arXiv subjects

Massimo Stella

Publications and source records attributed to Massimo Stella.

At least 19 recordsLinked to original sources

arXiv2026

Natural Language Processing Psychometrics

Natural Language Processing (NLP) models predicting mental health outcomes rarely specify what they measure: contextual knowledge, emotional content, or syntactic structure. NLP Psychometrics treats psychological prediction from text as a psychometric problem, linking scores to interpretable linguistic evidence and testing beyond the training text format. Nine LLMs, conditioned on controlled personas (cognitive digital shadows), completed psychometric questionnaires with textual explanations per item. We extracted emotional profiles and syntactic-semantic structure via textual forma mentis networks, combined with personality and sociodemographic variables in ablated random forest (RF) regressors, using SHAP to identify which features drove performance and in which direction. Full RF models explained up to 70.8% of variance in life satisfaction (SWLS), 55.7% in depression (PHQ-9), and, for DASS-21, 68.5% depression, 76.0% anxiety, 72.4% stress. Sociodemographics alone explained no meaningful variance in depression, anxiety, or stress, but did so for life satisfaction, where emotion features and income were the strongest predictors; neuroticism and network topology instead dominated depression and anxiety, reversing direction between them. Without retraining, RF models separated diaries from low- and high-score personas ($r$ up to 0.91) and, using only network/emotion features, classified clinical from control participants in real transcripts with up to 68% accuracy. These results show the promise and limits of synthetic data: LLM personas can expose model biases, recover patterns consistent with clinical rumination, and support psychometric prediction from human text without a matched questionnaire, but cannot substitute for human validation. NLP Psychometrics makes these distinctions explicit, measurable, and testable through interpretable AI and network/emotional features.

cs.CL↗

arXiv2026

MyMentorLLM: A psychotherapy GenAI environment with multimodal voice/text patients, trainees and experts for deliberate practice

Psychotherapists need repeated training and supervision by experts; however, scalability is problematic. Here we present MyMentorLLM, a multimodal voice- and text-based simulation environment for deliberate practice, used to generate 2,100 complete Cognitive Behavioural Therapy (CBT) training sessions. Each session links a DSM-5-TR-grounded patient (with major depressive, generalised anxiety or borderline personality disorder), a therapist-in-training and an expert supervisor. As an initial implementation, we adopted CBT because its structured procedures and competency-based supervision facilitate standardised simulation and evaluation. Sessions were analysed for emotional dynamics, therapeutic competence and diagnostic accuracy. Simulated patients expressed disorder-congruent emotional profiles, which trainee therapists mirrored as in real human counselling. The quality of supervision differed across LLMs: while most models overestimated trainees' competences, native speech-to-speech was closest to human scores. Supervisors' feedback led to better diagnoses in simulated psychotherapists in 5 out of 7 LLMs, and symptom identification accuracy increased with model size. This work shows that simulation of deliberate practice is possible for CBT training, although patient fidelity, calibration of supervisors, and harmful feedback should be evaluated together.

cs.CL↗

arXiv2026

A Tensor Network Framework for Interpretable Graph Analysis of Brain Networks

Identifying robust neurobiological signatures of brain disorders requires machine learning approaches that combine predictive performance with interpretable representations of feature interactions. Here we introduce a quantum-inspired framework based on tensor network machine learning that learns distributed representations of gray-matter features encoded in a Matrix Product State representation, a variational ansatz originally developed for quantum many-body systems. The trained model is then used not only as a classifier but to extract quantum connected correlations between features, which encode higher-order feature interactions and which we map onto a weighted graph. This construction allows us to track, within a single representation, both: (i) the global spectral properties of the network (capturing collective learning dynamics), and (ii) node-level centrality measures (providing interpretable signatures of individual brain regions). Using repeated train-test sampling schemes, we analyze two classification tasks on structural MRI data as examples of complex brain disorders: healthy controls versus schizophrenia and versus bipolar disorder. Node-level analysis identifies a stable set of gray-matter features, most prominently Heschl gyrus, insular cortex, and frontal regions, that act as hubs across multiple centrality measures and across resamplings. These centralities display lower variability across resamplings than Shapley values, supporting the interpretive value of the network representation. The bipolar feature set emerges as a subset of the schizophrenia one, consistent with the hierarchically organized neuroanatomical alterations reported in neuroimaging studies and offering a network-based characterization of this hierarchy.

