Smoking-related cue reactivity in a virtual reality setting: association between craving and EEG measures

Abstract

Background: Cue-reactivity is the array of responses that smokers exhibit when exposed to conditioned and contextual stimuli previously associated to substance use. The difficulty to experimentally recreate the complexity of smokers' spatial experience and context requires more ecological models. Virtual reality (VR) creates a state of immersion close to reality allowing controlled assessments of behavioral responses. To date, no studies investigated brain activation associated to smoking cue-reactivity in VR using electroencephalography (EEG).

Aims: To investigate whether a VR cue-reactivity paradigm (a) may increase smoking craving, (b) is feasible with EEG recording, and (c) induces craving levels associated to EEG desynchronization.

Methods: Smokers (N = 20) and non-smokers (N = 20) were exposed to neutral and smoking-related VR scenarios, without and with smoking conditioned stimuli, respectively. EEG was recorded from occipital and parietal leads throughout the sessions to assess alpha band desynchronization. Smoking and food craving and presence visual analogue scales (VAS) were assessed during the session.

Results: To be smoker, but not non-smoker, significantly influenced smoking craving VAS induced by smoking cue VR but not by neutral VR. No significant food craving changes was observed during the VR sessions. The new finding was that EEG alpha band power in posterior leads was significantly increased by the smoking context scenario only in smokers, and that the degree of smoking (i.e., heavy vs. light) was significantly associated to this neurophysiological measure.

Conclusions: This study demonstrated, for the first time, the feasibility of EEG recording in a VR setting, suggesting that EEG desynchronization may be a neurophysiological marker of smoking cue-reactivity.

Keywords: Craving; Cue reactivity; Electroencephalography; Smoking; Virtual reality; Visual analogue scale (VAS).

Fig. 1

Screenshots of the virtual reality (VR) scenarios used in the study. Participants were exposed to a 3-min VR scenario (neutral, a mountain landscape; left panel), a 3-min VR office (control, non-smoking cues context; middle panel), and a 3-min VR office and smoking cues (smoking cue context containing conditioned stimuli; right panel)

Fig. 2

Experimental setup. Participants wore a standard EEG elastic cap, the virtual reality head-mounted display, and the headphones

Fig. 3

Schematic diagram of the study design. Each experimental session lasted about 45 min, including informed consent, screening procedure, baseline questionnaires, and virtual reality (VR) session. Visual analogue scale (VAS) for smoking craving, food craving, and sense of presence was filled out after the VR sessions. CX, context; CS, conditioned stimuli, i.e., smoking cues; VR-neutro, VR neutral scenario; VR-CX + CS, VR smoking context scenario

Fig. 4

Smoking craving. Left panel. Smoking craving score in the non-smokers (open columns; N = 20) and the smokers (solid columns; N = 20) groups at the different timepoints (VAS1 to VAS4). Right panel. Smoking craving score in the non-smokers (open columns; N = 20), light smokers (gray columns; N = 13), and heavy smokers (solid columns; N = 7) groups at VAS1 to VAS4. Data are expressed as mean + SEM. ****Significant within-group post hoc (p < 0.0001 vs. VAS1, VAS2, and VAS3; Sidak’s test). Between-groups significant post hoc comparisons are reported in the main text

Fig. 5

Food craving. Food craving score in the non-smokers (open columns; N = 20) and the smokers (solid columns; N = 20) groups at the different timepoints (VAS1 to VAS4). Data are expressed as mean + SEM. *Significant within-group post hoc (p < 0.05 vs. VAS1; Dunnett’s test)

Fig. 6

Sense of presence. Left panel. Presence scores in the non-smokers (open columns; N = 20) and the smokers (solid columns; N = 20) groups at different timepoints (VAS2 to VAS4). Right panel. Presence scores in non-smokers (open columns; N = 20), light smokers (gray columns; N = 13), and heavy smokers (solid columns; N = 7) groups at VAS2 to VAS4. Data are expressed as mean + SEM. *p < 0.05 and **p < 0.01 significant within-group post hoc vs. VAS2; Sidak’s test. #Significant between-groups post hoc (p < 0.05 vs. VAS2 in non-smokers; Dunnett’s test)

Fig. 7

EEG Alpha band power. Upper/central panel. Log₁₀ alpha band power in the non-smokers (open columns; N = 20) and smokers (solid columns; N = 20) groups during 3-min VR mountain landscape scenario (VR-neutro), 3-min VR non-smoking cues context office (VR-CX), and 3-min VR smoking cue context + conditioned stimuli (VR-CX + CS) measured at P4 and Pz electrodes. Lower panel. Log₁₀ alpha band power in non-smokers (open columns; N = 20), light smokers (gray columns; N = 13), and heavy smokers (solid columns; N = 7) measured at Pz electrode. Data are expressed as mean + SEM. ***Significant within-group post hoc (p < 0.01 vs. VR-CX; Sidak’s test)