To create neural representations of external stimuli, the brain performs a number of processing steps that transform its inputs. For fundamental attributes, such as stimulus contrast, this involves one or ore nonlinearities that are believed to optimise the neural code to represent features of the natural environment. Here we ask if the same is also true of more complex stimulus dimensions, such as emotional facial expression. We report the results of three experiments combining morphed facial stimuli with electrophysiological and psychophysical methods to measure the function mapping emotional expression intensity to internal response. The results converge on a nonlinearity that accelerates over weak expressions, and then becomes shallower for stronger expressions, similar to the situation for lower level stimulus properties. We further demonstrate that the nonlinearity is not attributable to the morphing procedure used in stimulus generation.
University of Lincoln, College of Social Science Research
Katie L.H. Gray, University of Reading, School of Psychology and Clinical Language Sciences
Tessa R. Flack, University of Lincoln, School of Psychology
Miaomiao Yu, University of York, Department of Psychology
Freya A. Lygo, University of York, Department of Psychology
Daniel H. Baker, University of York, Department of Psychology and York Biomedical Research Institute