Work
Since December 2022
The research focused on detecting "windows of responsiveness" in Minimally Conscious State
patients, using Artificial Intelligence models.
Since September 2022
Teacher in the Data Analytics course, I am in charge of Statistics, SQL and Python & ML modules.
Since June 2022
I am engaged in the study and implementation of Artificial Intelligence models for creating
algorithms that can improve indoor environmental quality (IEQ) and energy efficiency of
buildings using IoT sensors.
External Data Scientist @P4Future
May 2022 - July 2022
I contributed to the creation of neural network algorithms to identify and classify fashion
objects.
Independent Consultant
Since 2022
I develop artificial intelligence, statistical modeling and data science solutions for industrial,
sustainability and public-sector projects. My consulting work connects predictive analytics and
technical teaching with the practical implementation of reliable machine learning systems.
External Data Scientist @BrainSigns
September 2020 - August 2022
I finalized research projects that stemmed from my master's thesis on Neuroscience, by taking
part in the research and development of models and methodologies for the classification of
mental states. I am currently collaborating on the creation of a portable device able to
classify mental states in different contexts, using machine learning algorithms.
Senior Data Scientist @Be
October 2020 - January 2022
I managed and lead a team of Data Scientists, and supervised the development of Machine and Deep
Learning models. I collaborated in several projects and independently carried out a project on
behavioural analysis, habits and points of interest of drivers for an insurance group.
September 2019 - July 2020
I carried out thesis research aimed at the analysis and creation of Machine Learning methods for
the classification of mental states (such as Stress, Workload and Vigilance) through the
analysis of bioelectric signals such as EEG, ECG and GSR.
Senior Data Scientist @Jakala
September 2018 - September 2020
I was Referent for the realisation of an internal Project focused on the creation of a data
platform, in order to enrich the partners' customer data in an anonymous way using Statistical
Matching functions and Advanced Data Anonymization and Privacy Tool.
April 2015 - August 2018
As part of a team of 5 people, I participated in the realisation of 3 projects, each lasting one
year. In the first project, for a banking group, I was responsible for creating a loyalty model
for classifying users using the Cloudera platform with Spark Scala. In the second Project,
Energy Group, I was in charge of creating Python scripts for the transformation of data included
in an ETL process. In the third Project, Telecommunications Group, I worked on the creation of
python scripts for
data analysis and the creation of churn models on Google cloud Platform technology.
Academic
Since November 2023
My PhD research in bioengineering focuses on the implementation of artificial intelligence models for biosignal analysis in the context of passive brain-computer interfaces (pBCI). The main objective is the development of artificial intelligence models that enable continuous interaction and analysis with brain signals, under different working conditions, especially in out-of-laboratory contexts.
September 2015 - July 2020
The Master's programme in Data Science provided me with an interdisciplinary approach that
incorporates contributions from Computer Science, Statistics, Bioinformatics, and Neuroscience.
Through this programme, I gained the necessary professional knowledge to develop big data
collection, management, processing, and analysis technologies, and translate this data into
valuable insights for innovative business and social sectors. In addition, elective courses such
as Bioinformatics, Digital Epidemiology, and Neuroscience provided me with a better
understanding of how Data Science can be applied in the medical field. Overall, this programme
equipped me with a broad range of skills and knowledge to succeed in the dynamic and rapidly
evolving field of Data Science.
Master in Big Data & Business Intelligence @BIOGEM
January 2015 - April 2015
The UIIP course provided me with a comprehensive set of skills and competencies in the field of
ICT. Through this program, I gained expertise in the latest technologies and trends in the
industry, including data analytics, machine learning, and software development. The program also
equipped me with valuable project management skills, enabling me to effectively lead teams and
oversee complex projects. Additionally, I developed strong problem-solving and critical thinking
skills, which allow me to analyze complex situations and identify optimal solutions. Finally, I
gained a deep understanding of the importance of collaboration and effective communication in
the field of ICT, which is crucial for success in today's fast-paced and interconnected world.
