Table 1.
Glossary: Embodiment Terms and Their Definitions
| Term | Definition | Citation |
|---|---|---|
| Active inference | Inferring the state of the environment, ie, perception, is not a passive response to sensory stimuli but an active process. Animals perceive by making predictions based on internal models (ie, “hypotheses” or “expectations”) and resolving the resulting prediction error. This resolution can be through changing predictions or changing the sensations that are sampled. Inference is thus embodied: an active process of inferring the causes of sensations, which are selected by action. | Hobson and Friston22 |
| Bayesian belief updating | The working mechanism of ↑ active inference: while executing a motor plan, prior beliefs and incoming sensory data—solicited by executing the plan—are continuously combined. In case of discrepancies, the actions (or predictions) may be adapted in order to minimize discrepancies (eg, prediction error). | Pezzulo et al30 |
| Bayesian statistics | An approach to inference that enables one to evaluate the probability of a hypothesis H, given some evidence D. The posterior probability that H is true, given data D, depends on the prior probability, ie, the probability of H before we see D, and on the likelihood, ie, the probability of observing D given that H is true. In ↑ Bayesian belief updating, posteriors can be used as priors—for new data—to accumulate evidence for one’s hypotheses. | |
| Embodiment | A concept emphasizing the reciprocal (bidirectional) relationship between mind (cognition, emotion) and body (motor behavior, nonverbal behavior, physiological processes). | Tschacher and Bergomi33 |
| Enactivism | Enactivism claims that cognition and perception result from an active engagement with environmental constraints (“affordances”), rather than from the representation of the environment—a position closely related to ↑ active inference. The enactive approach views mind, body, and environment as highly interdependent elements of an (ecological) system. | Thompson and Varela9 |
| Interoceptive inference | The notion of ↑ active inference can be generalized to interoception, ie, to perception and regulation of inner (physiological and visceral) processes. Interoceptive inference describes the (homeostatic) regulation of inner (autonomic) processes in the context of emotion and self-awareness. | Seth27 and Gu and FitzGerald28 |
| Mirror neuron system | Neuronal networks in the brain often have multiple functions: the same neurons may be active when executing an action and when perceiving the same action of another. Thus, action and action observation recruit the same neuronal (active inference) processes. This may be a neuronal correlate of empathy and mentalizing. | Rizzolatti and Craighero17 |
| Nonverbal synchrony | In the interaction of A and B, nonverbal behavior of A tends to resonate with the nonverbal behavior of B, commonly without the intention to imitate or mimic. Synchrony (of motor behavior, physiological arousal, prosody) signals sensorimotor coupling—and is a bodily expression of prosocial emotions. | Chartrand and Bargh34 |
| Optimal motor control | A formulation of motor control that rests on forward (and inverse) models of sensorimotor coupling. This usually involves generating an efference copy that is used to predict the sensory consequences of action. Discrepancies between the predicted and proprioceptive input are then used to adapt and fine-tune execution. | Wolpert et al.19 |
| Predictive coding | Predictive coding is a ↑ Bayesian belief updating scheme for estimating the causes of (sensory) data by minimizing prediction error. It is a popular metaphor for (neuronal) message passing in the brain. This recurrent message passing takes place at different hierarchical levels of the cortex; where top-down processes generate predictions, and bottom-up signals report prediction errors. | Friston20 |