Detecting Genuine and Deliberate Displays of Surprise in Static and Dynamic Faces

Abstract

People are good at recognizing emotions from facial expressions, but less accurate at determining the authenticity of such expressions. We investigated whether this depends upon the technique that senders use to produce deliberate expressions, and on decoders seeing these in a dynamic or static format. Senders were filmed as they experienced genuine surprise in response to a jack-in-the-box (Genuine). Other senders faked surprise with no preparation (Improvised) or after having first experienced genuine surprise themselves (Rehearsed). Decoders rated the genuineness and intensity of these expressions, and the confidence of their judgment. It was found that both expression type and presentation format impacted decoder perception and accurate discrimination. Genuine surprise achieved the highest ratings of genuineness, intensity, and judgmental confidence (dynamic only), and was fairly accurately discriminated from deliberate surprise expressions. In line with our predictions, Rehearsed expressions were perceived as more genuine (in dynamic presentation), whereas Improvised were seen as more intense (in static presentation). However, both were poorly discriminated as not being genuine. In general, dynamic stimuli improved authenticity discrimination accuracy and perceptual differences between expressions. While decoders could perceive subtle differences between different expressions (especially from dynamic displays), they were not adept at detecting if these were genuine or deliberate. We argue that senders are capable of producing genuine-looking expressions of surprise, enough to fool others as to their veracity.

Distinct temporal features of genuine and deliberate facial expressions of surprise

“…The speed of raising the eyebrow and the eyelid. The dynamic information in the facial displays serves to discriminate the perceptual differences between spontaneous and deliberate displays, as suggested in prior work 8,13,49 . To understand the role of dynamic information in discrimination and perception, we further investigate the features which differentiate genuine surprise from the different types of deliberate surprise.…”

“…To avoid misleading inferences regarding the genuine-deliberate facial display dimension, research on the encoding of facial displays should clarify how the production method employed to generate the expressions may create differences in appearance. Zloteanu et al 8,19 have already recorded four different deliberate surprise displays as well as spontaneous surprise displays: Spontaneous surprise display in response to a jack-in-the-box, which can be described as the unexpected and schema-discrepant eliciting event, was filmed (Genuine condition). This manipulation is justified as Ekman et al 36 suggest that the surprised expressions are produced by the unexpectedness of such an event.…”

“…The differences between deliberate displays elicited using different production methods also remain elusive. The current study, thus, focused on the core facial movement (i.e., eyebrow and eyelid movements) in surprise expressions under Genuine, Improvised, External, and Rehearsed conditions recorded in the two studies by Zloteanu et al 8,19 . In the current study, the momentto-moment x and y coordinates of eyelids and eyebrows in the 127 videos of surprised reactions were annotated with deep learning-based tracking software ('DeepLabCut').…”

“…More specifically, it was predicted that raising movements of the eyebrows and eyelids composed all surprise displays and that the eyelid movements preceded eyebrow movements in genuine surprise displays, consistent with Namba et al 10 . Furthermore, according to Zloteanu et al 8,19 , it was assumed that the different deliberate displays will also differ from each other but be overall more similar than with genuine displays; the exploratory nature of the study does not permit directional hypotheses for these differences.…”

“…An overview of the stimuli creation process is provided below. The full procedure is described in the previous studies 8,19 . The camera in all conditions was arranged at eye-level and recorded participants from the beginning of the startle reactions through the end of these responses.…”