Overall, the UIIP program has provided me with the skills and competencies necessary to succeed
in the dynamic and rapidly evolving field of ICT.
September 2011 - October 2014
My Bachelor's degree in Statistics for Business has equipped me with a strong foundation in
methodological skills in Statistics and Actuarial Science. The program's focus on Statistical
Science has also provided me with a comprehensive understanding of Computational, Data
Managment, and Algorithmic skills, as well as Financial and Insurance skills. Through this
program, I gained a deep understanding of the underlying theories and principles of statistical
analysis, and developed proficiency in advanced statistical techniques such as regression
analysis, time series analysis, and hypothesis testing. Furthermore, I honed my skills in data
management, visualization, and interpretation, which are essential for generating actionable
insights in a business setting. Overall, my Bachelor's degree has provided me with a rigorous
and comprehensive education in Statistics for Business, giving me the skills and knowledge
necessary to succeed in a variety of data-driven roles.
Research
Beyond the Lab: Real-World Benchmarking of Wearable EEGs for Passive Brain-Computer Interfaces
V. Ronca, M. Cecchetti, R. Capotorto, G. Di Flumeri, A. Giorgi, D. Germano, G. Borghini, F. Babiloni, P. Aricò
Brain Informatics 13.1 (2026)
View the abstract
Wearable EEG systems are increasingly used for brain–computer interface (BCI) applications beyond controlled laboratory environments. However, there is still limited evidence on their reliability in real-world cognitive monitoring, especially for deriving robust mental-state indicators. This study investigates the signal quality, computational stability, and neurometric consistency of two widely used consumer-grade EEG devices (Emotiv EPOC X and Muse S) compared to a validated research-grade system (Mindtooth Touch) during naturalistic tasks relevant to passive BCIs and brain-machine intelligence. The research grade system demonstrated higher signal stability, fewer artefacts, and more consistent neurometric responses to cognitive variations, with high significant correlation with subjective measures. Post-processing improved data continuity in consumer devices, but neurometrics remained less sensitive to task demands and less aligned with subjective ratings. Each device reflected different trade-offs between data quality, usability, and cost.
Brain Cortical Area Characterization and Machine Learning-Based Measure of Rasmussen's S-R-K Model
D. Amore, D. Germano, G. Di Flumeri, P. Aricò, V. Ronca, A. Giorgi, A. Vozzi, R. Capotorto, S. Bonelli, F. Drogoul, J.-P. Imbert, G. Granger, F. Babiloni, G. Borghini
Brain Sciences 15.9 (2025)
View the abstract
Background: the Skill, Rule, and Knowledge (S-R-K) model is a framework used to describe and analyze human behaviour and decision-making in complex environments based on the nature of the task and kind of cognitive control required. The S-R-K model is particularly useful in fields like human factor engineering, system design, and safety-critical industries because it helps to understand human errors and how they relate to different levels of cognitive control. However, the S-R-K model is still qualitative and lacks specific and quantifiable metrics for determining what kind of cognitive control a person is using at any given time. This aspect makes difficult to directly measure and compare performance across the three levels. This study aimed therefore to characterize the S-R-K model from a neurophysiological perspective by analyzing the operator’s cerebral cortical activity. Methods: in this study, participants carried out experimental tasks able to replicate the Skill (tracking task), Rule (rule-based navigation) and Knowledge conditions (unfamiliar situations). Results: participants’ Electroencephalogram (EEG) was recorded during tasks execution and then Global Field Power (GFP) was estimated in the different EEG frequency bands. Brodmann areas (BAs) and EEG features were then used to characterize the S-R-K pattern over the cerebral cortex and as inputs to build up the machine learning-based model to estimate participants’ cognitive control behaviours while dealing with tasks. Conclusions: the results demonstrate the possibility of objectively measuring the different S, R and K levels in terms of brain activations. Furthermore, such evidence is consistent with the scientific literature in terms of cognitive functions corresponding to the different levels of cognitive control.
