Exploring the Role of Theory of Mind in Human Decision Making: Cognitive, Spatial, and Emotional Influences in the Adversarial Rock-Paper-Scissors Game
Thuy Ngoc Nguyen, Jeffrey Flagg, Cleotilde Gonzalez · 2025
Read Exploring the Role of Theory of Mind in Human Decision Making: Cognitive, Spatial, and Emotional Influences in the Adversarial Rock-Paper-Scissors Game by Thuy Ngoc Nguyen, Jeffrey Flagg, Cleotilde Gonzalez (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-05699.txt.
Introduction
Thuy Ngoc Nguyen, Jeffrey Flagg, and Cleotilde Gonzalez study how reading other minds shapes play in rock-paper-scissors. Theory of mind, the skill of tracking others' beliefs and aims, matters when people face off against people or machines. How its parts interact during such contests was unclear.
Players took standard theory-of-mind tests, then played repeated games against bots and humans. The team measured how reasoning, spatial skill, and emotion reading related to performance. People did spot patterns in opponents' moves. Yet most single test scores failed to predict who played well. Recursive thinking, thinking about what the other thinks, was the one moderate predictor. Modeling split the influences into two hidden factors with opposite effects.
Catalog record
| CREATOR | Thuy Ngoc Nguyen, Jeffrey Flagg, Cleotilde Gonzalez |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-05699
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Exploring the Role of Theory of Mind in Human Decision Making: Cognitive, Spatial, and Emotional Influences in the Adversarial Rock-Paper-Scissors Game Thuy Ngoc Nguyen, Jeffrey Flagg, Cleotilde Gonzalez Abstract Understanding how humans attribute beliefs, goals, and intentions to others—known as theory of mind (ToM)— is critical in the context of human-computer interaction. Despite various metrics used to assess ToM, the interplay between cognitive, spatial, and emotional factors in influencing human decision making during adversarial interactions remains underexplored. This paper investigates these relationships using the Rock-Paper-Scissors (RPS) game as a testbed. Through established ToM tests, we analyze how cognitive reasoning, spatial awareness, and emotional perceptiveness affect human performance when interacting with bots and human opponents in repeated RPS settings. Our findings reveal significant correlations among certain ToM metrics and highlight humans’ ability to detect patterns in opponents’ actions. However, most individual ToM metrics proved insufficient for predicting performance variations, with recursive thinking being the only metric moderately associated with decision effectiveness. Through exploratory factor analysis (EFA) and structural equation modeling (SEM), we identified two latent factors influencing decision effectiveness: Factor 1, characterized by recursive thinking, emotional perceptiveness, and spatial reasoning, positively affects decision-making against dynamic bots and human players, while Factor 2, linked to interpersonal skills and rational ability, has a negative impact. These insights lay the groundwork for further research on ToM metrics and for designing more intuitive, adaptive systems that better anticipate and adapt to human behavior, ultimately enhancing human-machine collaboration. Keywords: Theory of mind; Human decision making; Human-AI interaction; Repeated adversarial games; Rock paper scissors 1 Introduction Theory of Mind (ToM) refers to the ability to understand and infer the mental states of others, including their knowledge, emotions, and intentions, which are crucial for human social interactions. This capability develops at an early stage in life and is applicable in various situations. ToM has philosophical and psychological foundations and has been the subject of extensive research [ 71 , 72 ] . Despite extensive research, there remains a lack of consensus on the precise definition and measurement of ToM [ 6 , 27 , 12 , 64 , 72 ] . This gap in understanding is crucial, as accurate measurement of ToM is essential to advance research in both human and artificial intelligence (AI) contexts, where improving interactions between AI systems and humans is increasingly important [ 48 , 80 ] . Research on ToM ability in humans has progressed markedly, including neuroimaging studies that have identified specific brain regions involved in ToM activation [ 74 ] and investigations across various vulnerable groups such as children, individuals with autism, dementia, brain injuries, and schizophrenia, all of which demonstrated ToM impairments [ 85 , 76 , 19 , 5 , 32 ] . These studies highlight the importance of ToM in understanding cognitive and social functions and