The research, led by Lily Chen and Eva Kimonis of the University of New South Wales together with colleagues at Macquarie University, is the first to disentangle the unique contributions of ASD symptoms and CU traits to different components of empathy in clinic-referred children as young as one to eight years old. The distinction matters enormously for treatment. Children with conduct problems, which include the symptoms of oppositional defiant disorder and conduct disorder, are among the most common referrals to psychological services, and an estimated 20 to 50 percent of them also show CU traits, a pattern marked by lacking guilt, remorse, and concern for others. Meanwhile, between 22 and 51 percent of young people with ASD also display elevated CU traits, creating a diagnostic tangle that has long frustrated clinicians trying to decide which intervention a child actually needs.

The theoretical backbone of the study is a hierarchical model of empathy proposed by Nuske and colleagues, which treats empathic responding as the final product of an assembly line. First, a child must attend to another person’s emotional cues. Next, the child must recognise which emotion those cues express and understand the underlying emotional state, a capacity related to theory of mind. Only then can the child produce a behavioural empathic response, such as comforting or helping. According to the model, empathy can fail at any link in this chain, and crucially, different disorders may break different links. Prior meta-analyses suggested that CU traits blunt affective resonance, the emotional arousal that motivates prosocial action, whereas ASD primarily disrupts cognitive empathy, including emotion recognition and perspective taking. But almost no previous study had tested both conditions simultaneously in the same young clinical sample.

The researchers recruited 317 children referred to a university-based or community-based clinic in metropolitan Sydney for assessment and treatment of conduct problems, plus 91 typically developing children from the same communities. Parents completed well-validated questionnaires: the Eyberg Child Behaviour Inventory for conduct problems, the Social Communication Questionnaire for ASD symptoms, and the preschool version of the Inventory of Callous-Unemotional Traits. Children themselves then completed a battery of laboratory tasks. A preschool dot-probe task measured how quickly children orient their attention toward images of distress. A dynamic emotion recognition task presented 40 one-second video clips of adults expressing anger, fear, happiness, sadness, or pain, and asked children to identify the emotion. Understanding of emotional states was probed with an ingenious task built around four emotional scenes from the animated film The Lion King, in which children were asked how Simba felt and why. Prosocial responding was captured through parent-report measures of empathic behaviour.

The statistical strategy was deliberately layered. Hierarchical regressions tested whether ASD symptoms or CU traits were uniquely associated with each empathy component, over and above the severity of conduct problems and the influence of components lower in the hierarchy. The results drew a strikingly clean dividing line. ASD symptoms, but not CU traits, were uniquely associated with poorer emotion recognition, a small-to-moderate effect. When the researchers examined specific emotions, higher ASD symptoms predicted deficits in recognising anger, fear, and pain, though not happiness or sadness in this sample. In contrast, CU traits, but not ASD symptoms, were uniquely and negatively associated with empathic responding, and this association was small to moderate in size. Neither condition showed a unique association with attention to distress cues or with understanding of emotional states once the lower components of the hierarchy were accounted for.

The person-centred analyses told the same story from a different angle. Children were classified into clinical groups: conduct problems alone, conduct problems plus CU traits, conduct problems plus ASD, and conduct problems with both. The two groups containing ASD, CP+ASD and CP+CU+ASD, showed significantly poorer emotion recognition than the CP+CU group, whose accuracy did not differ from typically developing children. The pattern reversed for empathic responding. Here the CP+CU+ASD group showed the lowest prosocial behaviour of all, significantly below the CP+CU group, which in turn fell significantly below the CP-only group. Children with conduct problems alone did not differ from typically developing children. In other words, ASD erodes the cognitive front end of the empathy chain, while CU traits erode the affective back end, and when both are present the damage compounds.

One of the most consequential implications of the findings is that emotion recognition deficits long attributed to psychopathy and CU traits may actually have been driven by undetected, co-occurring ASD symptoms. Earlier meta-analyses had reported that children with CU or psychopathic traits struggle to recognise anger, fear, and sadness. But those analyses never controlled for ASD. The current results, consistent with a 2021 study of seven-year-old community children, suggest that once ASD symptoms are accounted for, the link between CU traits and emotion recognition largely evaporates. This reframing matters because it redirects clinical attention: a child with conduct problems and poor emotion reading may need support with social perception, not the interventions designed for shallow affect.