A Novel Multi-Stage Algorithm for Real-Time Detection and Correction of Ocular Artifacts in EEG: A Calibration-Free Approach
V. Ronca, G. Di Flumeri, L. Lungarini, R. Capotorto, D. Germano, A. Giorgi, G. Borghini, F. Babiloni, P. Aricò
Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBS) (2025)
View the abstract
Ocular artifacts, particularly blinks, significantly affect the integrity of electroencephalographic (EEG) signals, posing a challenge for real-time applications. Traditional correction methods often require a calibration phase or additional electrooculogram (EOG) channels, limiting their applicability in mobile and real-world settings. This study presents a novel detection and correction method, designed for online ocular artifact correction without the need for prior calibration: the CFo-CLEAN. The proposed method integrates an Enhanced Adaptive Data-driven Algorithm (eADA) for real-time identification and correction of ocular artifacts directly from EEG signals. Unlike conventional approaches, this implementation adapts dynamically to ongoing EEG variations, enhancing flexibility and performance. The study evaluates the CFo-CLEAN method using EEG data recorded from 38 participants during real-world driving scenarios. Performance comparisons were conducted against established correction techniques, including Independent Component Analysis (ICA), regression-based methods, and subspace reconstruction approaches. The evaluation considered both artifact removal efficiency and EEG signal preservation across different experimental conditions. Results demonstrated that the method effectively reduced ocular artifact contamination while preserving neurophysiological content. Specifically, two implementations of the method, utilizing 60-second and 90-second time windows, were analyzed, revealing that longer windows provided superior EEG signal preservation, particularly in higher frequency bands. These findings validate the effectiveness of the CFo-CLEAN method for real-time applications, making it a valuable tool for brain-computer interfaces (BCIs), neuroergonomics, and cognitive state monitoring. By avoiding the need for a calibration phase and incorporating adaptive processing, this method represents a significant advancement in real-time EEG artifact correction, facilitating its deployment in dynamic, real-world environments.
Towards the Correction of Covariate Shift in EEG-Based Passive Brain-Computer Interfaces for Out-of-Lab Applications
D. Germano, V. Ronca, R. Capotorto, G. Di Flumeri, G. Borghini, A. Giorgi, F. Babiloni, P. Aricò
47th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC) (2025)
View the abstract
The increasing adoption of wearable EEG technology is enabling the development of passive Brain-Computer Interface (pBCI) systems for real-world applications, in the near future, such as Industry 5.0. However, one major challenge in classifying electroencephalographic (EEG) signals in these settings is covariate shift, which occurs when the distribution of the data changes between training and testing sessions due to variations in EEG headset positioning. This study investigates the effectiveness of a linear transformation approach to mitigate the negative effect of covariate shift. Simulations were conducted by using different shift conditions (i.e. deviation of the headset position from the original one), to evaluate (i) the performance of the transformation function used for mitigating the covariate shift occurrence and (ii) the importance that the change of reference and/or channels has on the classification performance. Results show that normalizing covariate shift-affected data (i.e., target) using shift-free data as a template (i.e., source) helps mitigate the negative impact of covariate shift, leading to improved classification performanceThe accuracy loss drops from 14% to 6% in the worst configuration and from 5% to 4% in the best configuration. This improvement is more pronounced when the shift is larger, i.e., when both the reference and channels change between the control dataset and the test dataset. These findings have significant implications for the development of robust and reliable pBCI models for out-of-the-lab contexts.
