked afterward, three of the 22 participants reported no discomfort, six a tiny bit and five slight discomfort.5 Four were moderately uncomfortable and four very uncomfortable, each with slight nausea.5 No one chose the most extreme answer, very uncomfortable and very nauseous.5
How it compares with rodent fear conditioning
In laboratory animals, the context is changed by moving the animal to a different conditioning box, or by changing its wall colors, floor texture or smell.5 Virtual reality places people in different environments under experimental control, much as rodent context conditioning does.5 Alvarez and colleagues take their brain scans as evidence that similar brain mechanisms may underlie context fear conditioning across species.4 Unpredictable stressors strengthen context conditioning in animals, and Grillon and colleagues found the same effect of unpredictable shocks in people.3
Learned fear shows as defensive responses such as freezing and a rise in sympathetic arousal.5 Human versions measure startle and skin conductance, and they add what a rodent cannot report, explicit memory and feelings of fear and anxiety.5 People in the guided task did not steer, which Kroes and colleagues list as a limitation, since rodent research suggests self-movement can affect how spatial representations form.5 They suggest that a longer first exploration might let people steer throughout in future studies.5
Build a virtual fear conditioning task with ConductScience VR
We write the task software in house and build each fear conditioning task to the protocol you plan to run.
- The rooms, their colors and furniture, the hallway, and cues such as colored panels or lights
- A guided path or free movement, with your room order, time in each room, return trips and avoidance tests
- The timing of startle probes and shocks in each room, with triggers for your shock stimulator and markers for your physiological recordings
- A time stamp for every room entry, probe and shock, with room ratings, shock estimates and avoidance choices, saved for analysis
- A desktop screen with a keyboard or joystick, or a headset
- The software alone, or a package with the headset and computer
Related pages
Rodent behavior tests rebuilt in virtual reality for people, and how we build them.
Our rodent fear conditioning chamber, for cued and context conditioning in mice and rats.
Two virtual rooms, one paired with a reward, and the time people choose to spend in each.
Open arm avoidance and anxiety in a human plus maze, in virtual and mixed reality.
Questions
- Is virtual context conditioning the same test as the rodent one?
- It keeps the core of the rodent test, a shock tied to a place with no discrete cue to signal it.5 Each person is compared between a room with shocks and a room without, so they serve as their own control.5,2
- What kind of shock do these studies use?
- A mild electric shock, which Kroes and colleagues set for each person at a level that was maximally uncomfortable but not painful.1,5
- How do you tell fear of the room from fear of a cue?
- Startle probes played when no cue is in view measure fear of the room, while probes played during a cue measure fear of the cue.2,3 Grillon and colleagues probed startle in the gaps between colored lights for the same purpose.3
- Do people avoid a room they have learned to fear?
- Yes, people later avoided the anxiety context in the study of Glotzbach and colleagues.1 Grillon and colleagues saw avoidance of the room with random shocks and a strong preference for the room with none.3
- Can virtual context conditioning run in a headset?
- Yes, Kroes and colleagues ran it in a headset with positional tracking. No one dropped out, and the authors judged it well tolerated on average, though some people felt slightly nauseous.5 We build it for a headset or for a desktop screen with a keyboard or joystick.
References
- Glotzbach E, Ewald H, Andreatta M, et al. Contextual fear conditioning predicts subsequent avoidance behaviour in a virtual reality environment. Cognition and Emotion. 2012. doi:10.1080/02699931.2012.656581
- Baas JM, Nugent M, Lissek S, et al. Fear conditioning in virtual reality contexts: a new tool for the study of anxiety. Biological Psychiatry. 2004. doi:10.1016/j.biopsych.2004.02.024
- Grillon C, Baas JM, Cornwell B, et al. Context conditioning and behavioral avoidance in a virtual reality environment: effect of predictability. Biological Psychiatry. 2006. doi:10.1016/j.biopsych.2006.03.072
- Alvarez RP, Biggs A, Chen G, et al. Contextual fear conditioning in humans: cortical-hippocampal and amygdala contributions. Journal of Neuroscience. 2008. doi:10.1523/jneurosci.1246-08.2008
- Kroes MCW, Dunsmoor JE, Mackey WE, et al. Context conditioning in humans using commercially available immersive Virtual Reality. Scientific Reports. 2017. doi:10.1038/s41598-017-08184-7
More virtual reality tests
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
ConductScience VR