q-bio.NC↗

arXiv2026

Introducing multiplex semantic networks as multifaceted representations of creative associative knowledge across multilingual samples

Creativity is a complex cognitive ability that relies on knowledge organisation and retrieval from semantic memory. Yet most research uses a single task to measure it, capturing only a fraction of this complexity. This study investigates multiplex networks - layered semantic networks obtained from six cognitive tasks - as a more comprehensive approach to modelling the associative knowledge underlying creativity. We collected data from N=518 individuals from four countries (Austria, USA, Singapore, Italy). From their responses to verbal fluency, sentence-chain, free association, and narrative writing tasks, we constructed semantic networks and assembled them in a multiplex structure. AI persona-based responses provided a comparison baseline. Structural reducibility analyses showed that different task layers captured distinct, non-redundant information about semantic organisation, supporting the use of multiple tasks over any single one. The networks from high- and low-creative groups remained structurally distinct, while AI-generated networks showed near-identical structures regardless of creativity group. Finally, we used 12 features (network measures, emotional scores, and spreading activation simulations) in a machine learning model using ridge regression to predict individual creativity scores. The combination of structurally similar layers, as identified in the previous stage, improved a proof-of-concept prediction accuracy by 50%. Structural measures showed the highest feature importance, with spreading activation dynamics providing additional predictive power. Together, these findings indicate that multiplex semantic networks capture a richer, cross-cultural picture of associative knowledge underlying creativity. We also release our diverse dataset and code to foster diverse computational approaches within the creativity community.

cs.CL↗

arXiv2026

Math Education Digital Shadows for Investigating Learning with GenAI: Mathematics Performance, Anxiety, and Confidence in LLMs

Understanding the impact of large language models (LLMs) on mathematics education requires data on LLMs' mathematical performance and biases. To this end, we introduce Math Education Digital Shadows (MEDS), a dataset mapping how LLMs reason about mathematics across human- and AI-like personifications. MEDS comprises 28,000 runs from 14 LLMs (i.e., Mistral, Qwen, DeepSeek, IBM Granite, Microsoft Phi, and xAI Grok) generated under human-shadow and AI-assistant conditions. Each record (digital shadow) includes a set of prompts; psychological and sociodemographic metadata; and four mathematics tasks: (i) interviews about relationships with mathematics, (ii) three psychometric questionnaires on mathematics self-efficacy and anxiety, (iii) one cognitive network capturing attitudes towards mathematics, and (iv) 18 high-school mathematics quiz items enriched with reasoning explanations and confidence scores. Analyses of the data show that LLMs exhibit differences in attitudes and performance across human-shadow and AI-assistant modes, including human-like negative attitudes towards mathematics, logical fallacies, and overconfidence in mathematics. As a data resource, MEDS can benefit learning scientists and developers of safer AI tutors in mathematics.

cs.AI↗

arXiv2026

Mapping how LLMs debate societal issues when shadowing human personality traits, sociodemographics and social media behavior

Large Language Models (LLMs) can strongly shape social discourse, yet datasets investigating how LLM outputs vary across controlled social and contextual prompting remain sparse. Cognitive Digital Shadows (CDS) is a 190,000-record synthetic corpus supporting analyses of LLM-generated discourse. Each CDS record is generated by one of 19 LLMs, prompted to shadow either a human persona or an AI-assistant role. CDS contains LLM responses on 4 controversial societal topics: vaccines/healthcare, social media disinformation, the gender gap in science, and STEM stereotypes. Persona-conditioned records encode 17 sociodemographic and psychological attributes, providing data linking LLMs' prompts, language, stances and reasoning. Texts are validated for topic anchoring and can support emotional analyses via interpretable NLP (e.g. textual forma mentis networks). CDS is enriched by a pooling platform with user-friendly dashboards, enabling easy, interactive group-level comparisons of emotional and semantic framing across personas, topics and models. The CDS prompting framework supports future audits of LLMs' bias, social sensitivity and alignment.

cs.CL↗

arXiv2026

The TEA Nets framework combines AI and cognitive network science to model targets, events and actors in text