Understanding the Unexplored: A Review on the Gap in Human Factors Characterization for Industry 5.0
A. Ricci, V. Ronca, R. Capotorto, A. Giorgi, A. Vozzi, D. Germano, G. Borghini, G. Di Flumeri, F. Babiloni, P. Aricò
Applied Sciences 15.4 (2025)
View the abstract
The integration of neurophysiological techniques into Industry 5.0 represents a transformative approach to assessing human factors in real-world operational settings. This study presents a systematic review of existing literature to evaluate the application of neurophysiological methods in assessing cognitive and emotional states, such as workload, stress, attention, and trust, within industrial environments. A total of X peer-reviewed articles published between 2018 and 2024 were analyzed following a structured methodology. The findings reveal that EEG (45%), eye-tracking (30%), EDA (20%), and ECG (15%) are the most frequently adopted techniques for monitoring cognitive and emotional responses. Additionally, 60% of the studies focused on stress and workload assessment, while only 25% examined trust and collaboration in human–robot interaction, highlighting a gap in comprehensive teamwork analysis. Furthermore, only 35% of the studies validated their approaches in real-world industrial settings, emphasizing a significant limitation in ecological validity. The review also identifies that multimodal integration remains underexplored, with just 15% of studies combining multiple neurophysiological signals for a more holistic assessment. These results indicate a growing but still fragmented research landscape, with clear opportunities for expanding real-world applications, improving methodological standardization, and fostering interdisciplinary collaboration. Future research should prioritize validation in dynamic, real-life work environments and explore the synergistic potential of multimodal neurophysiological approaches to enhance human-centred industrial systems.
Finestre di Responsività nello Stato di Minima Coscienza: Un Modello Neurofisiologico per un'Interfaccia Cervello-Computer Passiva
A. Riccio, J. Toppi, P. Aricò, V. Caracci, I. Quattrociocchi, V. Galiotta, M. D'Ippolito, D. Germano, F. Pichiorri, F. Cincotti, R. Formisano, D. Mattia
Nutrire il cervello, curare il pianeta. Innovazioni di neuroriabilitazione all'Insegna di One Health (2025)
View the abstract
This contribution proposes a neurophysiological model for identifying windows of responsiveness in
patients in a minimally conscious state, with implications for passive brain-computer interfaces in
neurorehabilitation.
Single-Subject Unsupervised Detection of Covariate Shift in EEG: A Step Towards Real-World Passive Brain-Computer Interface
D. Germano, V. Ronca, R. Capotorto, G. Di Flumeri, G. Borghini, A. Giorgi, A. Vozzi, F. Babiloni, P. Aricò
IX National Congress of Bioengineering (2025)
View the abstract
Thanks to significant advancements in wearable EEG technology, the application of passive Brain-Computer Interface (pBCI) systems is now becoming feasible in settings beyond the laboratory, such as Industry 5.0, where human machine interaction is central. However, in these real-world and less controlled settings, a common challenge in classifying EEG signals is the variation in EEG headset positioning across sessions, which can induce a phenomenon known as covariate shift, leading to reduced accuracy between training and subsequent recording sessions. This study aims to detect covariate shift, using an unsupervised Machine Learning model (i.e. Isolation Forest), at the single subject level for online applications, focusing on evaluating how quickly the model identifies anomalous records that deviate from the distribution of the control dataset. Simulations were conducted by using different shift conditions (i.e. deviation of the headset position from the original one), to evaluate (i) the performance of the detection model used for identifying the occurrence of covariate shift and (ii) the number of needed data samples to provide a statistically significant detection. The results show that it is possible to detect Covariate Shift in real-time, using less than 20 seconds of recording. This result was obtained in 9 out of 10 participants, across all test configurations and it indicates that with only 20 seconds of recording, the number of anomalies detected by the method is statistically significant compared to the number of anomalies in a control dataset, with a significance level of 5%. Only in one participant the method fails to reach this level of accuracy for two out of three test configurations. This pilot study represents an important step towards the generation of robust and reliable pBCI models for applications in real-world contexts.
Optimizing EEG Signal Integrity: A Comprehensive Guide to Ocular Artifact Correction
V. Ronca, R. Capotorto, G. Di Flumeri, A. Giorgi, A. Vozzi, D. Germano, V. Di Virgilio, G. Borghini, G. Cartocci, D. Rossi, B. M. S. Inguscio, F. Babiloni, P. Aricò
Bioengineering 11.10 (2024)
View the abstract
Ocular artifacts, including blinks and saccades, pose significant challenges in the analysis of electroencephalographic (EEG) data, often obscuring crucial neural signals. This tutorial provides a comprehensive guide to the most effective methods for correcting these artifacts, with a focus on algorithms designed for both laboratory and real-world settings. We review traditional approaches, such as regression-based techniques and Independent Component Analysis (ICA), alongside more advanced methods like Artifact Subspace Reconstruction (ASR) and deep learning-based algorithms. Through detailed step-by-step instructions and comparative analysis, this tutorial equips researchers with the tools necessary to maintain the integrity of EEG data, ensuring accurate and reliable results in neurophysiological studies. The strategies discussed are particularly relevant for wearable EEG systems and real-time applications, reflecting the growing demand for robust and adaptable solutions in applied neuroscience.
