VSS, May 13-18
Talk 1, 8:15 am, 41.21
Neural underpinnings of biological motion perception under attentional load
Biological motion (BM) perception under selective attention is supported by a network of regions including the occipito-temporal, parietal, and premotor cortices. Considering the effortless processing and social importance of BM for humans, it can be assumed to occur automatically. Consistently, Thornton and Vuong (2004) showed that task-irrelevant BM stimuli at the periphery impaired the performance at the fovea; indicating incidental processing of BM stimulus. Another study examining retinotopic mapping of BM stimuli showed distinct maps for BM processing under and away from attention (Saygin & Sereno, 2008). However, the neural underpinnings of how BM is processed when attention is directed away from it lacks thorough examination. In the current study, we investigated whether the BM would be processed when shown away from the selective attention through an attentional load paradigm. Participants (N=31) underwent an fMRI study in which they performed an attentionally demanding visual detection task at the fovea while intact or scrambled point-light displays of BM were shown at the periphery. Our results showed the main effect of attentional load in fronto-parietal regions; whereas, the main effect of distractor was present in occipito-temporal regions. More importantly, in the presence of task-irrelevant peripheral BM stimuli, the motion-sensitive areas in the occipito-temporal cortex were activated even when participants attended at the fovea. Furthermore, MVPA results showed the decrease in the number of regions that could decode BM under high compared to low attentional load, indicating the attentional load modulation on BM processing. ROI results have expanded this finding by showing that the attentional load modulation of BM were the strongest in occipito-temporal regions compared to parietal and premotor cortices in both left and right hemispheres. In conclusion, BM was found to be processed within the occipito-temporal cortex when shown away from the selective attention; and was modulated by attentional load.
Talk 2, 8:30 am, 41.22
Social action understanding after late sight recovery from congenital near-blindness
Understanding actions performed by others and interpreting their emotional context is commonplace in our daily life: We readily assess complex social settings from minimal visual cues. Theories of action understanding typically assume extensive experience of action observation in early life, for its assimilation. But what happens if pattern vision in these years is extremely limited. We studied Ethiopian children, born with dense bilateral cataracts, who were surgically treated only years later. These patients have relatively poor visual acuity, even after surgery, and typically have difficulty in interpreting facial gestures. We tested whether other visual cues, such as body-configuration and motion signals, which are relatively preserved despite image blur, allow them to understand social situations. Seven newly-sighted patients viewed videos or still images of human interactions, categorizing them as "friendly" or "aggressive". The conditions were (1) animations ("Full-Body" condition); (2) point-light displays (containing only motion information; "PLD"), and (3) a static snapshot from the animation (preserving only configural information, "ST"). The patients performed worse than normally-developing subjects in all conditions. However, they performed significantly above chance in the Full-Body condition (79.8% correct, p=0.001), as well as in the impoverished conditions (67.9% correct, p=0.023 and 66.7% correct, p=0.001 in the ST and PLD conditions, respectively). In another test, the patients were asked to categorize images of people as "scared" or "angry" based on their facial expressions. Facial expression and body posture were manipulated orthogonally. Unlike controls, patients’ response was solely affected by the body posture (62.1% correct, p=0.015), and not by the facial expression (47.1% correct, p=0.203). We conclude that interpreting social situations can be acquired despite prolonged early-onset visual deprivation, by using biological motion or configural-body cues. Future experiments will clarify whether this capability had developed despite near-blindness conditions before surgery, or was acquired through visual experience following pattern vision recovery.
Acknowledgements: Supported by the DFG German-Israeli Project Cooperation grant #Z0 349/1
Talk 3, 8:45 am, 41.23
Characterization of visual response properties and connectivity of Wide-field vertical neurons in the mouse superior colliculus
Vision is a sense that serves different ethological functions across species. Visual information can be used to either detect preys or predators and generate orienting or escaping responses. The superior colliculus has been implicated in the orchestration of these behaviors. Within this structure, at the interface between the visual and the multisensory layers, lay the wide-field vertical neurons (WFV). WFV are a conserved morphologically defined cell-type found across many taxa from reptiles to highly visual mammals, independently of their visual ecology. In mammals, these neurons receive direct inputs from the retina and the primary visual cortex, and have the pulvinar as a known target, however, their brain-wide connectivity remains to be identified. These cells display peculiar response properties, but despite their ubiquity and unique morphology, their contribution to vision remains elusive. In this study, we have characterized WFV morphology, connection, and visual response properties in vivo in the laboratory mouse. Viral intersectional approaches revealed their projection patterns. Two-photon calcium imaging in awake mice using a transgenic Cre-line targeting specifically WFV (NTSR1-Cre) confirmed their unique response properties. WFV showed a strong preference for small moving stimuli and a slight bias for certain directions. Interestingly, visual responses to distinct visual stimuli showed variable degrees of adaptation upon repeated presentation. In addition, Cre-dependent rabies tracing unraveled a variety of sources of input which had not been described previously. These results place WFV as a major integrator of visual information in the superior colliculus. Their adaptation to looming stimulus suggest a role in ethologically relevant function in avoiding inappropriate responses, while sustained responses to small moving stimulus might be linked to attentional processes and orienting behaviors. Taken together, this initial characterization paves the way to elucidate the role of these neurons in visual behavior.