We introduce Target-Event-Agent Networks (TEA Nets) as a computational framework to extract subjects (``Agents"), verbs (``Events"), and objects (``Targets") from texts. Grounded in cognitive network science and artificial intelligence, TEA Nets are implemented as an open-source Python library. We test TEA Nets in three case studies, demonstrating the framework's ability to perform interpretable emotion detection, semantic frame analyses, and linguistic inquiries across conspiracy texts and textual responses generated by LLMs. In the LOCO conspiracy corpus, TEA Nets revealed that highly conspiratorial narratives (4,227 texts) linked personal pronouns (``I", ``you", ``we") with the same actions twice as frequently as low-similarity conspiracy narratives. High-conspiracy narratives connected person-focused elements (``you", ``people") through actions eliciting anger above the random baseline ($z = 2.63, p < .05$), a trend absent in low-similarity conspiracy narratives, which emphasized scientific actors (``researcher", ``scientist"). In the HOPE and CounseLLMe datasets of 212 (human) and 200 (LLM-based) psychotherapy transcripts, respectively, TEA Nets highlighted emotional differences. When expressing feelings, Claude 3 Haiku, GPT-3.5, and humans used sad words with higher frequency than random expectations but Haiku expressed sadness with lower emotional intensity than humans ($U = 1243.5, p = .036$). We discuss these differences in the context of psychotherapy training on LLM-simulated patients. Our results show that Target-Event-Agent Networks can extract relevant emotional, syntactic, and semantic insights from narratives, opening new avenues for text analysis with cognitive network science.

cs.AI↗

arXiv2026

LLMs can persuade only psychologically susceptible humans on societal issues, via trust in AI and emotional appeals, amid logical fallacies

Scarce longitudinal evidence examines LLMs' persuasiveness and humanness along time-evolving psychological frameworks. We introduce Talk2AI, a longitudinal framework quantifying psycho-social, reasoning and affective dimensions of LLMs' persuasiveness about polarizing societal topics. In a four-way longitudinal setup, Talk2AI's 770 participants engaged in structured conversations with one of four leading LLMs on topics like climate change, social media misinformation, and math anxiety. This produced 3,080 conversations over 60,000 turns. After each wave, participants reported conviction in their initial topic stance, perceived opinion change, LLM's perceived humanness, a self-donation to the topic and a textual explanation. Feedback time series showed longitudinal inertia in convictions, indicating some human anchoring to initial opinions even after repeated exposure to AI-generated arguments. Interestingly, NLP analyses revealed that both humans and LLMs relied on fallacious reasoning in 1 conversational quip every 6, countering the ``LLMs as superior systems" stereotype behind LLMs' cognitive surrender. LLMs' perceived humanness was most learnable from sociodemographic, psychological and engagement features ($R^2=0.44$), followed by opinion change ($R^2=0.34$), conviction ($R^2=0.26$) and personal endowment ($R^2=0.24$). Crucially, explainable AI (XAI) indicated: (i) the presence of individuals more susceptible to LLM-based opinion changes; (ii) psychological susceptibility to LLM-convincing consisted of having more trust in LLMs, being more agreeable and extraverted and with a higher need for cognition. A multiverse approach with mixed-effects models confirmed XAI results, alongside strong individual differences. Talk2AI provides a grounded framework and evidence for detecting how GenAI can influence human opinions via multiple psycho-social pathways in AI-human digital platforms.

cs.AI↗

arXiv2026

The role of System 1 and System 2 semantic memory structure in human and LLM biases

Implicit biases in both humans and large language models (LLMs) pose significant societal risks. Dual process theories propose that biases arise primarily from associative System 1 thinking, while deliberative System 2 thinking mitigates bias, but the cognitive mechanisms that give rise to this phenomenon remain poorly understood. To better understand what underlies this duality in humans, and possibly in LLMs, we model System 1 and System 2 thinking as semantic memory networks with distinct structures, built from comparable datasets generated by both humans and LLMs. We then investigate how these distinct semantic memory structures relate to implicit gender bias using network-based evaluation metrics. We find that semantic memory structures are irreducible only in humans, suggesting that LLMs lack certain types of human-like conceptual knowledge. Moreover, semantic memory structure relates consistently to implicit bias only in humans, with lower levels of bias in System~2 structures. These findings suggest that certain types of conceptual knowledge contribute to bias regulation in humans, but not in LLMs, highlighting fundamental differences between human and machine cognition.