Equally intriguing is what the study found about children with ASD whose empathic responding remains intact. Despite measurable emotion recognition deficits, these children engaged in prosocial behaviour at levels comparable to typically developing peers. Emerging research suggests that some autistic individuals experience differences rather than deficits in emotional processing, including hypersensitivity to others’ emotions that can heighten emotional resonance in some situations, even when recognising facial expressions is harder. This challenges the popular but inaccurate stereotype that autism is defined by an absence of empathy, and it aligns with systematic reviews showing that autistic children comfort, help, and share to the same degree as their neurotypical peers.

The study is not without limitations, which the authors acknowledge candidly. Some laboratory tasks, particularly the dot-probe measure of attention and the Lion King task of emotional understanding, may have been too demanding for children as young as one to four years old, since the attention task is validated only from age three and the understanding task from age five. Low accuracy rates and high variance on these measures may explain why no unique associations emerged for those components, contradicting some earlier work with older children and adolescents. Laboratory data were also not available for every participant, leaving some analyses underpowered, and the sample contained no children with ASD without comorbid conduct problems, so the findings cannot yet be generalised to that population. The authors suggest that mobile eye tracking, which requires no active response from the child, could provide a more developmentally appropriate window onto attentional processes in future research.

The clinical payoff is concrete. For children with conduct problems and co-occurring ASD, the hierarchical model implies that adding an emotion recognition module to treatment could produce flow-on benefits to emotional understanding and empathic behaviour. For children with CU traits, whose recognition and understanding appear relatively preserved, the evidence points toward directly training empathic responding itself. One promising example is the Coaching and Rewarding Emotional Skills module adapted for Parent-Child Interaction Therapy for children with CU traits, which uses parent modelling and an intensive reward-based token economy to reinforce prosocial behaviour and has been shown to significantly improve child empathy. Given that CU traits predict earlier-onset, more severe, and more treatment-resistant antisocial behaviour, identifying which children need which kind of support, and doing so before school age, could change developmental trajectories that are otherwise difficult to redirect. The authors call for replication in larger samples and in populations beyond clinic-referred children, but the central message is already clear: empathy fails in different places for different children, and treatment should target the exact link that is broken.

Subject of Research: Differential associations of childhood callous-unemotional traits and autism spectrum disorder symptoms with components of empathy in children with conduct problems

Article Title: The Empathy Hierarchy: Unique Associations of Childhood Callous Unemotional Traits and Autism Spectrum Disorder Symptoms

Article References: The Empathy Hierarchy: Unique Associations of Childhood Callous Unemotional Traits and Autism Spectrum Disorder Symptoms. (n.d.). https://doi.org/10.1007/s10578-026-02104-5

Image Credits: AI Generated

DOI: 10.1007/s10578-026-02104-5

Keywords: autism spectrum disorder, callous-unemotional traits, empathy, emotion recognition, conduct problems, prosocial behaviour, theory of mind, child psychopathology, hierarchical model of empathy, clinical assessment, developmental psychology, Parent-Child Interaction Therapy

The scale of the underlying dataset is considerable. The researchers conducted a genetic meta-analysis drawing on nearly 500,000 social behavioural observations from more than 73,000 individuals, many of whom were assessed repeatedly between the ages of 2 and 29. This longitudinal depth allowed the team to examine developmental change directly, rather than inferring it from cross-sectional snapshots. Their analytical strategy focused on two broad domains of social behaviour. The first, low prosocial behaviour, captures individual differences in helping, sharing, and comforting others, qualities that emerge early in childhood and continue to mature through adolescence. The second, peer and social difficulties, encompasses problems forming or maintaining relationships, such as difficulties making friends or navigating group dynamics.

Methodologically, the study departs from the conventions of most large genome-wide association studies. Where typical analyses pool information across individuals and measurements to obtain a single overall genetic estimate, the team adopted a meta-regression approach that explicitly models how genetic effects vary across social measures, multiple informants, and different ages. The researchers refer to this combination of domain, reporter, and developmental stage as social context. By treating context as a variable to be modelled rather than noise to be averaged away, the analysis could detect genetic signals that conventional designs would have obscured, and could test formally whether those signals behave differently under different observational conditions.