o-CLEAN: A Novel Multi-Stage Algorithm for the Ocular Artifacts' Correction from EEG Data in Out-of-the-Lab Applications
V. Ronca, G. Di Flumeri, A. Giorgi, A. Vozzi, R. Capotorto, D. Germano, N. Sciaraffa, G. Borghini, F. Babiloni, P. Aricò
Journal of Neural Engineering 21.5 (2024)
View the abstract
In the context of electroencephalographic (EEG) signal processing, artifacts generated by ocular movements, such as blinks, are significant confounding factors. These artifacts overwhelm informative EEG features and may occur too frequently to simply remove affected epochs without losing valuable data. Correcting these artifacts remains a challenge, particularly in out-of-lab and online applications using wearable EEG systems (i.e. with low number of EEG channels, without any additional channels to track EOG). Objective. The main objective of the present work consisted in validating a novel ocular blinks artefacts correction method, named multi-stage OCuLar artEfActs deNoising algorithm (o-CLEAN), suitable for online processing with minimal EEG channels. Approach. The research was conducted considering one EEG dataset collected in highly controlled environment, and a second one collected in real environment. The analysis was performed by comparing the o-CLEAN method with previously validated state-of-art techniques, and by evaluating its performance along two dimensions: (a) the ocular artefacts correction performance (IN-Blink), and (b) the EEG signal preservation when the method was applied without any ocular artefacts occurrence (OUT-Blink). Main results. Results highlighted that (i) o-CLEAN algorithm resulted to be, at least, significantly reliable as the most validated approaches identified in scientific literature in terms of ocular blink artifacts correction, (ii) o-CLEAN showed the best performances in terms of EEG signal preservation especially with a low number of EEG channels. Significance. The testing and validation of the o-CLEAN addresses a relevant open issue in bioengineering EEG processing, especially within out-of-the-lab application. In fact, the method offers an effective solution for correcting ocular artifacts in EEG signals with a low number of available channels, for online processing, and without any specific template of the EOG. It was demonstrated to be particularly effective for EEG data gathered in real environments using wearable systems, a rapidly expanding area within applied neuroscience.
Cooperation Objective Evaluation in Aviation: Validation and Comparison of Two Novel Approaches in Simulated Environment
R. Capotorto, V. Ronca, N. Sciaraffa, G. Borghini, G. Di Flumeri, L. Mezzadri, A. Vozzi, A. Giorgi, D. Germano, F. Babiloni, P. Aricò
Frontiers in Neuroinformatics 18 (2024)
View the abstract
This study validates and compares two approaches for evaluating cooperation in aviation simulations,
combining behavioral and neurophysiological measures to characterize collaboration and human factors
in operationally relevant environments.
How Immersed Are You? State of the Art of the Neurophysiological Characterization of Embodiment in Mixed Reality for Out-of-the-Lab Applications
V. Ronca, A. Ricci, R. Capotorto, L. Di Donato, D. Freda, M. Pirozzi, E. Palermo, L. Mattioli, G. Di Gironimo, D. Coccorese, S. Buonocore, F. Massa, D. Germano, G. Di Flumeri, G. Borghini, F. Babiloni, P. Aricò
Applied Sciences 14.18 (2024)
View the abstract
Mixed Reality (MR) environments hold immense potential for inducing a sense of embodiment, where users feel like their bodies are present within the virtual space. This subjective experience has been traditionally assessed using subjective reports and behavioral measures. However, neurophysiological approaches offer unique advantages in objectively characterizing embodiment. This review article explores the current state of the art in utilizing neurophysiological techniques, particularly Electroencephalography (EEG), Photoplethysmography (PPG), and Electrodermal activity (EDA), to investigate the neural and autonomic correlates of embodiment in MR for out-of-the-lab applications. More specifically, it was investigated how EEG, with its high temporal resolution, PPG, and EDA, can capture transient brain activity associated with specific aspects of embodiment, such as visuomotor synchrony, visual feedback of a virtual body, and manipulations of virtual body parts. The potential of such neurophysiological signals to differentiate between subjective experiences of embodiment was discussed, with a particular regard to identify the neural and autonomic markers of early embodiment formation during MR exposure in real settings. Finally, the strengths and limitations of the neurophysiological approach in the context of MR embodiment research were discussed, in order to achieve a more comprehensive understanding of this multifaceted phenomenon.
