Time will tell: Object categorization and emotional engagement during processing of degraded natural scenes

The aim of the present study was to examine the relationship between object categorization in natural scenes and the engagement of cortico-limbic appetitive and defensive systems (emotional engagement) by manipulating both the bottom-up information and the top-down context. Concerning the bottom-up information, we manipulated the computational load by scrambling the phase of the spatial frequency spectrum, and asked participants to classify natural scenes as containing an animal or a person. The role of the top-down context was assessed by comparing an incremental condition, in which pictures were progressively revealed, to a condition in which no probabilistic relationship existed between each stimulus and the following one. In two experiments, the categorization and response to emotional and neutral scenes were similarly modulated by the computational load. The Late Positive Potential (LPP) was affected by the emotional content of the scenes, and by categorization accuracy. When the phase of the spatial frequency spectrum was scrambled by a large amount (>58%), chance categorization resulted, and affective LPP modulation was eliminated. With less degraded scenes, categorization accuracy was higher (.82 in Experiment 1, .86 in Experiment 2) and affective modulation of the LPP was observed at a late window (>800 ms), indicating that it is possible to delay the time of engagement of the motivational systems which are responsible for the LPP affective modulation. The present data strongly support the view that semantic analysis of visual scenes, operationalized here as object categorization, is a necessary condition for emotional engagement at the electrocortical level (LPP).