糖心Volg

Roberto Caldara

photo

 roberto.caldara@unifr.ch
 +41 26 300 7636
 

Professor
Department of Psychology

RM 01 bu. C-3.117
Rue P.A. de Faucigny 2
1700 Fribourg
RM 01, C-3.117
+41 26 300 7636

Research and publications

  • Publication list
    99 publications


    Fazilet Zeynep Yildirim-Keles, Lisa Stacchi, Roberto Caldara, eneuro (2025) | Journal article

  • Research projects

    Faces in motion: understanding and mapping the decoding of dynamic facial expressions of emotion
    Status: Ongoing

    Start 01.03.2022
    End 28.02.2026
    Funding SNSF 
    Humans communicate social and motivational internal states through complex dynamic facial signals that have been shaped by biological and evolutionary constraints. Everyday human social interactions are characterized by facial expressions of emotion (FEE) that progressively unfold from one expressive or neutral signal to another over a short period of time, before returning to a neutral baseline. During the last decade, we have demonstrated that the spatiotemporal dynamics of FEE are finely tuned by culture and are optimized to rapidly transmit to the decoder orthogonal unambiguous signals for effective communication. However, surprisingly, while real life social interactions are flooded with dynamic signals, most of the scientific literature and knowledge on facial expression recognition (FER) has been developed from the use of static face images. This scientific bias can be partly accounted for by both the technological limitations typical of the early years of FER studies and the subsequent replicability of those studies. Nowadays, technology has massively evolved, and dynamic stimuli can easily be acquired and implemented in experimental designs. Astonishingly, this progress in technology is still not paired with the use of those means in this scientific field, as the large majority of studies continue to use static face images instead of the more ecologically valid dynamic ones. This is even more critical, as frequency of exposure with dynamic faces, evolutionary and ontogenetic evidence all predict a special status of dynamic over static signals for FER. This dynamic over static advantage is objectively confirmed by emergent results in the literature, from us and others, pointing towards a dissociation in terms of development, performance, processes, and neural mechanisms engaged during the decoding of FEE. For instance, in a recent large cross-sectional study, we showed that dynamic compared to static FEE are processed better in early and late life. With the current project, we would like to break from the prevailing practice to progress and further understand the differences between static and dynamic FEE and boost the use of more ecological dynamic faces for scientific research. As such, the main objective of the current research proposal is to expand the understanding of how static and dynamic FEE are processed and decoded from the primary entry point of qualitative and quantitative visual signals to their decoding at the neural level, across different populations: Developing children and adolescents, Adults and Patients (DAP). The overarching aims of the research program include: i. Identifying facial signal recognition thresholds for static and dynamic expressions in children and adolescents (D), adults (A), and patients (P) by using novel psychophysical approaches ii. Isolating the perceptual strategies used to recognize static and dynamic expressions in the DAP populations by mapping eye movement fixation patterns in typical and atypical performance iii. Elucidating further the underlying spatiotemporal brain mechanisms of static and dynamic FEE processing, and tracking the neural markers of specific recognition abilities and deficits To achieve these aims, we will use a multidisciplinary approach with novel psychophysical techniques, behavioral measures, eye tracking, electrophysiology, and functional Magnetic Resonance Imaging, as well as developing a new toolbox for the statistical analysis of eye movements with dynamic stimuli (iMap Motion). We genuinely believe that highlighting the differences between the processing of static and dynamic FEE has the potential for profound theoretical, social, clinical, and economic impact. From a theoretical point of view, we are truly convinced that investigating the use of dynamic faces will be a major step forward in the understanding of face perception in ecological settings, paving the way to reinforce scientific research in that direction. The research planned here will provide a tighter link between behavioral, neuropsychological, and functional neuroimaging findings and the way in which faces are processed by humans in everyday life. Modern social interactions are characterized by the decoding of both realistic dynamic FEE and static ones, typically used in the virtual social networks. Thus, at the societal level, outlining the specific processes involved in both types of signals could promote the use of dynamic signals for a given target audience, for example, the elderly and fragile populations, which show more effective processing of dynamic signals. Ultimately, this could optimize and enhance intergenerational affective communication. Importantly, the knowledge and techniques developed here could also be of use in a variety of clinical settings to tailor early interventions or rehabilitation training programs, as well as in the realm of patient care. Finally, since dynamic expressions are increasingly used, the research here has the potential for knowledge transfer and economic impact, as it could attract interest in the advertising, animation, computing, and robotic industry.

