Introduction: Positively conditioned Pavlovian cues tend to promote approach and negative cues promote withdrawal in a Pavlovian-to-instrumental transfer (PIT) paradigm, and the strength of this PIT effect was associated with the subsequent relapse risk in alcohol-dependent (AD) patients. When investigating the effect of alcohol-related background cues, instrumental approach behavior was inhibited in subsequent abstainers but not relapsers. An automatic approach bias towards alcohol can be modified using a cognitive bias modification (CBM) intervention, which has previously been shown to reduce the relapse risk in AD patients. Here we examined the effects of such CBM training on PIT effects and explored its effect on the relapse risk in detoxified AD patients. Methods: N = 81 recently detoxified AD patients performed non-drug-related and drug-related PIT tasks before and after CBM versus placebo training. In addition, an alcohol approach/avoidance task (aAAT) was performed before and after the training to assess the alcohol approach bias. Patients were followed up for 6 months. Results: A stronger alcohol approach bias as well as a stronger non-drug-related PIT effect predicted relapse status in AD patients. No significant difference regarding relapse status or the number of heavy drinking days was found when comparing the CBM training group versus the placebo group. Moreover, there was no significant modulation effect of CBM training on any PIT effect or the aAAT. Conclusion: A higher alcohol approach bias in the aAAT and a stronger non-drug-related PIT effect both predicted relapse in AD patients, while treatment outcome was not associated with the drug-related PIT effect. Unlike expected, CBM training did not significantly interact with the non-drug-related or the drug-related PIT effects or the alcohol approach bias.
