- BACKGROUND: Cognitive Bias Modification (CBM) has garnered interest as a potential addiction treatment. CBM interventions such as Approach Avoidance Training (AAT) are designed to alter automatic tendencies to approach drugs or drug-related cues. In our previous work, the cannabis AAT (CAAT) reduced cannabis approach bias, which was related to reduced cannabis use, among 80 non-treatment-seeking cannabis-using youth (Jacobus et al., 2018). In this preliminary examination, a subsample of these youth underwent neuroimaging to explore CAAT's effect on cannabis cue-related neural activation. METHODS: Sub-study participants were 41 cannabis-using youth ages 17-21 (mean age = 18.83; 47.5% female). Participants completed a cannabis cue-reactivity task during a functional MRI scan pre- and post CAAT-training or CAAT-sham to examine CAAT-related neural changes. RESULTS: Thirty-seven youth completed all six CAAT (n = 19) or CAAT-sham (n = 18) training sessions and had usable neuroimaging data. The group*time interaction on cannabis approach bias reached trend-level significance (p = .055). Change in approach bias slopes from pre-to post-treatment was positive for CAAT-sham (increased approach bias) and negative for CAAT-training (change to avoidance bias), consistent with the larger study. No significant changes emerged for cannabis cue-induced activation following CAAT-training or CAAT-sham in whole brain or region of interest analyses. However, active CAAT-training was associated with small-to-medium decreases in amygdala (Cohen's dz = 0.36) and medial prefrontal cortex (Cohen's dz = 0.48) activation to cannabis cues. CONCLUSIONS: Despite reducing cannabis use in the larger sample, CAAT-training did not alter neural cannabis cue-reactivity in the sub-study compared to CAAT-sham. More research is needed to understand neural mechanisms underlying AAT-related changes in substance use.