Acknowledgements: NIH/NEI K99EY031783 and R01EY026286-05
Talk 4, 9:00 am, 41.24
What makes an elegant walk: Aesthetic preferences for prototypical movements in human walking actions
A common sight in life is seeing other people walk, and our visual systems specialize in processing such actions. Notably, we are not only quick to recognize actions, but also quick to judge how elegantly people walk. What movements lead to an appealing impression, and why do we have such aesthetic experiences? Do aesthetic preferences for body movements arise from perceiving others’ positive emotions? To answer these questions, we showed 150 observers different point-light walkers who expressed neutral, happy, angry, or sad emotions through their movements. Three experiments were conducted to measure the observers’ impressions of aesthetic appeal, emotion positivity, and naturalness of the intact walkers and spatially scrambled point-light displays. We discovered three patterns: (a) When the emotion categories are rendered unrecognizable through spatial scrambling, observers preferred stimuli with faster joint motions. (b) For intact walkers, the more positive the walkers’ emotions appeared, the more aesthetically pleasing they were. (c) Intriguingly, despite the general aesthetic preference for positive emotions, observers gave higher aesthetic ratings for neutral walkers than walkers with other emotions both before and after regressing out the influences of emotion positivity. The aesthetic preference for neutral walkers can be explained by a computational model that captures movement prototypicality. The model used a dynamic time warping algorithm to assess trajectory dissimilarity among walkers, and showed that an aesthetic preference for prototype accounted for the neutral walker preference. The model also predicted both the perceived typicality of walkers (measured by their naturalness ratings) and the aesthetic experiences from different walkers in general. These findings imply possible functions for action aesthetics beyond preferring happy conspecifics: Atypical actions may signal sickness or ill intentions, and thus, liking prototypical agents (or disliking atypical agents) may aid us in avoidance of potential dangers.
Acknowledgements: This project was funded by NSF BSC-1655300 awarded to HL.
Talk 5, 9:15 am, 41.25
The role of motion in the neural representation of social interactions
Humans are inherently social, with dedicated brain regions sensitive to social cues such as faces, bodies, and biological motion. More recently, research has begun to investigate how the brain responds to more complex social scenes. This work has identified the posterior superior temporal sulcus (pSTS) as a key region for processing dynamic social interactions. Findings from static vs dynamic paradigms differ, however, suggesting that the extrastriate body area (EBA) - but not the pSTS - is central to processing simple static dyadic interactions. Despite an upsurge in work investigating social interactions both behaviourally and neurally, the crucial role of motion in interaction perception has not yet been investigated directly. Here, 23 participants viewed videos, image sequences, scrambled image sequences, and static images of dyadic social interactions or non-interactive independent actions. Both bilateral pSTS and left EBA showed sensitivity to motion and interactive content. Indeed, both regions showed higher responses to interactions than independent actions in videos and intact sequences, but not in other conditions. While both regions show this “dynamic-specific” interaction sensitivity, it is seen most strongly in pSTS. Contrary to our expectations, EBA was not sensitive to interactive content in static pictures. Intriguingly, exploratory multivariate regression analyses suggest that, for bilateral pSTS, both interaction selectivity and body selectivity (measured in separate localisers) but not motion sensitivity drive interaction selectivity for videos in our main task. In contrast, interaction selectivity in EBA appears to be driven primarily by body selectivity. More work is needed to understand the different roles EBA may play in processing interactions conveyed by prototypical static dyads (facing dyads) as opposed to our complex and varied dynamic displays. Altogether, these findings support the existence of a third visual stream supporting dynamic social scene perception, where EBA may play a supporting, and pSTS a central role.
Acknowledgements: This work was funded by a European Research Council (ERC) Starting Grant: Grant ID #716974, ‘Becoming Social’
Talk 6, 9:30 am, 41.26
Distributed representations of natural body pose in visual cortex
Hongru Zhu1 (), Yijun Ge2, Alexander Bratch3, Alan Yuille1, Kendrick Kay4, Daniel Kersten4; 1Johns Hopkins University, 2RIKEN Center for Brain Science, 3Stanford University, 4University of Minnesota Twin Cities
Human pose, defined as the spatial relationships between body parts, carries critical visual information about the underlying motion and action of a person. A substantial body of previous work has identified cortical areas responsive to images of different body parts and their spatial relationships. However, these studies were done on a very limited range of poses and with fairly simple stimuli. Our paper investigates high-resolution fMRI responses to a broad range of poses present in over 4,000 complex natural images of people from the Natural Scene Dataset. To help analyze these data, we exploit detailed ground truth annotations created by the computer vision community. Using these annotations, we built models that contrasted view-dependent vs. view-independent and 2D vs. 3D parameterizations of body pose. We compared the similarity of patterns of cortical activity with similarities computed from model activations, thereby identifying how cortical areas represent viewpoint and 2D/3D body pose. We found distributed patterns of cortical activity that captured the similarity structure of the natural pose space in lateral occipital-temporal cortex (LOTC), fusiform gyrus and posterior parietal cortex, including previously studied areas (EBA and FBA). In particular, we found near the right superior temporal sulcus (STS) neural representations that exclusively encode intrinsic, view-independent 3D pose dissimilarity structures. Together, these results reveal a distributed cortical network, encoding both view-dependent and view-independent representations of pose.
Acknowledgements: NIH/NEI grant R01EY029700; NSF CRCNS grant IIS-1822683