The results were striking. The analysis identified six regions of the genome associated with social behaviour, and every one of them showed evidence that its associations differed according to behavioural domain, reporter, and age. The strongest association signal fell within CADM2, a gene involved in organising connections between nerve cells. This signal was particularly pronounced for self-reported low prosocial behaviour during late childhood and early adolescence, a period when peer relationships intensify and self-perception of social functioning becomes more reliable. Notably, different genetic variations within CADM2 have previously been linked to autism and other traits, placing the new finding within a broader literature on the genetic overlap between social behaviour and neurodevelopmental conditions.

Beyond individual genomic regions, the team estimated how much of the variation in social behaviour can be attributed to common genetic variation, measured through single-nucleotide polymorphisms, or SNPs. Between 2 and 7 percent of variation in social behaviour was captured at this level, a figure consistent with the highly polygenic nature of behavioural traits, in which thousands of variants each contribute a small effect. At this broader level, the genetic architecture resolved into four interrelated overarching genetic factors. Two of these factors corresponded to the behavioural domains themselves, one for low prosocial behaviour and one for peer and social problems. The remaining two factors reflected systematic differences between parent and teacher perspectives, suggesting that observer viewpoint carries genetic signal in its own right rather than simply adding measurement error.

This reporter effect is one of the study’s most provocative implications. A mother describing her toddler learning to share, a teacher watching a schoolchild navigate friendships, and a teenager reflecting on their own social relations are each reporting on behaviour within a distinct social context, and the genetic signals underlying those reports are not interchangeable. Children move through many different social worlds requiring different types of behaviour, at home and at school, and their behaviour changes with age. Senior investigator Beate St Pourcain argued that genome-wide studies need to take these behavioural differences into account rather than averaging effects into a single estimate. First author Lucía de Hoyos framed the methodological lesson in developmental terms: human development is dynamic, and the new analyses offer an opportunity to make genome-wide studies dynamic too.

To test whether the context-specific findings held beyond the discovery sample, the researchers constructed polygenic scores, which summarise the combined effects of many small SNP effects into a single predictive index for each individual. These scores were evaluated in 16,305 people from independent validation cohorts of European and African ancestry. The results showed that predictions of social behaviour improved when the social context of the studied cohorts was matched, meaning that a score built from, for example, teacher-reported data predicted teacher-reported outcomes better than it predicted self-reported or parent-reported outcomes. This context-matching effect provides practical confirmation that social context is not a statistical nuisance but a real structural feature of behavioural genetic data.

The team then turned to the genetic relationships between childhood social behaviour and traits measured later in life, including mental health outcomes such as autism and ADHD. Genetic correlations, which quantify the extent to which the same variants influence two different traits, became detectable at different developmental stages depending on the particular social context studied. In other words, the window during which genetic links to conditions like ADHD or autism emerge is itself context-dependent. The authors suggest this offers new insight into neurodevelopment, hinting that the developmental trajectories connecting early social difficulties with later clinical outcomes may be more nuanced and stage-specific than aggregate analyses have implied.

The researchers are careful to draw a firm line between population-level genetic insight and individual prediction. They emphasise that the findings cannot be used to predict how any individual child will behave. Polygenic scores for behavioural traits remain far too weak for forecasting, and the study’s central message is conceptual rather than clinical: understanding the biology of human behaviour requires understanding the context in which behaviour develops. Thomas Bourgeron, coordinator of the R2D2-MH consortium, underscored this point, reminding that genetic variants matter but are not destiny, and that with the right environment, support, and opportunities, people can adapt and flourish. St Pourcain noted that the findings can help better understand the developmental trajectories of social behaviour and their links with later-life and mental-health outcomes, and announced that the team will follow up with a sex-specific meta-regression approach studying differences across males and females.

The work was carried out under the EAGLE and R2D2-MH consortia, with R2D2-MH funded by Horizon Europe under grant agreement number 101057385, by UK Research and Innovation under the UK government’s Horizon Europe funding guarantee with grant number 10039383, and by the Swiss State Secretariat for Education, Research and Innovation under contract number 22.00277. For a field that has long treated social behaviour as a single measurable quantity, the study delivers a technically grounded corrective: genetic influence on how we help, share, and connect is real, heritable, and heterogeneous, and it is woven into the specific social worlds, developmental moments, and relationships in which behaviour unfolds. Averaging across those worlds, the authors show, risks erasing the very biology scientists are trying to find.

