BACKGROUND: Generalized anxiety disorder (GAD) is one of the most prevalent anxiety disorders but the least successfully treated. The search for accessible clinical, psychological and biological markers is crucial for developing more effective and personalized interventions. AIMS: To evaluate if changes in heart rate variability (HRV) between rest and stress conditions before interventions could predict improvement in emotional interference (EI) in a cognitive task after three different treatment modalities in patients with GAD. METHOD: This is a post-hoc analysis study reporting data from a larger randomized controlled trial (NCT03072264) assessing a mindfulness-based intervention (BMT), fluoxetine (FLX), and an active comparison group (QoL) in adult patients diagnosed with GAD. We assessed pulse plethysmography (PPG) data using a Shimmer3 GSR to measure HRV. Regression analyses were performed using the variation between baseline and endpoint EI scores as dependent variables and contrasts considering changing in HRV*group interaction in the baseline. RESULTS: 106 individuals were included. The correlations between HRV changing from rest to task predicted improvement in IE only in the FLX versus control group contrast (estimated = -80.24; SE = 27.31; p = 0.005) and not in the BMT and control group contrast. CONCLUSION: More flexible HRV at baseline predicted EI improvement only in the FLX group. This finding is clinically relevant since it may help us develop more personalized interventions for GAD.