    The Speed of Race
    Status: Completed

    Start 01.03.2020
    End 28.02.2025
    Funding SNSF 
    The face is arguably the first and foremost visual category of human expertise. Through repeated exposure, we develop the ability to recognise within a fraction of a second a person鈥檚 emotional states, age, gender, race and identity. The relevance of experience with faces for the development of this expertise is illustrated in an effect that can be observed in everyday life 鈥� the striking inability people have in recognising faces belonging to other-races compared to faces of their own race, the so-called Other-Race Effect (ORE). The ORE is important at the theoretical level for the understanding of how the neural face system and its processing tunes with experience, as well as on the very nature of face representations. This phenomenon also has practical consequences that range from the trivial to the deadly, such as eyewitness misidentification of other-race faces in criminal cases, as in the police shooting of Brazilian expat Jean Charles de Menzes in the day following a failed bombing of the London Underground. Interestingly, this decreased ability to recognise other-race faces due to decreased visual and social experience is paired by a seemingly paradoxical speed advantage humans have when the task requires categorising faces by race: the so-called Other-race Categorisation Advantage (ORCA). The ORCA relates to the faster response times elicited by other-race faces when observers categorise faces by race. However, while the ORE scientific literature is sizeable, only a few studies have investigated the ORCA, and none have explored the extent to which both effects are linked and tap into the very same mechanisms. Perceptual learning has been suggested to play a key role in diverse mechanisms and theoretical explanations of this phenomenon. Critically, however, we do not yet know at which stage in the early processing of faces that race effects begin to modulate our behaviour, or how attention to different aspects of faces as either a function of task or visual sampling bias may additionally modulate these effects. Thus, it is still unclear whether the ORE results either from suboptimal visual information processing for other-race faces or from a categorization that would occur at an earlier stage (i.e., ORCA) and would shape the subsequent processes, or both. In fact, the mere rapid categorization of people into ethnic groups might be sufficient to increase attention to in-group members and lead to a devaluation of out-group members, resulting in a decrease in face processing performance. We are thus at a point in the current state of research where a series of targeted experimental investigations should shed greatly-needed light on when and where race is processed in the brain, and how the early stages of this processing lead to the behavioural patterns observed in the other-race effects. To understand the key factors of both theoretical frameworks, we have thus planned a series of studies that aim to clearly identify the origin of the ORE and the ORCA, by: i. Tracking the neural dynamics of the ORCA. ii. Investigating the role of attention in the modulation of the ORCA. iii. Identifying the nature of same- and other-race face representations in the brain. We will use a psychophysical approach with experiments involving behavioural measures, electrophysiological, eye-movement signals, functional Magnetic Resonance Imaging (fMRI), as well as brain damaged patients. When appropriate, an original dimension of our approach will be the co-registration of these three data-streams (behaviour, EEG and eye-movements), providing information on the spatiotemporal neural activity and information use in the ORE and ORCA. This innovative integrative framework will reveal critical insights on the nature of the other-race effects, with potential benefits of the research outlined here going well beyond the research community, as they could have a significant impact in criminal and justice settings. We also genuinely believe that in an increasingly globalized world and other-race migration flows identifying neural sensitivity and further characterizing the perceptual differences leading to the extraction of race information has the potential for profound theoretical and social impact for human interactions of our modern multi-racial societies.

Teaching and courses

Show in course program