The specificity of action knowledge in sensory and motor systems

Abstract

Neuroimaging studies have found that sensorimotor systems are engaged when participants observe actions or comprehend action language. However, most of these studies have asked the binary question of whether action concepts are embodied or not, rather than whether sensory and motor areas of the brain contain graded amounts of information during putative action simulations. To address this question, we used repetition suppression (RS) functional magnetic resonance imaging to determine if functionally-localized motor movement and visual motion regions-of-interest (ROI) and two anatomical ROIs (inferior frontal gyrus, IFG; left posterior middle temporal gyrus, pMTG) were sensitive to changes in the exemplar (e.g., two different people “kicking”) or representational format (e.g., photograph or schematic drawing of someone “kicking”) within pairs of action images. We also investigated whether concrete versus more symbolic depictions of actions (i.e., photographs or schematic drawings) yielded different patterns of activation throughout the brain. We found that during a conceptual task, sensory and motor systems represent actions at different levels of specificity. While the visual motion ROI did not exhibit RS to different exemplars of the same action or to the same action depicted by different formats, the motor movement ROI did. These effects are consistent with “person-specific” action simulations: if the motor system is recruited for action understanding, it does so by activating one's own motor program for an action. We also observed significant repetition enhancement within the IFG ROI to different exemplars or formats of the same action, a result that may indicate additional cognitive processing on these trials. Finally, we found that the recruitment of posterior brain regions by action concepts depends on the format of the input: left lateral occipital cortex and right supramarginal gyrus responded more strongly to symbolic depictions of actions than concrete ones.

Neural Representations of Observed Actions Generalize across Static and Dynamic Visual Input

“…To perceive categories of actions, one must have representations that distinguish between them (e.g., biting is different from pushing) yet show invariance to different instantiations of the same action. Although previous work has described a network of regions involved in coding observed actions (the "action observation network" or AON; Caspers et al, 2010; Rizzolatti and Sinigaglia, 2010;Kilner, 2011;Urgesi et al, 2014), the extent to which these regions abstract across differences between action exemplars is not well understood.Previous research has addressed the question of abstraction (i.e., invariance) in two ways. First, many neuroimaging and neuropsychological studies have explored generalization between observed and executed actions in an effort to resolve a debate over motor system involvement in action understanding (Chong et al., 2008; Dinstein et al, 2008;Kilner et al, 2009;Oosterhof et al, 2012a Oosterhof et al, , 2012b Tarhan et al, 2015;Tucciarelli et al, 2015; for review, see Rizzolatti and Sinigaglia, 2010;Oosterhof et al, 2013;.…”

“…Although previous work has described a network of regions involved in coding observed actions (the "action observation network" or AON; Caspers et al, 2010; Rizzolatti and Sinigaglia, 2010;Kilner, 2011;Urgesi et al, 2014), the extent to which these regions abstract across differences between action exemplars is not well understood.…”

“…Second, these representations should be elicited both by dynamic visual sequences, in which the entire action is observed, and static snapshots, from which the causal sequence must be inferred (Hafri et al, 2013). Several studies have found action-specific representations using static images (Ogawa and Inui, 2011;Watson et al, 2014) but, crucially, none has demonstrated common representations across dynamic and static input. Beyond testing these invariances, we also wished to examine actions performed with a wide variety of effectors (e.g., foot, mouth), not just hand/arm actions that are commonly investigated in the literature (see also Kable and Chatterjee, 2006;Jastorff et al, 2010;Watson et al, 2014).…”

“…Several studies have found action-specific representations using static images (Ogawa and Inui, 2011;Watson et al, 2014) but, crucially, none has demonstrated common representations across dynamic and static input. Beyond testing these invariances, we also wished to examine actions performed with a wide variety of effectors (e.g., foot, mouth), not just hand/arm actions that are commonly investigated in the literature (see also Kable and Chatterjee, 2006;Jastorff et al, 2010;Watson et al, 2014).To these ends, we used multivoxel pattern analysis (MVPA) of fMRI data to identify regions supporting abstract action representations. We scanned subjects while they viewed eight action categories in two visual formats ( Fig.…”

“…First, many neuroimaging and neuropsychological studies have explored generalization between observed and executed actions in an effort to resolve a debate over motor system involvement in action understanding (Chong et al., 2008; Dinstein et al, 2008;Kilner et al, 2009;Oosterhof et al, 2012a Oosterhof et al, , 2012b Tarhan et al, 2015;Tucciarelli et al, 2015; for review, see Rizzolatti and Sinigaglia, 2010;Oosterhof et al, 2013;. Second, other studies have examined invariance to different perceptual instantiations of observed actions (Kable and Chatterjee, 2006;Oosterhof et al, 2012a;Watson et al, 2014;Tucciarelli et al, 2015). In an especially direct test of such invariance, Wurm and Lingnau (2015) found that representations in several AON regions distinguished between opening Received Aug. 5, 2016; revised Jan. 16, 2017; accepted Feb. 3, 2017 …”