Opposing roles for amygdala and vmPFC in the return of appetitive conditioned responses in humans

Learning accounts of addiction and obesity emphasize the persistent power of Pavlovian reward cues to trigger craving and increase relapse risk. While extinction can reduce conditioned responding, Pavlovian relapse phenomena—the return of conditioned responding following successful extinction—challenge the long-term success of extinction-based treatments. Translational laboratory models of Pavlovian relapse could therefore represent a valuable tool to investigate the mechanisms mediating relapse, although so far human research has mostly focused on return of fear phenomena. To this end we developed an appetitive conditioning paradigm with liquid food rewards in combination with a 3-day design to investigate the return of appetitive Pavlovian responses and the involved neural structures in healthy subjects. Pavlovian conditioning (day 1) was assessed in 62 participants, and a subsample (n = 33) further completed extinction (day 2) and a reinstatement test (day 3). Conditioned responding was assessed on explicit (pleasantness ratings) and implicit measures (reaction time, skin conductance, heart rate, startle response) and reinstatement effects were further evaluated using fMRI. We observed a return of conditioned responding during the reinstatement test, evident by enhanced skin conductance responses, accompanied by enhanced BOLD responses in the amygdala. On an individual level, psychophysiological reinstatement intensity was significantly anticorrelated with ventromedial prefrontal cortex (vmPFC) activation, and marginally anticorrelated with enhanced amygdala-vmPFC connectivity during late reinstatement. Our results extend evidence from return of fear phenomena to the appetitive domain, and highlight the role of the vmPFC and its functional connection with the amygdala in regulating appetitive Pavlovian relapse.

intervals (PP intervals on day 3, respectively) was visually inspected and manually corrected, if necessary. Sequences with artefacts or low signal in the data preventing reliable heart beat detection were treated as missing data points. The time series of non-uniform inter-beat intervals was converted to HR and interpolated to the sample rate of acquisition. Trials with missing data in a window from -1 to 4 s with respect to CS onset were excluded from further analyses. Mean HR was calculated for the time window 1-3 s after cue onset.
Trialwise HR data were normalized and aggregated over each phase and cue type. In the conditioning sample technical failures during recording caused data loss in three subjects, while for day 3 another five subjects were excluded due to failed PP detection (low signal-tonoise ratio/ PPG dislocation). Because swallowing causes prolonged HR changes after US delivery 2 , only trials without preceding reinforcement were considered for analysis of HR during conditioning (day 1).
Reaction Time. RTs from the cue side detection task were collected in the laboratory and fMRI with a 2-button keypad and MR-compatible response buttons, respectively. RTs between 200-2400 ms were considered valid responses. Data were log transformed to reduce skewness, and averaged over each phase and cue type. Data from one subject are missing due to technical malfunction.  Table S1: Exploratory whole-brain results during reinstatement test, displayed at p<.001 uncorrected using a cluster-forming threshold of k=10 contiguous voxels

Exploratory connectivity analysis using the right NAcc as seed region
Following a reviewer's suggestion, we further explored functional cue-dependent connectivity between right NAcc, which showed a time-dependent effect during the reinstatement test with stronger BOLD response towards CS+ compared to CS-during the early compared to the late reinstatement phase, and the vmPFC by applying a similar gPPI analysis as for the amygdala but using the right NAcc as seed region. This analysis revealed no significant cuedependent NAcc-vmPFC involvement across or within phases of the reinstatement test (p FWE ROI ≥.648). Interestingly, we instead observed heightened functional connectivity between right NAcc and amygdala ([x:23, y:0, z:-26]; Z=3.40; p FWE ROI =.035) during the early reinstatement phase, suggesting these two structures to closely interact upon CS+ compared to CSpresentation.

Effects of contingency awareness on conditioning measures
Following a worthwhile reviewer comment, we explored possible associations between contingency awareness and measured indices of conditioning on day 1. As our study was not designed to unambiguously classify participants as contingency aware or unaware, we based our analyses on a median split on difference scores of rated reward probabilities (CS+ minus CS-) acquired after conditioning (see Figure S2A). We then re-evaluated each conditioning measure by introducing a between subject group factor (aware vs. unaware), i.e. CS pleasantness ratings, SCRs, RTs, and HR were analyzed in separate mixed ANOVAs with within subject factors cue type (CS+ vs. CS-) and time (pre/early vs. post/late) and between subject factor group (aware vs. unaware), and startle responses were analyzed in a mixed 4 ANOVA with within subject factor cue type and between subject factor group. Significant interactions with contingency awareness were followed up by groupwise post-hoc analyses.
RTs. In line with our main analysis, no significant main or interaction effects were observed in