<p>During social interactions, we perceive emotions from facial expressions and the accompanying body postures of individuals. When facial and body expressions display different emotions, an ambiguity arises, which makes the recognition of facial emotions more challenging than when they are similar. The dimensional model relates this difficulty to interference caused by the closeness of emotions on the valence and arousal dimensions. In contrast, the emotion seed model proposes that these confusions result from shared physical characteristics between emotions. However, observations that neither model fully explains call for further investigation. To better understand discrepancies between models’ predictions and unexplained observations, we conducted a systematic review including thirteen studies (N = 470 participants) to (1) describe dimensional and emotion seed models as an explanatory framework for the incongruence effect, (2) identify variations in methodologies and (3) propose research directions. Methodological details such as participants’ characteristics, study design, sets of facial and body emotions, measured emotions, actor characteristics, focus types, and outcome variables were extracted from included studies. Three types of attentional focus were used in the included studies: the focus on facial emotion, on bodily expression, or on the whole person. Our review confirms that neither of the two theoretical models fully explains the emotional recognition errors during incongruence between facial and bodily expressions. Moreover, it suggests that the attentional focus mediates the influence of bodily expressions on facial emotion recognition and vice versa. These findings highlight the importance of building new theoretical models considering attentional focus to better understand incongruence effects.</p>