Unsupervised Detection of Covariate Shift Due to Changes in EEG Headset Position: Towards an Effective Out-of-Lab Use of Passive Brain–Computer Interface
D. Germano, N. Sciaraffa, V. Ronca, A. Giorgi, G. Trulli, G. Borghini, G. Di Flumeri, F. Babiloni, P. Aricò
Applied Sciences (MDPI) 2023
View the abstract
In the field of passive Brain–computer Interfaces (BCI), the need to develop systems that require rapid setup, suitable for use outside of laboratories is a fundamental challenge, especially now, that the market is flooded with novel EEG headsets with a good quality. However, the lack of control in operational conditions can compromise the performance of the machine learning model behind the BCI system. First, this study focuses on evaluating the performance loss of the BCI system, induced by a different positioning of the EEG headset (and of course sensors), so generating a variation in the control features used to calibrate the machine learning algorithm. This phenomenon is called covariate shift. Detecting covariate shift occurrences in advance allows for preventive measures, such as informing the user to adjust the position of the headset or applying specific corrections in new coming data. We used in this study an unsupervised Machine Learning model, the Isolation Forest, to detect covariate shift occurrence in new coming data. We tested the method on two different datasets, one in a controlled setting (9 participants), and the other in a more realistic setting (10 participants). In the controlled dataset, we simulated the movement of the EEG cap using different channel and reference configurations. For each test configuration, we selected a set of electrodes near the control electrodes. Regarding the realistic dataset, we aimed to simulate the use of the cap outside the laboratory, mimicking the removal and repositioning of the cap by a non-expert user. In both datasets, we recorded multiple test sessions for each configuration while executing a set of Workload tasks. The results obtained using the Isolation Forest model allowed the identification of covariate shift in the data, even with a 15-s recording sample. Moreover, the results showed a strong and significant negative correlation between the percentage of covariate shift detected by the method, and the accuracy of the passive BCI system (p-value < 0.01). This novel approach opens new perspectives for developing more robust and flexible BCI systems, with the potential to move these technologies towards out-of-the-lab use, without the need for supervision for use by a non-expert user.
A Neuroergonomic Approach Fostered by Wearable EEG for the Multimodal Assessment of Drivers Trainees
G. Di Flumeri, A. Giorgi,D. Germano, V. Ronca, A. Vozzi, G. Borghini, L. Tamborra, I. Simonetti, R. Capotorto, S. Ferrara, N. Sciaraffa, F. Babiloni, P. Aricò
Applied Sciences (MDPI) 2023
View the abstract
When assessing trainees’ progresses during a driving training program, instructors can only rely on the evaluation of a trainee’s explicit behavior and their performance, without having any insight about the training effects at a cognitive level. However, being able to drive does not imply knowing how to drive safely in a complex scenario such as the road traffic. Indeed, the latter point involves mental aspects, such as the ability to manage and allocate one’s mental effort appropriately, which are difficult to assess objectively. In this scenario, this study investigates the validity of deploying an electroencephalographic neurometric of mental effort, obtained through a wearable electroencephalographic device, to improve the assessment of the trainee. The study engaged 22 young people, without or with limited driving experience. They were asked to drive along five different but similar urban routes, while their brain activity was recorded through electroencephalography. Moreover, driving performance, subjective and reaction times measures were collected for a multimodal analysis. In terms of subjective and performance measures, no driving improvement could be detected either through the driver’s subjective measures or through their driving performance. On the other side, through the electroencephalographic neurometric of mental effort, it was possible to catch their improvement in terms of mental performance, with a decrease in experienced mental demand after three repetitions of the driving training tasks. These results were confirmed by the analysis of reaction times, that significantly improved from the third repetition as well. Therefore, being able to measure when a task is less mentally demanding, and so more automatic, allows to deduce the degree of users training, becoming capable of handling additional tasks and reacting to unexpected events.
