Individual differences in bottom-up and top-down emotion generation.

Research has demonstrated there are two primary ways emotions can be generated. These include emotions generated from perception of simple physical properties of stimuli in the environment, known as "bottom-up" generated emotions, and emotions generated from cognitive appraisals of a cue or situation, known as "top-down" generated emotions. Experimentally, it has been shown these two forms of emotion generation recruit distinct neural and psychological processes. However, the extent to which individuals differ in how their emotions are generated, and how this variation may relate to other relevant constructs, has yet to be determined. Previous research has used self-report measures to assess individual differences in emotional responding. However, no measures to date have investigated individual differences in the degree to which people respond to bottom-up and top-down generated emotions specifically. A novel self-report questionnaire was developed to assess individual differences in bottom-up and top-down generated emotions. We report on three independent studies that evaluate the measures' psychometric properties and exploratory factor structure (Study 1; N = 149), confirmatory factor structure, associations of the measure with relevant psychopathology symptom domains, relevant established measures of emotional processes, and an established experimental task of emotion regulation (Studies 2 N = 230 and 3 N = 238). Factor analyses consistently indicated a two-factor structure for the items (Comparative Fit Index range = .94-.96) with adequate reliabilities. Correlation analyses revealed both top-down and bottom-up generated emotions were positively associated with internalizing symptoms and other measures relevant to emotion processing (significant r ranges:.15-.55). The relationships with extant measures indicated that we developed a measure that characterizes the use of cognition in the generation of negative emotion, as opposed to the use of cognition to control emotion. Future work should examine how this measure may be useful for affective and clinical science, including mechanistic treatment targets.