Toward a valid electrocortical correlate of regulation of craving using single-trial regression.

Regulation of craving (ROC) is a promising method of intervention in addiction and has received attention in event-related potential (ERP) research investigating the late positive potential (LPP). However, in the face of previously unaligned behavioral and LPP results, the present work aimed to take a first step toward more valid electrocortical markers of ROC. Twenty-seven smokers (16 female), abstinent from smoking for 3 h prior to testing, viewed smoking-related pictures while EEG was recorded. During picture viewing, participants were asked to either imagine the short-term benefits (NOW condition) or long-term negative consequences (LATER condition) of smoking, or engage in an arithmetic task (DISTRACT condition). After each picture, participants indicated picture-related subjective craving. In addition to analyzing classic ERPs, we employed single-trial multiple robust regression within subjects to relate EEG activity at each time point and electrode to trial-wise craving ratings. Craving was lower in the LATER and DISTRACT than the NOW condition. LPP amplitudes (300-1000 ms at centro-parietal electrodes) were lower in the DISTRACT but not the LATER compared to the NOW condition. Trial-wise craving was captured between 500 and 1100 ms and around 2400 ms at central midline and left sites for DISTRACT and associated with a very brief and local signal (~900 ms; at C3) for LATER. This replicates the mismatch between self-reported craving and LPP. Single-trial results suggest that the correlates of ROC are not entirely compatible with the focus on centro-parietal slow waves (i.e. the LPP) and that different strategies of ROC are characterized by distinct electrocortical markers.