Niedenthal et al. diverge from an increasing trend for localizationalism in cognitive neuroscience, and present a model of emotional expression and understanding that spans the somatosensory system, motor system, reward system, amydgdala, basal ganglia, and prefrontal cortex, as well as systems involved in eye gaze and body posture. This model enfolds ideas relating to mirror neurons and the link between perception and production, but as a subset of a larger system for understanding emotions in others.

Not only is this model of emotion unlocalized in space, it is also not localized in time. The systems involved must communicate and pass activation through time in order to fully engage this emotion system, akin to a closed-loop dynamic process. It may be crucial that the timing of the interactions of these structures be right. Disorders of temporal processing at any level of this system, whether in the perception or the production of smiles, or in passing activation between brain structures, could disrupt this system and hence disrupt emotion understanding.

Consideration of the temporal dimension of such a model may lead to a better understanding of social deficits, such as those in autism spectrum disorders as well as those reported in association with attention deficit disorder. Indeed, Gepner and Feron (Reference Gepner and Feron2009) describe a theory of temporal processing deficits that may underlie a range of deficits in autism. Is this theory at odds with the mirror neuron hypothesis (e.g., Williams et al. Reference Williams, Whitenb, Suddendorfc and Perrett2001), that a disruption of the motor neuron response to the perception of movement in others underlies social processing dysfunction in autism? Probably not. Both may be components of a larger emotion understanding system that involves multiple structures and their interactions in time. In support of this perspective, Oberman et al. (Reference Oberman, Winkielman and Ramachandran2009) showed that individuals with autism do indeed show spontaneous mimicry of facial expressions, but that the response is delayed.

Such models inform the development of interventions to help people with social dysfunction. The Niedenthal et al. model, together with theories such as the one proposed by Gepner and Feron (Reference Gepner and Feron2009), suggests that social processing interventions should tap multiple processes, not individually, but together and at the right temporal intervals. Recent technology for automatically recognizing and responding to facial expression, head pose, and eye gaze in real-time opens up new possibilities for intervention systems that link perception and production on timescales related to social responding (Bartlett & Whitehill, in press; Cockburn et al. Reference Cockburn, Bartlett, Tanaka, Movellan, Pierce and Schultz2008). Such technology contributes not only to clinical research, but also to the study of learning and plasticity in perception and production systems, and to understanding the cognitive neuroscience of emotion.