Are paired or single stimuli better to recognize genuine and posed smiles from observers' galvanic skin response?