Depressed persons have demonstrated emotion based cognitive biases, specifically surrounding vigilance of negative information and avoidance of positivity. These biases are sometimes operationalized through emotion recognition tasks. However, previous emotion recognition tasks lack in their ability to accurately measure and decompose positivity avoidance with enhanced negativity, while accounting for basic cognitive processes that can drive the results. Therefore, we developed a novel emotion recognition task that examines emotional intensity thresholds, while accounting for general response bias. Linear mixed effects modeling revealed substantial individual differences on all conditions in the task, using both frequentist and Bayesian approaches. Additionally, the findings suggest that those with higher depression scores exhibit a greater cognitive response bias on emotion recognition tasks. Ultimately, this study provides evidence that there is variability in performance on the morphing paradigm, as further researcher is needed to assess the influence general response biases have on emotion recognition performance in depression.