cs.CL↗

arXiv2026

Talk2AI: A Longitudinal Dataset of Human--AI Persuasive Conversations

Talk2AI is a large-scale longitudinal dataset of 3,080 conversations (totaling 30,800 turns) between human participants and Large Language Models (LLMs), designed to support research on persuasion, opinion change, and human-AI interaction. The corpus was collected from 770 profiled Italian adults across four weekly sessions in Spring 2025, using a within-subject design in which each participant conversed with a single model (GPT-4o, Claude Sonnet 3.7, DeepSeek-chat V3, or Mistral Large) on three socially relevant topics: climate change, math anxiety, and health misinformation. Each conversation is linked to rich contextual data, including sociodemographic characteristics and psychometric profiles. After each session, participants reported on opinion change, conviction stability, perceived humanness of the AI, and behavioral intentions, enabling fine-grained longitudinal analysis of how AI-mediated dialogue shapes beliefs and attitudes over time.

cs.HC↗

arXiv2026

Cognitive networks reconstruct mindsets about STEM subjects and educational contexts in almost 1000 high-schoolers, University students and LLM-based digital twins

Attitudes toward STEM develop from the interaction of conceptual knowledge, educational experiences, and affect. Here we use cognitive network science to reconstruct group mindsets as behavioural forma mentis networks (BFMNs). In this case, nodes are cue words and free associations, edges are empirical associative links, and each concept is annotated with perceived valence. We analyse BFMNs from N = 994 observations spanning high school students, university students, and early-career STEM experts, alongside LLM (GPT-oss) "digital twins" prompted to emulate comparable profiles. Focusing also on semantic neighbourhoods ("frames") around key target concepts (e.g., STEM subjects or educational actors/places), we quantify frames in terms of valence auras, emotional profiles, network overlap (Jaccard similarity), and concreteness relative to null baselines. Across student groups, science and research are consistently framed positively, while their core quantitative subjects (mathematics and statistics) exhibit more negative and anxiety related auras, amplified in higher math-anxiety subgroups, evidencing a STEM-science cognitive and emotional dissonance. High-anxiety frames are also less concrete than chance, suggesting more abstract and decontextualised representations of threatening quantitative domains. Human networks show greater overlapping between mathematics and anxiety than GPT-oss. The results highlight how BFMNs capture cognitive-affective signatures of mindsets towards the target domains and indicate that LLM-based digital twins approximate cultural attitudes but miss key context-sensitive, experience-based components relevant to replicate human educational anxiety.

cs.CL↗

arXiv2026

Complex networks map test anxiety and wellbeing levels in students and ChatGPT

Academic STEM evaluation can elicit anxiety, yet routine grading rarely captures how students semantically frame exams and wellbeing. We reconstruct these framings using behavioural forma mentis networks (BFMNs), that is, feature-rich networks of concepts linked by memory recalls and enriched with affective ratings and concreteness norms. We build BFMNs from 994 participants spanning STEM experts (N1 = 59), Italian high-schoolers (N2 = 206), physics undergraduates (N3 = 10), psychology undergraduates with math-anxiety levels (N4 = 301), and simulated students (N5 = 497) personified by a large language model (GPT-OSS 20B). Across all human groups, the concepts "exam" and "grade" were (i) perceived negatively, (ii) connected primarily to negatively valenced memory recalls, indicating a clustering of negative emotions around assessment, and (iii) framed through concepts eliciting fear and anticipation in most groups, including physics undergraduates (z-scores in the range [2.04, 2.53]). The semantic neighbourhoods of "anxiety" and "exam" overlapped three times more in human students than in GPT-based simulations, providing structural evidence of test anxiety in student populations. By contrast, experts displayed a neutral and more concrete network neighbourhood for "exam" (z = 1.87), with no clear trace of test anxiety. These negative assessment framings coexisted with positive representations of "wellbeing", which were rich in concrete associations in humans but linked to more abstract concepts in GPT digital twins. Overall, our results show that BFMNs offer a quantitative and interpretable framework to study academic anxiety and to distinguish human affective framing from current AI-based simulations

physics.ed-ph↗

arXiv2026

How to predict creativity ratings from written narratives: A comparison of co-occurrence and textual forma mentis networks