Subject of Research: Genome-wide meta-regression of social behaviour across social domains, reporters, and developmental stages

Article Title: Social context matters when studying the genetics of social behavior

Article References: Social context matters when studying the genetics of social behavior. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: social behaviour, genetics, genome-wide association study, meta-regression, prosocial behaviour, peer difficulties, CADM2, autism, ADHD, polygenic scores, neurodevelopment, Nature Human Behaviour

The research, conducted by Matthew Furnell, James Elander and William Van Gordon of the University of Derby, compared two eight-week second-generation mindfulness-based interventions, or SG-MBIs. Unlike first-generation programmes such as Mindfulness-Based Stress Reduction, which centre almost exclusively on attentional training, second-generation approaches explicitly incorporate practices derived from Buddhist ethics and wisdom traditions. The study exploited a novel threefold classification system grounded in the Buddhist three trainings of concentration (samādhi), ethics (śīla) and wisdom (prajñā). One intervention, labelled MBI-CE, combined concentration-based practices such as breath-focused awareness with ethics-based practices including loving-kindness and compassion meditation. The second, MBI-CEW, retained those elements but replaced three weeks of ethics training with wisdom-based practices exploring impermanence, attachment and interdependence.

Participants were 38 non-Buddhist students aged 18 to 24 at a British overseas university in mainland China, none of whom had previously completed a formal mindfulness programme. The researchers gathered written course reflections from the entire cohort as exploratory scoping material, then conducted semi-structured interviews with twelve randomly selected participants, six from each condition. Interview transcripts were analysed using reflexive thematic analysis, a method that treats the researcher’s own perspective as an explicit part of the interpretive process rather than a bias to be eliminated. The lead analyst, himself a Buddhist meditation practitioner and the designer of both interventions, documented how his familiarity with the material might have shaped what he noticed in the data, and deliberately attended to accounts of confusion, ambivalence and difficulty alongside positive reports.

Four themes emerged from the analysis. The first was that Buddhist-derived concepts were experienced as meaningful but not uniformly accessible. Far from recoiling from explicitly Buddhist content, participants frequently described it as helpful and educational, clarifying the purpose of the meditations rather than obscuring it. Storytelling played a striking role here: parables such as A Chinese Farmer (Maybe) and The Meaning of True Tranquillity helped translate abstract philosophy into relatable narrative. One participant captured the prevailing sentiment neatly, explaining that the course was not about believing in Buddhism but about showing very good perspectives on how this philosophy can positively impact life. The concepts functioned as an educational resource, not a religious demand.

Yet accessibility varied considerably across practice types. Some students found the later ethics-based practices, which asked them to extend gratitude and kindness towards strangers, conceptually and emotionally harder than simple breath focus. One participant admitted confusion about why one should feel grateful for strangers and difficulty feeling into the body during those meditations, remarking that it would be easier to just focus on the breath. In the wisdom-based condition, another student described the content as positive but difficult because it stressed the inside self. These accounts point to uneven terrain rather than outright rejection: practices probing the nature of self, long regarded as among the most demanding dimensions of Buddhist training, asked participants to entertain ideas that their existing frameworks did not necessarily accommodate.

Cultural and linguistic context added further layers. Several participants noted that Buddhist-derived practices felt culturally familiar because of their deep historical roots in Chinese philosophy, yet they distinguished between surface familiarity and genuine understanding, with one observing that many people in China know of the Buddha without knowing what the teachings actually contain. Language mattered too: one participant highlighted how the Chinese term zhèngniàn, often translated as mindfulness, carries nuances that English struggles to capture. The study’s setting is significant, since the students lived within a culture historically shaped by Buddhist ideas but also by Confucian ethical traditions, which prioritise hierarchical obligations to family and close relations in ways that can sit uneasily with the Buddhist ideal of extending compassion equally to all beings. That tension may partly explain why kindness towards strangers proved challenging for some.

