Angelika Lingnau

The recognition of actions presented as static photographs has been suggested to be modulated by the scene context—or image background —in which an action is presented, particularly for actions that are hard to recognize. This suggests scene context as an important feature when retrieving action representations, and poses the question whether scene context directly modulates the representational space of actions. Therefore, in this preregistered virtual reality (VR) study, we investigated in a group of N = 24 healthy participants how scene context by means of the (virtual) environment affects the representation of actions that do or do not correspond to the current environment. Participants performed multi-arrangement tasks on a set of kitchen and workshop actions while the virtual environment was manipulated to represent either a kitchen, workshop or neutral environment. Representational structure was estimated from the multi-arrangement tasks separately for each virtual environment and action domain (kitchen, workshop). Results showed increased consistency of the representation of kitchen actions over participants when the task was performed in a congruent (i.e., kitchen) compared with an incongruent (i.e., workshop) or neutral virtual environment, however, the same effect was not observed for workshop actions. Representational structure, however, was not affected by the environment. In accordance with previous studies on representational changes in domain-specific experts within various action and object domains the results suggest that an agent’s environment may modulate the action space for corresponding actions by increasing robustness of represented familiar actions. These findings contribute to our understanding of contextual effects on mental representations within the domain of actions and beyond.

Everyday action perception unfolds in complex contexts, requiring coordinated processing from low-level visual information to higher-level inferences about action types and social interactions. Aging is accompanied by sensory and cognitive changes, raising the question of how these processes may be altered. Here, we combined action similarity judgments with analyses of time-resolved and source-localized brain activity measured with magnetoencephalography (MEG) to investigate age differences in action perception when younger and older adults viewed naturalistic actions. Using cross-validated variance partitioning within a multivariate representational framework, three key findings were uncovered. First, affective-communicative features predicted similarity judgments better than, and independent of, visual and action-related features in both age groups. Second, neural processes in both groups followed a hierarchy from visual to action-related and affective-communicative features and recruited similar functional regions. However, older adults showed delayed visual processing but earlier affective-communicative processing, indicating selective prioritization rather than uniform slowing of subprocesses. Third, when sensory uncertainty was increased through visual blurring, older adults showed hierarchical reorganization, with affective-communicative features emerging earlier than action-related features, whereas younger adults preserved the original hierarchy. Together, these findings unveil adapted information prioritization in healthy aging, whereby socially meaningful information increasingly guides action perception when sensory input becomes unreliable.

Action recognition relies on efficiently prioritizing relevant sources of information. This study examined the degree to which category-based attention facilitates the identification of basic action categories (climb, jump, run, walk) depicted as static avatar images (Exp. 1, 2) and naturalistic images (Exp. 3, 4). Participants received prior information about the target action category through a word cue (Exp. 1, 2, 4) or an avatar image cue (Exp. 3), with a 3:1 ratio of valid to invalid trials. Performance was assessed in terms of accuracy, reaction time (RT), and sensitivity (d-prime). Exp. 1 revealed a cueing effect of word cues on processing subsequent avatar stimuli that depicted different actions. Exp. 2 confirmed that this facilitation was due to the cue itself rather than specific motor associations. Exp. 3 indicated that avatar cues aid action recognition in naturalistic images, though the effect was weaker and limited to accuracy. Exp. 4 combined word cues with the same naturalistic images and revealed robust cueing effects on both accuracy and RT, indicating that top-down facilitation is preserved for naturalistic stimuli when word cues are used. Together, these findings provide an important extension of previous studies, suggesting that category-based attention enhances action recognition by pre-activating relevant action representations.

A growing body of theoretical and experimental accounts suggest that observed actions are represented in multidimensional representational spaces. So far there is little evidence how these representations are shaped by experience. Here we investigate whether and how action spaces are modulated by long-term experience with particular action domains by comparing action spaces of gymnast actions in action experts, i.e. gymnasts, and non-experts. The similarity structure of gymnast actions was characterized in N = 40 experts and non-experts using multi-arrangement tasks on short videos of gymnast actions. Within-group leave-one-out correlations suggested that experts’ similarity structures are more consistent than those of non-experts. Furthermore, hierarchical clustering and representational similarity analyses suggest that, whereas non-experts are more strongly driven by visually salient features of actions, experts’ similarity ratings are based more strongly on functionally and possibly conceptually relevant movement features of the same actions. These results provide first evidence that action spaces are not static entities but subject to long-term experience with action domains.

In the current issue of Cerebral Cortex, Smekal et al. (2026) argue that the extrastriate body area (EBA) is not only involved in the visual perception of the human body, but that it is a multisensory and multifunctional region that supports the processing of body-related information “across sensory, conceptual and linguistic domains”. In this commentary, I argue that a better understanding of the role of the EBA profits from a network perspective that takes different subdivisions of the EBA into account.

Writing systems are cultural inventions that differ in how they represent spoken language. We tested how the brain learns to map arbitrary visual symbols to sound and meaning by comparing neural activity for artificial writing systems that were either alphabetic (systematic symbol-sound mappings) or logographic (arbitrary symbol-sound mappings). Twenty-four adults learned to read aloud and comprehend novel words written in each system. After 2 weeks of training, functional magnetic resonance imaging during reading comprehension revealed that the dorsal pathway (inferior parietal gyrus, inferior frontal gyrus), which is involved in mapping from print-to-sound, was more active for the alphabetic system, whereas the ventral pathway (anterior fusiform gyrus, middle temporal gyrus), which is involved in mapping from print-to-meaning, was more active for the logographic system. Combined with performance differences, these findings indicate that systematic symbol-sound mappings allowed the brain to bridge the interface between vision and meaning via sound, whereas an absence of systematicity made it more efficient to link vision directly to meaning. Thus, the same brain finds different solutions to the problem of reading that capitalise on the statistical properties of culturally invented writing systems.