This tutorial paper provides a step-by-step workflow for building and analysing semantic networks from short creative texts. We introduce and compare two widely used text-to-network approaches: word co-occurrence networks and textual forma mentis networks (TFMNs). We also demonstrate how they can be used in machine learning to predict human creativity ratings. Using a corpus of 1029 short stories, we guide readers through text preprocessing, network construction, feature extraction (structural measures, spreading-activation indices, and emotion scores), and application of regression models. We evaluate how network-construction choices influence both network topology and predictive performance. Across all modelling settings, TFMNs consistently outperformed co-occurrence networks through lower prediction errors (best MAE = 0.581 for TFMN, vs 0.592 for co-occurrence with window size 3). Network-structural features dominated predictive performance (MAE = 0.591 for TFMN), whereas emotion features performed worse (MAE = 0.711 for TFMN) and spreading-activation measures contributed little (MAE = 0.788 for TFMN). This paper offers practical guidance for researchers interested in applying network-based methods for cognitive fields like creativity research. we show when syntactic networks are preferable to surface co-occurrence models, and provide an open, reproducible workflow accessible to newcomers in the field, while also offering deeper methodological insight for experienced researchers.

cs.CL↗

arXiv2025

Math anxiety and associative knowledge structure are entwined in psychology students but not in Large Language Models like GPT-3.5 and GPT-4o

Math anxiety poses significant challenges for university psychology students, affecting their career choices and overall well-being. This study employs a framework based on behavioural forma mentis networks (i.e. cognitive models that map how individuals structure their associative knowledge and emotional perceptions of concepts) to explore individual and group differences in the perception and association of concepts related to math and anxiety. We conducted 4 experiments involving psychology undergraduates from 2 samples (n1 = 70, n2 = 57) compared against GPT-simulated students (GPT-3.5: n2 = 300; GPT-4o: n4 = 300). Experiments 1, 2, and 3 employ individual-level network features to predict psychometric scores for math anxiety and its facets (observational, social and evaluational) from the Math Anxiety Scale. Experiment 4 focuses on group-level perceptions extracted from human students, GPT-3.5 and GPT-4o's networks. Results indicate that, in students, positive valence ratings and higher network degree for "anxiety", together with negative ratings for "math", can predict higher total and evaluative math anxiety. In contrast, these models do not work on GPT-based data because of differences in simulated networks and psychometric scores compared to humans. These results were also reconciled with differences found in the ways that high/low subgroups of simulated and real students framed semantically and emotionally STEM concepts. High math-anxiety students collectively framed "anxiety" in an emotionally polarising way, absent in the negative perception of low math-anxiety students. "Science" was rated positively, but contrasted against the negative perception of "math". These findings underscore the importance of understanding concept perception and associations in managing students' math anxiety.

cs.CL↗

arXiv2025

A word association network methodology for evaluating implicit biases in LLMs compared to humans

As Large language models (LLMs) become increasingly integrated into our lives, their inherent social biases remain a pressing concern. Detecting and evaluating these biases can be challenging because they are often implicit rather than explicit in nature, so developing evaluation methods that assess the implicit knowledge representations of LLMs is essential. We present a novel word association network methodology for evaluating implicit biases in LLMs based on simulating semantic priming within LLM-generated word association networks. Our prompt-based approach taps into the implicit relational structures encoded in LLMs, providing both quantitative and qualitative assessments of bias. Unlike most prompt-based evaluation methods, our method enables direct comparisons between various LLMs and humans, providing a valuable point of reference and offering new insights into the alignment of LLMs with human cognition. To demonstrate the utility of our methodology, we apply it to both humans and several widely used LLMs to investigate social biases related to gender, religion, ethnicity, sexual orientation, and political party. Our results reveal both convergences and divergences between LLM and human biases, providing new perspectives on the potential risks of using LLMs. Our methodology contributes to a systematic, scalable, and generalizable framework for evaluating and comparing biases across multiple LLMs and humans, advancing the goal of transparent and socially responsible language technologies.