The second major theme concerned well-being, awareness and connectedness. Participants in both conditions reported that the practices helped them regulate anger, exhaustion and academic pressure, with several describing short meditative breaks used to defuse stress between study sessions. Crucially, the benefits did not stop at stress reduction. Students in the wisdom-based condition described contemplations of impermanence prompting them to question the stability of their own identity, with one participant noting that the class led them to think about philosophy concerning who they really are. Participants in both groups also reported heightened attentiveness to other people’s emotions, but the flavour of connectedness differed. The ethics-focused group tended to describe empathy and observation of others’ feelings, while the wisdom group more often articulated a broader sense of interdependence extending to family, community, nature and the environment, with one student describing how meditation practice shifted her attention from rushing down the street to noticing trees and birds.

The third theme offered a small but revealing behavioural window. At the end of the course, participants entered a raffle for 1000 RMB and could keep their ticket or donate it to the World Food Programme. Many framed their donation as an expression of connection, care and even moral responsibility, with one student stating simply that he felt responsible to make the world more kind, and another linking the decision to the course’s opening primer, which framed meditation as a practice undertaken to better oneself, one’s family and one’s community. Not every participant reasoned in these terms, however; one cited the low odds of winning, a reminder that prosocial behaviour rarely follows a single motivational script.

The fourth theme concerned structure. Students valued the eight-week, two-hour format as a protected space they might otherwise have filled with less meaningful activity, and one first-year student credited the course with easing her transition into university life more effectively than a standard induction week. But the programme’s limits were also visible. The weekly meditation journal, intended as a reflective tool, sometimes felt like homework, with one participant confessing to making up entries when time ran short. Others felt the course was too compressed to explore certain practices in real depth, exposing a tension between the breadth of the contemplative repertoire and the time available to embody it.

The study’s broader message is methodological as much as practical. Because the two interventions produced distinct experiential patterns, with ethics-based practices aligning with empathy and interpersonal warmth and wisdom-based practices aligning with existential reflection and ecological connectedness, the authors argue that second-generation MBIs should not be treated as an interchangeable category. Grouping interventions with different contemplative compositions under one label risks masking meaningful variation in both content and outcome. At the same time, the findings suggest no trade-off: adding ethics and wisdom elements did not erode the stress-regulation and attentional benefits associated with classical mindfulness training, and a companion quantitative study found greater improvements in well-being, connectedness and prosocial attitudes in the wisdom-inclusive condition. The researchers caution that their sample was culturally homogeneous and the interview sub-samples small, so the findings remain exploratory. But the core insight travels well: what people get from meditation depends substantially on what, precisely, they are asked to practise, and careful pedagogical framing, through stories, metaphors and experiential activities, may matter as much as the practices themselves.

Subject of Research: Participant experience of second-generation mindfulness-based interventions incorporating Buddhist-derived ethics- and wisdom-based practices among non-Buddhist university students

Article Title: How Practice Differences Shape Participant Experience in Second-Generation Mindfulness-Based Interventions: A Qualitative Study of Non-Buddhist University Students

Article References: Furnell, M., Elander, J., & Van Gordon, W. (2026). How Practice Differences Shape Participant Experience in Second-Generation Mindfulness-Based Interventions: A Qualitative Study of Non-Buddhist University Students. Mindfulness. https://doi.org/10.1007/s12671-026-02997-2

Image Credits: AI Generated

DOI: 10.1007/s12671-026-02997-2

Keywords: mindfulness, meditation, Buddhism, second-generation mindfulness-based interventions, qualitative research, university students, loving-kindness, wisdom practices, prosocial behaviour, well-being, connectedness, thematic analysis

The research, conducted by Marius C. Vollberg of the University of Zurich and University of Geneva and Jörg Gross of the University of Zurich, spans nine preregistered online experiments involving 8,769 adults, predominantly residents of the United States. The central finding is striking in its symmetry: both selfish and prosocial individuals distort the social consequences of their choices, but they do so in predictably different directions. Selfish people shrink the apparent harm of their actions, while prosocial people inflate the apparent benefit of theirs. Crucially, when these distorted interpretations are transmitted to others, only the selfish-serving version reliably spreads, nudging those who encounter it toward more selfish behaviour.