AIR SAFE: Leveraging IoT Sensors and AI Models to Foster Optimal Indoor Conditions
M. Viviani, S. Colace, D. Germano, S. Laurita, G. Papuzzo, A. Forestiero
Numerical Computations: Theory and Algorithms - 4th International Conference, NUMTA 2023
View the abstract
The need of safe and livable indoor environments has intensified recently, given the great amount of time people spend
indoor. In addition, the recent COVID-19 pandemic has moved the interest from the outdoor to indoor spaces. To guarantee
an optimal indoor environmental quality, monitoring and regulating many variables (such as indoor and outdoor
temperature, pollutants concentration, noise, and brightness) is necessary. In this context, we have developed AIR SAFE,
an IoT and AI based infrastructure to monitor and control environmental quality in closed spaces. AIR SAFE uses Machine
Learning models to make predictions of temperature, relative humidity, and CO2 concentration. These predictions,
together with data from a network of IoT sensors, are used to take actions on windows and air conditioning system with
the aim of modifying for the better the room environment. We show the results of the AI model we have developed for
predicting indoor concentration of CO2, relative humidity, and temperature. Our Long Short Term Memory (LSTM) model
has been tested against literature models using simulated data at first, and then testing the best models on real data.
Using real data, the LSTM network performs best at forecasting temperature and relative humidity, while Random Forest is
the best CO2 concentration predictor.
Evaluation of a New Lightweight EEG Technology for Translational Applications of Passive
Brain-Computer Interfaces
N. Sciaraffa, G. Di Flumeri, D. Germano, A. Giorgi, A. Di Florio, G. Borghini, A. Vozzi, V.
Ronca, F. Babiloni, P. Aricò
Frontiers in Human Neuroscience 16 (2022)
View the abstract
Technologies like passive brain-computer interfaces (BCI) can enhance human-machine
interaction. Anyhow, there are still shortcomings in terms of easiness of use, reliability,
and generalizability that prevent passive-BCI from entering real-life situations. The
current work aimed to technologically and methodologically design a new gel-free passive-BCI
system for out-of-the-lab employment. The choice of the water-based electrodes and the
design of a new lightweight headset met the need for easy-to-wear, comfortable, and highly
acceptable technology. The proposed system showed high reliability in both laboratory and
realistic settings, performing not significantly different from the gold standard based on
gel electrodes. In both cases, the proposed system allowed effective discrimination (AUC
> 0.9) between low and high levels of workload, vigilance, and stress even for high
temporal resolution (<10 s). Finally, the generalizability of the proposed system has
been tested through a cross-task calibration. The system calibrated with the data recorded
during the laboratory tasks was able to discriminate the targeted human factors during the
realistic task reaching AUC values higher than 0.8 at 40 s of temporal resolution in case of
vigilance and workload, and 20 s of temporal resolution for the stress monitoring. These
results pave the way for ecologic use of the system, where calibration data of the realistic
task are difficult to obtain.