cs.CL↗

arXiv2025

SpreadPy: A Python tool for modelling spreading activation and superdiffusion in cognitive multiplex networks

We introduce SpreadPy as a Python library for simulating spreading activation in cognitive single-layer and multiplex networks. Our tool is designed to perform numerical simulations testing structure-function relationships in cognitive processes. By comparing simulation results with grounded theories in knowledge modelling, SpreadPy enables systematic investigations of how activation dynamics reflect cognitive, psychological and clinical phenomena. We demonstrate the library's utility through three case studies: (1) Spreading activation on associative knowledge networks distinguishes students with high versus low math anxiety, revealing anxiety-related structural differences in conceptual organization; (2) Simulations of a creativity task show that activation trajectories vary with task difficulty, exposing how cognitive load modulates lexical access; (3) In individuals with aphasia, simulated activation patterns on lexical networks correlate with empirical error types (semantic vs. phonological) during picture-naming tasks, linking network structure to clinical impairments. SpreadPy's flexible framework allows researchers to model these processes using empirically derived or theoretical networks, providing mechanistic insights into individual differences and cognitive impairments. The library is openly available, supporting reproducible research in psychology, neuroscience, and education research.

cs.CL↗

arXiv2025

Textual forma mentis networks bridge language structure, emotional content and psychopathology levels in adolescents

We introduce a network-based AI framework for predicting dimensions of psychopathology in adolescents using natural language. We focused on data capturing psychometric scores of social maladjustment, internalizing behaviors, and neurodevelopmental risk, assessed in 232 adolescents from the Healthy Brain Network. This dataset included structured interviews in which adolescents discussed a common emotion-inducing topic. To model conceptual associations within these interviews, we applied textual forma mentis networks (TFMNs)-a cognitive/AI approach integrating syntactic, semantic, and emotional word-word associations in language. From TFMNs, we extracted network features (semantic/syntactic structure) and emotional profiles to serve as predictors of latent psychopathology factor scores. Using Random Forest and XGBoost regression models, we found significant associations between language-derived features and clinical scores: social maladjustment (r = 0.37, p < .01), specific internalizing behaviors (r = 0.33, p < .05), and neurodevelopmental risk (r = 0.34, p < .05). Explainable AI analysis using SHAP values revealed that higher modularity and a pronounced core-periphery network structure-reflecting clustered conceptual organization in language-predicted increased social maladjustment. Internalizing scores were positively associated with higher betweenness centrality and stronger expressions of disgust, suggesting a linguistic signature of rumination. In contrast, neurodevelopmental risk was inversely related to local efficiency in syntactic/semantic networks, indicating disrupted conceptual integration. These findings demonstrated the potential of cognitive network approaches to capture meaningful links between psychopathology and language use in adolescents.

cs.CY↗

arXiv2025

A network psychometric analysis of maths anxiety factors in Italian psychology students

Dealing with mathematics can induce significant anxiety, strongly affecting psychology students' academic performance and career prospects. This phenomenon is known as maths anxiety and several scales can measure it. Most scales were created in English and abbreviated versions were translated and validated among Italian populations (e.g. Abbreviated Maths Anxiety Scale). This study translated the 3-factor MAS-UK scale in Italian to produce a new tool, MAS-IT, validated specifically in a sample of Italian undergraduates enrolled in psychology or related BSc programmes. A sample of 324 Italian undergraduates completed the MAS-IT. The data were analysed using confirmatory Factor Analysis (CFA), testing the original MAS-UK 3-factor model. CFA results revealed that the original MAS-UK 3-factor model did not fit the Italian data. A subsequent Exploratory Graph Analysis (EGA) identified 4 distinct components/factors of maths anxiety detected by MAS-IT. The items relative to "Passive Observation maths anxiety" factor remained stable across the analyses, whereas "Evaluation maths anxiety" and "Everyday/Social maths anxiety" items showed a reduced or poor item stability. Quantitative findings indicated potential cultural or contextual differences in the expression of maths anxiety in today's psychology undergraduates, underlining the need for more appropriate tools to be used among psychology students.

stat.AP↗