BRAVE Lab’s PI, Alberto Greco

The activities of BRAVE Lab follow a coherent scientific and technological pathway.

We begin with the design of innovative experimental protocols, move to the acquisition and modelling of physiological signals, use these signals to infer emotional states, extend the analysis to interpersonal physiological coupling, and ultimately translate these models into adaptive systems and biofeedback technologies.

This integrated approach allows us not only to measure and understand human affective dynamics, but also to design systems that can positively influence social interaction, emotional regulation, and well-being.

Research question

Understanding and recognising human emotion objectively and continuously remains one of the major challenges. Despite advances in psychophysiology and artificial intelligence, we still lack reliable tools to measure how internal emotional states unfold over time and shape human experience.

BRAVE Lab’s activity

At BRAVE Lab, we study emotion as a measurable physiological process. By combining wearable sensing, computational modelling, and adaptive technologies, we develop systems capable of estimating, reproducing, and modulating internal states directly from biosignals.

What we are building

Building on this foundation, we extend our research to how emotional and physiological dynamics unfold during social interaction, investigating how individuals influence each other and how this knowledge can be translated into technologies that enhance well-being and human experience.

Our research pathway is articulated through four tightly connected research areas ↓

Physiological Signal Processing & Modelling

At BRAVE Lab, physiological signals are treated as rich, dynamic representations of the human internal state.

We work with multimodal recordings including ECG (from which heart rate variability is derived), EDA (electrodermal activity), respiration, and EEG, acquired through wearable and unobtrusive sensing systems. Our activity focuses on advanced preprocessing, feature extraction, and time-series modelling to transform raw biosignals into informative variables that reflect autonomic nervous system (ANS) activity. We develop statistical, machine learning, and Bayesian models capable of capturing the temporal structure and variability of physiological processes.

This foundational work enables the translation of biological dynamics into quantitative representations of latent states that can be used for emotion estimation, synchronization analysis, and real-time adaptive technologies.

Here is our most representative paper about this research line

Read it online Download it!

Affective Computing

BRAVE Lab develops computational AI models to infer emotional states from psychophysiological data.

Rather than relying solely on self-report or behavioral observation, we estimate continuous dimensions such as arousal, valence, stress, and anxiety using probabilistic, machine learning, and deep learning approaches.

Our research focuses on identifying the physiological signatures that underlie emotional dynamics and on building models that can operate also in real time. This enables the reconstruction of an individual’s internal emotional trajectory during interaction or exposure to complex environments.

From our perspective, affective computing represents the bridge between physiological signals and interpretable emotional representations that can inform adaptive systems, clinical tools, and human–AI interaction.

Read it online Download it!

Adaptive Systems, Virtual Reality and Biofeedback

A central research theme at BRAVE Lab is the translation of computational models of physiology and emotion into adaptive technologies that interact with users in real time.

We develop biofeedback systems and virtual reality environments capable of adapting their behaviour according to the user’s estimated internal state. These systems are applied to scenarios such as social anxiety, phobias, therapeutic support, human–AI interaction, and social training.

By closing the loop between sensing, modelling, and actuation, we create environments that respond dynamically to the individual. This approach allows us not only to measure emotional processes, but also to influence and guide them through immersive, interactive technologies.

Here is our most representative paper about this research line

Read it online Download it!

Interpersonal Physiological Coupling

During social interaction, human physiological systems tend to align and influence each other in subtle but measurable ways.

At BRAVE Lab, we study this phenomenon as interpersonal physiological coupling, a process through which people affect each other emotionally and biologically. Using hyperscanning approaches and multimodal sensing, we analyse synchrony, directionality, and temporal alignment across heart rate variability, electrodermal activity, respiration, and neural signals. This work supports the idea that social connection can be described as a dynamical coupling between physiological processes.

Our models aim to quantify emotional contagion, bonding, and social coordination, providing an objective framework to study how humans connect beyond observable behaviour.

Read it online Download it!

Greco Alberto

Research Center "E.Piaggio", School of Engineering, University of Pisa, Italy

Verified email at centropiaggio.unipi.it

Biomedical System and Signal Processing Physiological models Wearable monitoring systems Electrodermal activity

A Deep Learning approach for estimating Time-Domain Heart Rate Variability parameters from wrist photoplethysmography during daily activity

G Rho, N Carbonaro, M Laurino, A Tognetti, A Greco Biomedical Signal Processing and Control 123, 110531, 2026

AbstractHeart rate variability (HRV) is a critical indicator of autonomic nervous system regulation and cardiovascular health, typically measured using electrocardiography (ECG). Wrist devices are gaining popularity as non-invasive alternatives to monitor heart rate (HR) and pulse rate variability (PRV) in unconstrained setting through photoplethysmography (PPG). However, movement artifacts severely deteriorate signal quality, making estimation reliability...

Multimodal Classification of Social Anxiety Using Continuous Self-Rating and Physiological Data in Virtual Reality

M Pardini, S Frumento, M Martini, M Alaimo, G Rho, N Paparo, ... IEEE Transactions on Affective Computing, 2026

Assessment of Social Anxiety Disorder (SAD) is limited by paradigms that often target performance-related fears and its reliance on static self-reports that fail to capture the dynamic, in-the-moment nature of anxiety. This study introduces a novel Virtual Reality (VR) framework to assess SAD that integrates continuous self-ratings of anxiety with objective physiological data. Sixty-three participants...

Emotional body odors alter autonomic nervous activity during mindfulness training in social anxiety

E Dal Bò, C Cecchetto, L Zurlo, L Lavezzo, ET Eliasson, E Vigna, ... Journal of Anxiety Disorders, 103201, 2026

Chemosignals found in human body odors (BOs) convey information about a person’s emotional state and elicit observable responses in others. While mindfulness typically enhances parasympathetic tone and reduces anxiety, little is known about how BOs modulate these effects. This study investigated how emotional and non-emotional human BOs influence autonomic nervous system activity – heart rate...

The volatile fatty-acid fingerprint of human fear

T Bruderer, A Greco, M Ripszam, AL Callara, A Gargano, S Reale, ... bioRxiv, 2026.05. 11.724357, 2026

Whether humans communicate fear through volatile chemical cues has remained unresolved, despite decades of behavioural and neuroimaging evidence. Here, we identify a reproducible molecular signature of acute human fear, bridging the gap between functional evidence and chemical mechanism. Using a low-background sampling approach combined with multiscale chemical analysis and hierarchical modelling, we link high-dimensional axillary...

Emotional Responses to AI Use: Development of the SAG-AI Scale for Shame and Guilt

E Cipriani, D Menicucci, A Greco, S Grassini International Journal of Human–Computer Interaction, 1-20, 2026

The rise of generative AI tools has introduced new ethical dilemmas and emotional responses, yet little research has examined users’ social emotions such as shame, guilt, and impostor syndrome. This article presents the SAG-AI (Shame and Guilt related to AI tools) scale, a 9-item psychometric instrument developed to measure these emotions in AI tool usage....

View All Publications

We will reply as soon as possible

Contact us

Feel free to reach us for any need – whenever you want to (freely!) use the tools we developed, or you want us to share with you our data, or you want to interview us