Tracking historical changes in trustworthiness using machine learning analyses of facial cues in paintings.
Ontology highlight
ABSTRACT:
SUBMITTER: Safra L
PROVIDER: S-EPMC7508927 | biostudies-literature | 2020 Sep
REPOSITORIES: biostudies-literature
Publications
Tracking historical changes in trustworthiness using machine learning analyses of facial cues in paintings.
Safra Lou L Chevallier Coralie C Grèzes Julie J Baumard Nicolas N
Nature communications 20200922 1
Social trust is linked to a host of positive societal outcomes, including improved economic performance, lower crime rates and more inclusive institutions. Yet, the origins of trust remain elusive, partly because social trust is difficult to document in time. Building on recent advances in social cognition, we design an algorithm to automatically generate trustworthiness evaluations for the facial action units (smile, eye brows, etc.) of European portraits in large historical databases. Our resu ...[more]
Similar Datasets
Project description:Appearance-based trustworthiness inferences may reflect the misinterpretation of emotional expression cues. Children and adults typically perceive faces that look happy to be relatively trustworthy and those that look angry to be relatively untrustworthy. Given reports of atypical expression perception in children with Autism Spectrum Disorder (ASD), the current study aimed to determine whether the modulation of trustworthiness judgments by emotional expression cues in children with ASD is also atypical. Cognitively-able children with and without ASD, aged 6-12 years, rated the trustworthiness of faces showing happy, angry and neutral expressions. Trust judgments in children with ASD were significantly modulated by overt happy and angry expressions, like those of typically-developing children. Furthermore, subtle emotion cues in neutral faces also influenced trust ratings of the children in both groups. These findings support a powerful influence of emotion cues on perceived trustworthiness, which even extends to children with social cognitive impairments.
| S-EPMC4039438 | biostudies-literature
Project description:There have been long-standing differences of opinion regarding the influence of the face relative to that of contextual information on how individuals process and judge facial expressions of emotion. However, developmental changes in how individuals use such information have remained largely unexplored and could be informative in attempting to reconcile these opposing views. The current study tested for age-related differences in how individuals prioritize viewing emotional faces versus contexts when making emotion judgments. To do so, we asked 4-, 8-, and 12-year-old children as well as college students to categorize facial expressions of emotion that were presented with scenes that were either congruent or incongruent with the facial displays. During this time, we recorded participants' gaze patterns via eye tracking. College students directed their visual attention primarily to the face, regardless of contextual information. Children, however, divided their attention between both the face and the context as sources of emotional information depending on the valence of the context. These findings reveal a developmental shift in how individuals process and integrate emotional cues.
| S-EPMC4824404 | biostudies-literature
Project description:BackgroundFaces play a key role in signaling social cues such as signals of trustworthiness. Although several studies identify the amygdala as a core brain region in social cognition, quantitative approaches evaluating its role are scarce.ObjectivesThis review aimed to assess the role of the amygdala in the processing of facial trustworthiness, by analyzing its amplitude BOLD response polarity to untrustworthy versus trustworthy facial signals under fMRI tasks through a Meta-analysis of effect sizes (MA). Activation Likelihood Estimation (ALE) analyses were also conducted.Data sourcesArticles were retrieved from MEDLINE, ScienceDirect and Web-of-Science in January 2016. Following the PRISMA statement guidelines, a systematic review of original research articles in English language using the search string "(face OR facial) AND (trustworthiness OR trustworthy OR untrustworthy OR trustee) AND fMRI" was conducted.Study selection and data extractionThe MA concerned amygdala responses to facial trustworthiness for the contrast Untrustworthy vs. trustworthy faces, and included whole-brain and ROI studies. To prevent potential bias, results were considered even when at the single study level they did not survive correction for multiple comparisons or provided