The experimental logic was elegantly simple. Participants, cast as ‘Deciders’, chose between two options with fully known personal payoffs: option A yielded 7 points for themselves, option B yielded 5. Each option also carried an ambiguous consequence for another anonymous participant, the ‘Receiver’. Option A gave the Receiver between 1 and 3 points, option B between 3 and 5. Because the Decider’s own payoff was certain while the Receiver’s was a range, the design captured a fundamental asymmetry in social life: personal outcomes are immediate and tangible, whereas social outcomes are shadowy and hard to verify. After choosing, participants estimated the Receiver’s actual outcome for both options, with real money riding on the accuracy of their guesses.

If participants treated each outcome within a range as equally likely, their estimates should cluster around the midpoint and should not depend on what they chose. That is not what happened. Selfish Deciders, those who took the higher payoff, reported the two options’ social consequences as more similar than prosocial Deciders did, effectively portraying their choice as socially harmless. Statistically, the difference was robust: prosocial Deciders’ estimates exceeded selfish Deciders’ by a coefficient of 0.24, with a 95 percent confidence interval of 0.14 to 0.34. Selfish Deciders exaggerated the Receiver’s outcome under the selfish option while downplaying it under the prosocial option, compressing the apparent moral gap between the two choices until their selfishness looked costless.

Prosocial Deciders distorted too, but differently. They exaggerated the social benefit of the option they chose, and, intriguingly, they also exaggerated the benefit of the selfish option they rejected. Rather than painting their sacrifice as heroic against a dark alternative, they construed a world in which even a modest social benefit justified forgoing personal gain, casting themselves as people who need little prompting to do good. In a neutral condition where the self-maximizing option also maximized the Receiver’s payoff, removing any conflict between self-interest and social interest, estimates drifted closer to the true midpoints, though some downplaying of the unchosen alternative persisted.

Sceptics might wonder whether these deviations reflect genuine beliefs about how the ambiguous outcomes were generated rather than motivated distortions. The team addressed this in a conceptual replication that made the outcome-generation process fully transparent: participants watched an animation showing numbers being sampled into an urn and one drawn at random, making clear that the expected value equalled the range midpoint and that their choice could not influence the process. Distortions persisted. The selfish pattern replicated in full, and the overall difference between Decider types remained strong, indicating that the effects cannot be dismissed as private theories about hidden sampling procedures.

Two further experiments probed the psychology behind the distortions. When participants reported their estimates before choosing, the distortions were already present, though attenuated, ruling out a purely post-hoc justification account. More decisive was a forced-choice design in which participants were randomly assigned to implement either the selfish or the prosocial option. If distortions served to justify specific actions, forcing a choice should have reshaped them accordingly. Instead, distortions tracked participants’ underlying social preferences, measured under free choice, better than the choice they had been forced to make; a Bayesian model comparison strongly favoured the preference-based account. Distortions, in other words, are not excuses manufactured after the fact but preference-consistent interpretations of social reality.

Yet the distortions also revealed a telling double standard. In a separate task, participants faced the very same ambiguous ranges, but this time the outcomes determined their own bonuses. Their earlier estimates about others’ outcomes did not predict how they treated equivalent ambiguity when their own money was at stake. People who had reported that a 3-to-5 range favoured the Receiver did not behave as if they believed it when choosing gambles for themselves. This suggests the distortions are not sincere beliefs about underlying distributions but selective interpretations reserved for the social domain, beliefs held too weakly to guide decisions where self-interest is directly measurable.

Do these distortions actually pay off socially? A fifth experiment asked naive observers to rate the trustworthiness of Deciders whose choices were presented with fully known social outcomes drawn from the distortion patterns observed earlier. Selfish Deciders whose outcomes implied both options were socially equivalent, the aggregate selfish distortion, were rated significantly more trustworthy than selfish Deciders in any other scenario, including the undistorted midpoint case. Prosocial Deciders likewise earned higher trust ratings when outcomes matched their aggregate distortion, implying a generous sacrifice. Distortion, when it goes undetected, appears to buy reputation, offering a concrete social function for what might otherwise look like mere self-deception.