Validation of a Light EEG-Based Measure for Real-Time Stress Monitoring during Realistic Driving
N. Sciaraffa, G. Di Flumeri, D. Germano, A. Giorgi, A. Di Florio, G. Borghini, A. Vozzi, V.
Ronca, R. Varga, M. van Gasteren, F. Babiloni, P. Aricò
Brain sciences 12.3 (2022)
View the abstract
Driver's stress affects decision-making and the probability of risk occurrence, and it is
therefore a key factor in road safety. This suggests the need for continuous stress
monitoring. This work aims at validating a stress neurophysiological measure—a
Neurometric—for out-of-the-lab use obtained from lightweight EEG relying on two wet sensors,
in real-time, and without calibration. The Neurometric was tested during a multitasking
experiment and validated with a realistic driving simulator. Twenty subjects participated in
the experiment, and the resulting stress Neurometric was compared with the Random Forest
(RF) model, calibrated by using EEG features and both intra-subject and cross-task
approaches. The Neurometric was also compared with a measure based on skin conductance level
(SCL), representing one of the physiological parameters investigated in the literature
mostly correlated with stress variations. We found that during both multitasking and
realistic driving experiments, the Neurometric was able to discriminate between low and high
levels of stress with an average Area Under Curve (AUC) value higher than 0.9. Furthermore,
the stress Neurometric showed higher AUC and stability than both the SCL measure and the RF
calibrated with a cross-task approach. In conclusion, the Neurometric proposed in this work
proved to be suitable for out-of-the-lab monitoring of stress levels.
Mental Effort Estimation by Passive BCI: A Cross-Subject Analysis
N. Sciaraffa, D. Germano, A. Giorgi, V. Ronca, A. Vozzi, G. Borghini, G. Di Flumeri, F.
Babiloni, P. Aricò
43rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society
(2021)
View the abstract
Despite the technological advancements, the employment of passive brain computer interface
(BCI) out of the laboratory context is still challenging. This is largely due to
methodological reasons. On the one hand, machine learning methods have shown their potential
in maximizing performance for user mental states classification. On the other hand, the
issues related to the necessary and frequent calibration of algorithms and to the temporal
resolution of the measurement (i.e. how long it takes to have a reliable state measure) are
still unsolved. This work explores the performances of a passive BCI system for mental
effort monitoring consisting of three frontal electroencephalographic (EEG) channels. In
particular, three calibration approaches have been tested: an intra-subject approach, a
cross-subject approach, and a free-calibration procedure based on the simple average of
theta activity over the three employed channels. A Random Forest model has been employed in
the first two cases. The results obtained during multi-tasking have shown that the
cross-subject approach allows the classification of low and high mental effort with an AUC
higher than 0.9, with a related time resolution of 45 seconds. Moreover, these performances
are not significantly different from the intra-subject approach although they are
significantly higher than the calibration-free approach. In conclusion, these results
suggest that a light (three EEG channels) passive BCI system based on a Random Forest
algorithm and cross-subject calibration could be a simple and reliable tool for
out-of-the-lab employment.
Expertise
Technical Expertise
Artificial Intelligence, Machine Learning, Generative AI and Statistical Modeling
My technical work combines statistical reasoning with modern artificial intelligence methods, from
machine learning and generative AI to quantitative analysis and predictive modeling.
Scientific Computing
Python, SQL, R and Quantitative Analysis
I use scientific programming and data analysis to turn complex datasets into reproducible analyses,
interpretable models and practical evidence for decision-making.
Specialization
Signal Processing, Neuroscience and Passive Brain-Computer Interfaces
My research specialization focuses on biosignal processing and neurotechnology, especially robust
passive BCI systems intended for use in realistic, out-of-laboratory settings.
Infrastructure
Spark, Google Cloud Platform, Cloudera and Big Data Systems
Experience with distributed data systems supports the development of scalable pipelines and machine
learning workflows for large and heterogeneous datasets.
Affiliations and Service
IEEE Member, AIxIA Member and National Group of Bioengineering
Since 2000, I have also served as an AGESCI Scout Leader, developing leadership, team management and
community-service experience alongside my academic and professional work.
Languages
Italian, English and Spanish
Italian is my native language; I work professionally in English and communicate conversationally in
Spanish.
Research Interests
AI for Healthcare, Brain-Computer Interfaces, Signal Processing, Neurotechnology, Generative AI and Applied Machine Learning
These interests bring together healthcare-oriented artificial intelligence, biosignal analysis and
practical machine learning systems designed to support people in realistic environments.