emphasize the need for clear and precise methods of measuring it. The operationalization of ToM has been long debated, particularly regarding the distinction between explicit and implicit processing of mental states [ 34 , 3 ] . Cognitive components play a prominent role in explicit ToM processing, as studies have shown that performance in ToM tasks is impaired when cognitive resources are taxed [ 4 ] . Moreover, many widely used ToM tests for humans focus on the attribution of beliefs through recursive thinking [ 60 ] , understanding false beliefs [ 9 ] , and strange stories [ 41 ] ; all require significant cognitive processing [ 84 , 80 , 40 ] . In contrast, facets related to empathy and emotion recognition are often considered implicit processing and distinct from ToM. Neuroimaging studies show that these emotional processes activate different brain regions than those used for cognitive ToM tasks [ 75 , 73 ] . Conversely, other studies suggest that social perception, action observation, and cognitive processing are connected through the posterior superior sulcus, promoting functional connectivity between these components [ 88 ] . Although fMRI studies highlight regional differences between cognitive, emotional, and spatial tasks [ 72 ] , they reveal significant overlap in brain activation during ToM tasks [ 24 , 17 ] . A meta-analysis confirms this overlap and specialization in brain functioning during ToM tasks [ 74 ] . Despite the mixed results surrounding the operationalization of ToM, recent studies on ToM abilities in humans indicate that while cognitive evaluations are important, they primarily focus on particular aspects of ToM and may not fully assess all relevant mental states, including spatial reasoning and emotional perceptiveness [ 77 , 70 , 51 ] . Previous research has highlighted the role of affective perspective-taking, which involves understanding emotions and interpersonal perceptions that influence individual beliefs and behaviors [ 13 , 28 , 63 ] . Furthermore, evidence shows that spatial reasoning is integral to ToM, especially in traditional abstract reasoning tests [ 84 ] . This type of reasoning, particularly motion perception, is associated with ToM development in children [ 67 , 54 ] . In adults, engaging with others’ visuospatial perspectives activates specific areas of the brain during ToM processing, unlike when focusing solely on one’s self-perspective [ 38 ] . In addition, understanding spatial relationships is essential for planning actions based on the intentions of another agent [ 43 ] . These insights highlight the interaction among the cognitive, emotional, and spatial components for understanding and interacting with others. In terms of applications, numerous studies have demonstrated the role of ToM capabilities in predicting others’ actions, strategic planning, and various aspects of social reasoning and decision making [ 43 , 62 , 69 , 7 , 50 ] . These abilities are crucial in predicting competitive behaviors and strategies within adversarial settings, which have wide-ranging applications in security and defense [ 30 ] . However, applying ToM reasoning in such contexts, where one person seeks to gain at another’s expense, is particularly challenging [ 21 , 91 , 15 ] . Specifically, the complexity of infinitely recursive guessing about an opponent’s beliefs necessitates an exploration of ToM’s role in adversarial behavior [ 33 ] . Importantly, previous studies have emphasized the adaptability of human decision behavior in response to different opponent strategies [ 55 ] , but it is not yet clear how these adaptive abilities are associated with ToM. In general, the literature indicates that cognitive, social, emotional, and spatial processes are crucial for a robust ability to understand the beliefs, desires, and actions of others. However, little is currently understood about the relationship between these elements, whether they are part of a single construct or if individual elements can predict human behavior in situations requiring ToM. To this end, our work investigates how these cognitive, spatial, and emotional perceptiveness metrics impact human decision making in human-AI interactions where people were paired with different bots and human opponents during repeated Rock-Paper-Scissors (RPS) games with imperfect information. The repeated RPS game, a mixed strategy equilibrium scenario, provides an ideal paradigm for studying how people adapt to an opponent’s sequential actions [ 14 , 37 ] . Our research sheds light on the interplay among the cognitive, emotional, and spatial components of ToM. We also explore their interconnected aspects in predicting human decision behavior in adversarial settings,