non-significant results. ALE considered whole-brain studies, using the same methodology to prevent bias. A summary of the methodological options (design and analysis) described in the articles was finally used to get further insight into the characteristics of the studies and to perform a subgroup analysis. Data were extracted by two authors and checked independently.Data synthesisTwenty fMRI studies were considered for systematic review. An MA of effect sizes with 11 articles (12 studies) showed high heterogeneity between studies [Q(11) = 265.68, p < .0001; I2 = 95.86%, 94.20% to 97.05%, with 95% confidence interval, CI]. Random effects analysis [RE(183) = 0.851, .422 to .969, 95% CI] supported the evidence that the (right) amygdala responds preferentially to untrustworthy faces. Moreover, two ALE analyses performed with 6 articles (7 studies) identified the amygdala, insula and medial dorsal nuclei of thalamus as structures with negative correlation with trustworthiness. Six articles/studies showed that posterior cingulate and medial frontal gyrus present positive correlations with increasing facial trustworthiness levels. Significant effects considering subgroup analysis based on methodological criteria were found for experiments using spatial smoothing, categorization of trustworthiness in 2 or 3 categories and paradigms which involve both explicit and implicit tasks.LimitationsSignificant heterogeneity between studies was found in MA, which might have arisen from inclusion of studies with smaller sample sizes and differences in methodological options. Studies using ROI analysis / small volume correction methods were more often devoted specifically to the amygdala region, with some results reporting uncorrected p-values based on mainly clinical a priori evidence of amygdala involvement in these processes. Nevertheless, we did not find significant evidence for publication bias.Conclusions and implications of key findingsOur results support the role of amygdala in facial trustworthiness judgment, emphasizing its predominant role during processing of negative social signals in (untrustworthy) faces. This systematic review suggests that little consistency exists among studies' methodology, and that larger sample sizes should be preferred.
| S-EPMC5127572 | biostudies-literature
Project description:Background: The inclusion of facial and bodily cues (clinical gestalt) in machine learning (ML) models improves the assessment of patients' health status, as shown in genetic syndromes and acute coronary syndrome. It is unknown if the inclusion of clinical gestalt improves ML-based classification of acutely ill patients. As in previous research in ML analysis of medical images, simulated or augmented data may be used to assess the usability of clinical gestalt. Objective: To assess whether a deep learning algorithm trained on a dataset of simulated and augmented facial photographs reflecting acutely ill patients can distinguish between healthy and LPS-infused, acutely ill individuals. Methods: Photographs from twenty-six volunteers whose facial features were manipulated to resemble a state of acute illness were used to extract features of illness and generate a synthetic dataset of acutely ill photographs, using a neural transfer convolutional neural network (NT-CNN) for data augmentation. Then, four distinct CNNs were trained on different parts of the facial photographs and concatenated into one final, stacked CNN which classified individuals as healthy or acutely ill. Finally, the stacked CNN was validated in an external dataset of volunteers injected with lipopolysaccharide (LPS). Results: In the external validation set, the four individual feature models distinguished acutely ill patients with sensitivities ranging from 10.5% (95% CI, 1.3-33.1% for the skin model) to 89.4% (66.9-98.7%, for the nose model). Specificity ranged from 42.1% (20.3-66.5%) for the nose model and 94.7% (73.9-99.9%) for skin. The stacked model combining all four facial features achieved an area under the receiver characteristic operating curve (AUROC) of 0.67 (0.62-0.71) and distinguished acutely ill patients with a sensitivity of 100% (82.35-100.00%) and specificity of 42.11% (20.25-66.50%). Conclusion: A deep learning algorithm trained on a synthetic, augmented dataset of facial photographs distinguished between healthy and simulated acutely ill individuals, demonstrating that synthetically generated data can be used to develop algorithms for health conditions in which large datasets are difficult to obtain. These results support the potential of facial feature analysis algorithms to support the diagnosis of acute illness.
