Emotional categorization of objects: A novel clustering approach and the effect of depression

Most everyday actions engender interactions with meaningful emotionally-laden stimuli. This study aimed to select pictures of objects as emotional stimulus of affordance to be grasped. The participant's depression trait was also assessed to examine its effect on the judgment of these pictures, and time spent in the classification was computed. Sixty-three participants joined this study. Self-Assessment-Manikin scale was used to classify pictures of the objects, and Beck Depression Inventory was applied to distribute the sample according depression trait. Cluster analysis was used in the classification of 123 objects based on valence and arousal values. Cluster results returned 102 classified pictures in three categories: pleasant (21), neutral (48) and unpleasant (33). Where cluster analysis did not agree, the picture was excluded and not used any further (21). Pleasant pictures presented the highest valence values and unpleasant pictures the lowest, and both categories returned the highest arousal level. In the middle of the valence range, the neutral category evoked the lowest arousal levels. Participants were slower to classify unpleasant pictures in valence sub-scale and faster to classify neutral pictures in arousal one. There was no effect of depression in the response time needed to score the pictures. Thus, agreement of high-performance soft clustering algorithms emerged as a good tool to classify pictures representing objects based on valence and arousal dimensions. Depression trait does not significantly affect the accuracy or time-order of emotional classification. Finally, we presented a set of emotional stimuli that can be employed to examine distinct aspects of emotion over physiology and behavior.