The most consequential finding came last. In a final experiment, participants saw estimates attributed to a previous participant, an ‘Influencer’, before making their own choices and estimates. When the Influencer’s estimates were selfish-serving, portraying the selfish option as no more harmful than the prosocial one, Followers adopted those estimates and became significantly more likely to choose selfishly themselves, an effect that held against every comparison condition. Prosocial-serving distortions, though sometimes echoed in individual estimates, produced no reliable shift in average estimates or choices. Because fewer than 10 percent of Deciders actually produce the selfish aggregate pattern, a small minority may exert a disproportionate influence on collective beliefs. The authors suggest that information lowering the psychological cost of selfishness appeals to undecided actors in a way that glorified sacrifice does not. In domains from climate mitigation to foreign aid, the study offers a psychological foundation for why polarized views about the impact, rather than merely the value, of our actions take hold: when social consequences are open to interpretation, people may predictably form divergent interpretations, and the most self-serving of them can spread.

Subject of Research: How selfish and prosocial individuals distort estimates of the ambiguous social consequences of their decisions, and how those distortions spread socially

Article Title: Selfish and prosocial individuals both distort their actions’ impact, but selfish distortions can be more contagious

Article References: Vollberg, M. C., & Gross, J. (2026). Selfish and prosocial individuals both distort their actions’ impact, but selfish distortions can be more contagious. Nature Human Behaviour. https://doi.org/10.1038/s41562-026-02585-3

Image Credits: AI Generated

DOI: 10.1038/s41562-026-02585-3

Keywords: social preferences, decision making, motivated reasoning, selfishness, prosocial behaviour, ambiguity, trustworthiness, social contagion, behavioral economics, moral psychology, Nature Human Behaviour, preregistered experiments

The review rests on a deceptively simple premise. Empathy, defined in the literature as the ability to understand and share another person’s emotional state, is widely assumed to be trainable, much like a muscle. If that assumption holds, and if stronger empathy leads to more prosocial behaviour, then tools that reliably enhance empathy could transform education, health care and workplace training. VR seems uniquely suited to the task because it combines immersion, embodiment and a sense of presence, allowing users to inhabit a perspective rather than merely imagine it. That removes the extra cognitive load of traditional perspective-taking exercises and sidesteps classic methodological headaches such as the trade-off between experimental control and real-world realism.

The theoretical case for VR empathy training draws on theory of mind, mirror neuron research and embodied cognition. Theory of mind describes our capacity to attribute thoughts, emotions and intentions to others, while mirror neurons fire both when we perform an action and when we watch someone else perform it. Embodied cognition holds that thinking is rooted in bodily experience rather than abstract computation alone. Together, these frameworks suggest that simulating another person’s situation in an immersive environment could genuinely reshape how users understand others. Some empirical results are striking: in one well-known study, participants who spent just a few minutes embodied in a virtual environment later spent twice as much time helping others as those who received conventional training. In another, people who experienced homelessness in VR did not report higher empathy on questionnaires than controls, yet they signed supportive petitions at significantly higher rates, hinting that behaviour change may outpace self-report.

But the evidence is far from uniform. A meta-analysis of 43 studies concluded that VR reliably arouses compassionate feelings, boosting emotional or affective empathy, yet does little to encourage users to adopt other people’s cognitive perspectives. Another meta-analysis using a narrower definition of VR found the opposite pattern, with gains in cognitive empathy but no significant effect on emotional empathy. The contradiction, the review’s authors argue, likely stems from the sheer variety of products lumped under the VR umbrella, from passive 360-degree films to fully interactive embodied simulations. Without a shared taxonomy, researchers comparing apples to oranges have produced a muddled literature, and the choice of tool directly shapes the quality of the evidence future studies can generate.

To bring order to the chaos, Khukalenko and Khanolainen ran a three-step search. They queried the largest VR distribution platforms, including Steam, the Epic Games Store, the MetaQuest online store, VIVEPORT, PlayStation VR, PICO VR and the Pimax store, using only empathy-related keywords. They then searched PubMed, EBSCOhost and ERIC for studies employing VR products, screening 270 records down to 127 full-text articles. Finally, a Google search combining VR, empathy and training terms generated more than seven million results, of which the first 300 were screened. The outcome was a five-part classification of empathy-related VR content: non-interactive video content such as the United Nations’ Waves of Grace series; semi-interactive VR experiences like Becoming Homeless and Notes on Blindness; structured training software; multisensory simulations used in dementia and disability education, which add props such as stiff gloves or full-body suits; and authoring platforms like Engage and Vizard that let institutions build their own content.