| S-EPMC8350122 | biostudies-literature
Project description:Automatic facial coding (AFC) is a promising new research tool to efficiently analyze emotional facial expressions. AFC is based on machine learning procedures to infer emotion categorization from facial movements (i.e., Action Units). State-of-the-art AFC accurately classifies intense and prototypical facial expressions, whereas it is less accurate for non-prototypical and less intense facial expressions. A potential reason might be that AFC is typically trained with standardized and prototypical facial expression inventories. Because AFC would be useful to analyze less prototypical research material as well, we set out to determine the role of prototypicality in the training material. We trained established machine learning algorithms either with standardized expressions from widely used research inventories or with unstandardized emotional facial expressions obtained in a typical laboratory setting and tested them on identical or cross-over material. All machine learning models' accuracies were comparable when trained and tested with held-out dataset from the same dataset (acc. = [83.4% to 92.5%]). Strikingly, we found a substantial drop in accuracies for models trained with the highly prototypical standardized dataset when tested in the unstandardized dataset (acc. = [52.8%; 69.8%]). However, when they were trained with unstandardized expressions and tested with standardized datasets, accuracies held up (acc. = [82.7%; 92.5%]). These findings demonstrate a strong impact of the training material's prototypicality on AFC's ability to classify emotional faces. Because AFC would be useful for analyzing emotional facial expressions in research or even naturalistic scenarios, future developments should include more naturalistic facial expressions for training. This approach will improve the generalizability of AFC to encode more naturalistic facial expressions and increase robustness for future applications of this promising technology.
| S-EPMC9916590 | biostudies-literature
Project description:Defensive system activation promotes heightened perception of threat signals, and excessive attention to threat signals has been discussed as a contributory factor in the etiology of anxiety disorders. However, a mechanistic account of attentional modulation during fear-relevant processes, especially during fear generalization remains elusive. To test the hypothesis that social fear generalization prompts sharpened tuning in the visuocortical representation of social threat cues, 67 healthy participants underwent differential fear conditioning, followed by a generalization test in which participants viewed faces varying in similarity with the threat-associated face. We found that generalization of social threat sharpens visuocortical tuning of social threat cues, whereas ratings of fearfulness showed generalization, linearly decreasing with decreasing similarity to the threat-associated face. Moreover, individuals who reported greater anxiety in social situations also showed heightened sharpened tuning of visuocortical neurons to facial identity cues, indicating the behavioral relevance of visuocortical tuning during generalization learning.
| S-EPMC7311168 | biostudies-literature
Project description:Current microbial source tracking techniques that rely on grab samples analyzed by individual endpoint assays are inadequate to explain microbial sources across space and time. Modeling and predicting host sources of microbial contamination could add a useful tool for watershed management. In this study, we tested and evaluated machine learning models to predict the major sources of microbial contamination in a watershed. We examined the relationship between microbial sources, land cover, weather, and hydrologic variables in a watershed in Northern California, United States. Six models, including K-nearest neighbors (KNN), Naïve Bayes, Support vector machine (SVM), simple neural network (NN), Random Forest, and XGBoost, were built to predict major microbial sources using land cover, weather and hydrologic variables. The results showed that these models successfully predicted microbial sources classified into two categories (human and non-human), with the average accuracy ranging from 69% (Naïve Bayes) to 88% (XGBoost). The area under curve (AUC) of the receiver operating characteristic (ROC) illustrated XGBoost had the best performance (average AUC = 0.88), followed by Random Forest (average AUC = 0.84), and KNN (average AUC = 0.74). The importance index obtained from Random Forest indicated that precipitation and temperature were the two most important factors to predict the dominant microbial source. These results suggest that machine learning models, particularly XGBoost, can predict the dominant sources of microbial contamination based on the relationship of microbial contaminants with daily weather and land cover, providing a powerful tool to understand microbial sources in water.