The recognition of actions presented as static photographs has been suggested to be modulated by the scene context - or image background - in which an action is presented, particularly for actions that are hard to recognize. This suggests scene context as an important feature when retrieving action representations, and poses the question whether scene context directly modulates the representational space of actions. Therefore, in this preregistered virtual reality (VR) study, we investigated in a group of N = 24 healthy participants how scene context by means of the (virtual) environment affects the representation of actions that do or do not correspond to the current environment. Participants performed multi-arrangement tasks on a set of kitchen and workshop actions while the virtual environment was manipulated to represent either a kitchen, workshop or neutral environment. Representational structure was estimated from the multi-arrangement tasks separately for each virtual environment and action domain (kitchen, workshop). Results showed increased consistency of the representation of kitchen actions over participants when the task was performed in a congruent (i.e. kitchen) compared to an incongruent (i.e. workshop) or neutral virtual environment, however, the same effect was not observed for workshop actions. Representational structure, however, was not affected by the environment. In accordance with previous studies on representational changes in domain-specific experts within various action- and object-domains the results suggest that an agent’s environment may modulate the action space for corresponding actions by increasing robustness of represented familiar actions. These findings contribute to our understanding of contextual effects on mental representations within the domain of actions and beyond.

The lateral occipitotemporal cortex (LOTC) has been suggested to host action representations that are thought to contribute to accessing meaning of observed actions. In line with this proposal, previous studies have shown spatially overlapping representations of various action-related features including objects, scenes, social properties, and kinematics, as well as abstract representations pertaining to action semantics. Less is known, however, about the way in which these features are integrated during action recognition. To address this question, we aimed to examine the temporal order in which different action-related features emerge in the EEG signal using EEG-based representational similarity analysis (RSA). Additionally, to investigate the spatiotemporal characteristics of these neural representations, we employed fMRI-EEG fusion. Static images spanning 27 everyday actions (e.g., riding a bike, washing dishes, brushing hair) were shown to participants (N = 24) in a delayed matching task, in a total of 648 trials (using different combinations of locations, actors, and viewpoints). Participants were asked to judge stimuli with respect to the type of action, location, and actor identity. Temporally specific neural representational dissimilarity matrices (RDMs) pertaining to neural representations of action categories were generated from 64-channel EEG recordings using time-resolved RSA. These were correlated with behavioral RDMs pertaining to semantic action (dis)similarities, revealing a time course of neural semantic action representations. Additionally, we considered RDMs capturing (dis)similarities between actions in terms of involved body postures, scenes, social aspects, and lower-level visual information. Results suggest a temporally ordered hierarchical buildup of neural representations related to visual, contextual, body-related, and semantic action information. fMRI-EEG fusion analysis further suggests that striate and extrastriate areas along the lateral visual pathway encode lower-level visual and body-related properties of actions, and that contextual and semantic information is integrated in the LOTC. These outcomes provide a spatiotemporal characterization of the neural processes enabling humans to recognize goal-directed actions.

How do we understand actions performed by others? Recent studies suggest that actions can be represented as points in a multidimensional space, where the perceived similarity of two actions is thought to be related to their proximity within this space. Here we present a data-driven approach to reveal key dimensions underlying this space using a carefully selected stimulus database of 768 one-second video clips spanning 256 action categories. We gathered similarity ratings for these videos from 6,036 participants and used a computational modeling procedure to identify key dimensions underlying these ratings. This approach revealed 28 meaningful dimensions (e.g. interaction, sport and craft) which capture information concerning human actions as well as a broad range of related domains (e.g. living and non-living things). Explicit ratings of actions along these dimensions gathered in a separate group of participants revealed a high correlation between ratings and weights along these dimensions, demonstrating that these dimensions are interpretable and can be used by participants. The multidimensional action space established in the current study enables the quantification of the similarity between different actions, which will be useful for the generation of hypotheses and future experimental manipulations. Together, our results provide a window into the nature of the representations underlying the ability to interpret other people’s actions and pave the way for future lines of research.

The lateral occipitotemporal cortex (LOTC) has been suggested to host action representations that are thought to contribute to accessing meaning of observed actions. In line with this proposal, previous studies have shown spatially overlapping representations of various action-related features including objects, scenes, social properties and kinematics, as well as abstract representations pertaining to action semantics. Less is known, however, about the way in which these features are integrated during action recognition. To address this question, we aimed to examine the temporal order in which different action-related features emerge in the EEG signal using EEG-based representational similarity analysis (RSA). Additionally, to investigate the spatiotemporal characteristics of these neural representations, we employed fMRI-EEG fusion. Static images spanning 27 everyday actions (e.g. riding a bike, washing dishes, brushing hair) were shown to participants (N=24) in a delayed matching task, in a total of 648 trials (using different combinations of locations, actors, and viewpoints). Participants were asked to judge stimuli with respect to the type of action, location, and actor identity. Temporally specific neural representational dissimilarity matrices (RDMs) pertaining to neural representations of action categories were generated from 64-channel EEG recordings using time-resolved RSA. These were correlated with behavioral RDMs pertaining to semantic action (dis)similarities, revealing a time course of neural semantic action representations. Additionally, we considered RDMs capturing (dis)similarities between actions in terms of involved body postures, scenes, social aspects and lower-level visual information. Results suggest a temporally ordered hierarchical buildup of neural representations related to visual, contextual, body-related, and semantic action information. fMRI-EEG fusion analysis further suggests that striate and extrastriate areas along the lateral visual pathway encode lower-level visual and body-related properties of actions, and that contextual and semantic information is integrated in the LOTC. These outcomes provide a spatiotemporal characterization of the neural processes enabling humans to recognize goal-directed actions.