The behavioral and electrophysiological study of the emotional intensity of facial expressions of emotions has relied on image processing techniques termed ‘morphing’ to generate realistic facial stimuli in which emotional intensity can be manipulated. This is achieved by blending neutral and emotional facial displays and treating the percent of morphing between the two stimuli as an objective measure of emotional intensity. Here we argue that the percentage of morphing between stimuli does not provide an objective measure of emotional intensity and present supporting evidence from affective ratings and neural (event-related potential) responses. We show that 50% morphs created from high or moderate arousal stimuli differ in subjective and neural responses in a sensible way: 50% morphs are perceived as having approximately half of the emotional intensity of the original stimuli, but if the original stimuli differed in emotional intensity to begin with, then so will the morphs. We suggest a re-examination of previous studies that used percentage of morphing as a measure of emotional intensity and highlight the value of more careful experimental control of emotional stimuli and inclusion of proper manipulation checks.