| S-EPMC7854693 | biostudies-literature
Project description:Creativity is a compelling yet elusive phenomenon, especially when manifested in visual art, where its evaluation is often a subjective and complex process. Understanding how individuals judge creativity in visual art is a particularly intriguing question. Conventional linear approaches often fail to capture the intricate nature of human behavior underlying such judgments. Therefore, in this study, we employed interpretable machine learning to probe complex associations between 17 subjective art-attributes and creativity judgments across a diverse range of artworks. A cohort of 78 non-art expert participants assessed 54 artworks varying in styles and motifs. The applied Random Forests regressor models accounted for 30% of the variability in creativity judgments given our set of art-attributes. Our analyses revealed symbolism, emotionality, and imaginativeness as the primary attributes influencing creativity judgments. Abstractness, valence, and complexity also had an impact, albeit to a lesser degree. Notably, we observed non-linearity in the relationship between art-attribute scores and creativity judgments, indicating that changes in art-attributes did not consistently correspond to changes in creativity judgments. Employing statistical learning, this investigation presents the first attribute-integrating quantitative model of factors that contribute to creativity judgments in visual art among novice raters. Our research represents a significant stride forward building the groundwork for first causal models for future investigations in art and creativity research and offering implications for diverse practical applications. Beyond enhancing comprehension of the intricate interplay and specificity of attributes used in evaluating creativity, this work introduces machine learning as an innovative approach in the field of subjective judgment.
| S-EPMC10415252 | biostudies-literature
Project description:Neighborhood disorder (i.e., physical or social decay) is associated with decreased trust, which reinforces criminal behavior among some individuals in these communities. However, past research largely is descriptive and has not measured processes underlying trust. Using behavioral and neural indices [the late positive potential (LPP), a marker of salience elaboration] in a sample of adults (N = 55), we examined the association between perceived neighborhood disorder and facial trustworthiness perception as well as the potential moderating role of trustworthiness perception on the association between PND and criminal behavior. Individuals with higher perceived neighborhood disorder displayed less LPP differentiation between untrustworthy and trustworthy faces. Moreover, individuals with higher perceived neighborhood disorder and less LPP differentiation were less likely to commit a variety of crimes, whereas those with higher perceptions of neighborhood disorder and high LPP differentiation were more likely to commit a variety of crimes. Combined, these findings suggest that similarly processing trustworthy and untrustworthy faces, as indexed by less LPP differentiation, may reflect an adaptation among those with higher perceived neighborhood disorder that mitigates against deviant behavior and contacts with the law. Understanding the intersection between neighborhood characteristics and individual-level cognitive-affective processing may provide insight into what shapes beliefs and behaviors about important social information.
| S-EPMC7115212 | biostudies-literature
Project description:The impression of trustworthiness based on someone's facial appearance biases our subsequent behavior toward that subject in a variety of contexts. In this study, we investigated whether facial trustworthiness also biases the credibility of utterances associated with that face (H1). We explored whether this bias is mitigated by utterances eliciting reasoning, i.e. explanations (as opposed to factual statements; H2). Moreover, we hypothesized that overimposing facemasks on those faces could enhance/reduce utterance credibility due to social value of mask-wearing (H3), and that facemasks could counter the putative credibility bias introduced by facial trustworthiness (H4). If so, this may be either because facemasks remove the visual information necessary for trustworthiness impression (H4a), or because information is less salient, although it can be retrieved under different circumstances (H4b). An online study (N = 159) was conducted to test these hypotheses. In the first task, subjects saw 48 facial pictures coupled with one utterance and judged the truthfulness/falsity of this utterance. In the second task, they saw again 16 of the faces from the previous tasks and were asked to recall whether the associated utterance was true or false. Findings from the first task support H1 and H4, but not H2 and H3. However, in the second task, where the face is the only available cue, the credibility-mitigation bias exerted by facemask disappears, supporting H4b over H4a. Our results confirm the pervasivity of facial trustworthiness impressions in social cognition, and suggest that facemask can mitigate them, or at least their salience.Supplementary informationThe online version contains supplementary material available at 10.1007/s12144-022-03277-7.
| S-EPMC9178339 | biostudies-literature