Emotion Perception

The Heider-Simmel Task

Over seventy years ago a stop motion video was made in which two triangles and a circle move around for just under a minute and a half, on a flat surface with a box on it. Participants were asked to narrate the video as they watched it. Unprompted, the viewers talked fluidly about causal relationships, and spoke of the shapes as if they were animate organisms with intentions and emotions. Such experiences, mainstream psychologists declared, were illusory, and to be explained away: The video was made by a modeler, who moved the pieces little bits between each frame, with none of the causal relations the viewers described, never the less the animacy, the intentionality, nor the emotions present. Because of that view, this became seen as a firm example of "emotion attribution." Researchers now make similar displays fairly often, and several labs use similar tasks to try to better understand the neural underpinnings of emotional experience and the symptoms of those with emotion processing deficits. 

Despite the popularity of this task in assessing emotion attribution, there has been very little research determining exactly which display properties are crucial. That is, no one has identified the aspects of the displays that participants are responding to when they experience the shapes as having emotions. Presumably there are four types of variables that could be important in this process: 1) static features of the objects (size, shape, color, etc.), 2) dynamic features of the objects (speed, acceleration, erraticness of movements, etc.), 3) Relative features (smaller than social partner, faster than social partner, etc.), 4) Dynamic interaction features (stopping near a partner, mimicking a partner, moving straight towards a partner or curving around them, etc.)

In a series of experiments, using agent based modeling in the program NetLogo, we have been starting to investigate these different features. The work is a collaboration with Dr. Lee Rudolph, a topologist in the Math Department of Clark University. Preliminary results have been presented at The International Society for Developmental Psychobiology and the International Conference on Perception and Action. Much of the preliminary work served as an Honor's Thesis for Mark Singer. Also, bits of the theory and data have appeared in a few short papers.

More details soon.