“…With this space-by-time manifold decomposition, we address important questions in emotion communication using facial expressions. In particular, we determine which synergistic facial movements communicate emotions, and we isolate the specific movements that cause confusions between specific emotion categories (e.g., fear and surprise; see Ekman, 1992 ; Gagnon, Gosselin, Hudon-ven Der Buhs, Larocque, & Milliard, 2010 ; Jack, Blais, Scheepers, Schyns, & Caldara, 2009 ; Matsumoto & Ekman, 1989 ; Moriguchi et al, 2005 ; Roy-Charland, Perron, Beaudry, & Eady, 2014 ) from those that resolve these confusions. That is, we can identify the low-dimensional space-by-time representation that predicts categorization behavior.…”