In an era where digital interaction is rapidly supplanting physical presence, understanding how avatars influence human behavior has become more crucial than ever. A groundbreaking study recently published in PLOS Biology sheds light on an intriguing phenomenon: people are more prone to take risks in gambling tasks when they receive feedback from an avatar rather than a real human counterpart. This finding not only challenges assumptions about virtual communication but also uncovers the intricate neural dynamics that underlie decision-making processes influenced by digital representations.
Researchers Toshiko Tanaka and Masahiko Haruno from Japan’s National Institute of Information and Communications Technology spearheaded this comprehensive investigation into how avatars modulate risk-taking behavior. Utilizing a carefully designed gambling task, they explored the impact of receiving dynamic facial feedback from either a real human or that same individual’s avatar counterpart. Participants were instructed to choose between a safe certain option and a probabilistic gamble, while continuously observing an observer’s face that mirrored either the actual human or their avatar version expressing win/no-win outcomes.
What emerged was a compelling pattern: participants exhibited increased gambling tendencies in the avatar condition compared to when observing the real human face. To quantify and decode this behavioral shift, the team employed state-of-the-art functional magnetic resonance imaging (fMRI) and computational modeling techniques. They found that the amygdala, a brain region long associated with emotional processing and risk assessment, was central to this augmented risk-taking behavior. Specifically, the amygdala’s response to uncertainty in the feedback—whether it be facial expressions signaling success or failure—was modulated when participants engaged with avatars.
The computational models used in this study dissected participants’ decision-making strategies, revealing that in the avatar condition, there was an amplified valuation of the uncertainty inherent in facial-expression feedback. In other words, participants interpreted the uncertain social cues from avatars differently, leading to a greater propensity for risk. This finding bridges a critical gap between behavioral economics and social neuroscience by demonstrating how virtual representations can recalibrate fundamental cognitive processes involved in valuing uncertainty.
Importantly, the data showed that this amygdala response to feedback uncertainty was negatively correlated with risk-taking: lower amygdala activation corresponded with higher gambling rates. This suggests that individuals may suppress or reinterpret the emotional signals emanating from avatar expressions, thus altering their natural caution toward uncertain outcomes. This neural modulation was consistent across both avatar and human conditions, which implies that while avatars intensify the effect, the underlying brain mechanisms remain fundamentally the same.
The investigation extended beyond group averages, uncovering substantial individual variability in sensitivity to feedback uncertainty. The researchers linked this variability to personality traits associated with emotional empathy and interpersonal consideration. Participants scoring higher on measures reflecting emotional responsiveness toward others demonstrated different patterns of neural and behavioral responses. This nuanced connection highlights the complex interplay between personality, brain function, and digital social interaction.
Such insights are particularly pertinent given the explosion of avatar-based communication platforms and virtual reality environments in everyday life. As society increasingly integrates avatars for work, socializing, and entertainment, understanding how these digital intermediaries affect our choices and emotional processing becomes vital. The study by Tanaka and Haruno provides empirical evidence that the “face” reacting to us—even if it’s a digital mimic—significantly shapes how we evaluate risk and uncertainty.
The methodology underpinning this study was rigorous. Twenty-eight participants engaged in behavioral experiments while fifty-one additional individuals underwent fMRI scanning, ensuring robust validation of both psychological and neural effects. The avatars were carefully programmed to mirror real human observers’ expressions in real-time, preserving ecological validity. This meticulous design allowed the team to isolate the influence of the avatar’s form without confounds related to differing feedback content.
Researchers were candid about the challenges encountered during the study. Tanaka noted the complexity involved in maintaining the illusion of a seamless human-avatar interaction, emphasizing the meticulous effort required to keep participants believing that avatars and humans provided genuinely dynamic responses. Such attention to detail underscores the growing necessity for experimental designs that faithfully replicate the immersive experiences users have in virtual settings.
From a neuroscientific perspective, the demonstration that the amygdala mediates sensitivity to avatar feedback uncertainty provides a mechanistic explanation for how virtual social cues alter human behavior. It points to a specific neural circuit through which avatars can modulate emotional responses and decision-making processes—a finding that blends computational modeling with brain imaging in an elegant, translational manner.
Furthermore, the broader implications extend to domains such as online gambling, virtual teamwork, and even mental health interventions where avatar interactions may subtly influence risk perception and choices. Understanding that avatars can unconsciously nudge users toward greater risk-taking opens avenues for designing safer and more effective virtual systems by leveraging or mitigating these neural effects.
Haruno reflects on these transformative insights with a reminder of their real-world importance: “People tend to take more risks when their partner responds through an avatar rather than showing their real face. This seems to be driven by a change in how they process uncertainty—and interestingly, that change is reflected in the amygdala.” His statement underscores the fundamental shift in human cognition as we transition toward increasingly avatar-mediated communications.
The study’s open-access publication in PLOS Biology ensures that these findings are widely accessible to scientists, policymakers, and technologists eager to grasp and harness the nuances of human-avatar interaction. As avatars become omnipresent, deciphering the subtle but powerful influences they exert on our behavior will be a cornerstone of future cognitive neuroscience and behavioral science research.
In conclusion, this pioneering research by Tanaka and Haruno elucidates a fascinating aspect of modern social interaction: avatars, while virtual and mediated, exert tangible impacts on our risk-taking behaviors through neural pathways involving the amygdala’s processing of feedback uncertainty. Their multidisciplinary approach, combining behavioral tasks, neuroimaging, and computational theory, charts a path forward for investigating the complex psychology of digital embodiment and virtual communication—a frontier that is unleashing profound changes in how humans connect, decide, and respond.
Subject of Research: People
Article Title: Feedback from an avatar facilitates risk-taking by modulating the amygdala response to feedback uncertainty
News Publication Date: Unknown (study published April 22, 2025)
Web References:
- DOI link: 10.1371/journal.pbio.3003122
- JST Moonshot R&D: https://projectdb.jst.go.jp/grant/JST-PROJECT-20339237/
- JST CREST: https://projectdb.jst.go.jp/grant/JST-PROJECT-22712882/
- KAKENHI: https://kaken.nii.ac.jp/ja/grant/KAKENHI-PLANNED-22H05155/
References:
Tanaka T, Haruno M (2025) Feedback from an avatar facilitates risk-taking by modulating the amygdala response to feedback uncertainty. PLoS Biol 23(4): e3003122.
Image Credits: Toshiko Tanaka (CC-BY 4.0)
Keywords: Avatar, risk-taking, amygdala, feedback uncertainty, virtual reality, social neuroscience, computational modeling, fMRI, decision-making, emotional processing, human-computer interaction