The heart of the review is its deep dive into the third category. Of 18 training applications identified, only five were available for hands-on testing, and the authors evaluated each with headsets in hand: Bodyswaps, People Tech Revolution, SkillsVR, Vrailexia and Kiin. Bodyswaps, a soft-skills platform with AI-generated feedback, includes a health care module in which users practise person-centred communication with a virtual dementia patient named Geoffrey, then receive automated analysis of their eye contact, body language and use of patronising language. People Tech Revolution places users in the embodied role of Lucy, a jobseeker with social anxiety confronting a hostile employment coach, using heartbeat audio and a mirror tutorial to reinforce identification. SkillsVR stages a late-night train conflict and then walks learners through multiple perspectives in a reflection room. Vrailexia, aimed at dyslexia awareness, traps users in an alchemy game with mismatched labels so they experience the pressure dyslexic students face. Kiin, co-founded by VR researchers Mel Slater and Mavi Sanchez-Vives, lets users swap bodies during a simulated meeting to spot and counter microaggressions.

What distinguishes these five products from the flood of videos and experiences is genuine instructional design: explicit learning objectives, structured activities and built-in assessment or feedback. That matters for two reasons, the authors argue. First, ready-made pedagogy spares educators from improvising their own learning scaffolds around a passive experience. Second, a consistent built-in design makes an application a valid, replicable research instrument. When researchers bolt their own objectives onto a product lacking them, every participant receives a slightly different intervention, and variability compounds across research teams, corroding the replicability of the field. The reviewed apps broadly align with established frameworks that combine experiential methods such as role-play with didactic teaching and skills practice.

Yet the review also sounds a note of caution that cuts against Silicon Valley’s techno-solutionist marketing. Companies promise cutting-edge, effective solutions, but rigorous proof remains scarce, and some design choices are questionable. In Vrailexia’s potions game, for example, the failure of a first attempt leads to the death of a virtual friend the user cannot save, a scenario that could plausibly trigger empathy distress rather than empathic concern, prompting self-centred withdrawal instead of altruistic motivation. Compounding the problem, free demo versions are rare, effectively forcing schools and clinics to make blind purchases, and intervention studies rarely disclose even the name of the software they tested, let alone open access to it, blocking replication efforts entirely.

The authors sketch a research agenda to close these gaps. They call for experiments comparing interventions with equivalent instructional designs but different levels of immersion and engagement, to determine whether a headset is actually necessary or whether a desktop screen achieves the same result. They urge head-to-head comparisons of VR training with traditional perspective-taking exercises and role-play. Most strikingly, they note that not a single longitudinal study of VR empathy training has been conducted to date, leaving open whether short-term gains persist, whether sustained use amplifies benefits, or whether technology fatigue could eventually reverse them. Accessibility, technological literacy, hardware disparities, data privacy and the psychological risks of intense immersive experiences all demand attention before the technology scales.

For now, the review functions as a roadmap for a market that has outgrown its evidence base. Teachers selecting tools for bullying prevention, health care providers seeking training for dementia care, researchers hunting for replicable instruments and investors sizing up the empathy-training segment all gain a systematic overview of what exists, what it claims and what has actually been tested. The five-category taxonomy, anchored in interactivity and instructional design, gives the field a common vocabulary it has lacked. As socio-emotional skills gain recognition across education and industry, the message from this analysis is clear: virtual reality’s empathy promise is real but unproven, and the next chapter of the story will be written not by better headsets alone, but by studies that finally hold the software to scientific account.

Subject of Research: Virtual reality software for empathy training and its instructional design

Article Title: The use of virtual reality training to enhance empathy: software review

Article References: Khukalenko, I., & Khanolainen, D. (2025). The use of virtual reality training to enhance empathy: software review. Journal of New Approaches in Educational Research, 14(1), Article 14. https://doi.org/10.1007/s44322-025-00034-z

Image Credits: AI Generated

DOI: 10.1007/s44322-025-00034-z

Keywords: virtual reality, empathy training, instructional design, immersive learning, embodiment, perspective-taking, prosocial behaviour, educational technology, healthcare training, software review, embodied cognition, dyslexia awareness