My research focuses on intelligent methods for analysing physiological and neurophysiological signals, with particular emphasis on multimodal learning, neurophysiological decoding, wearable biosensing, and embedded and real-time AI.
I am interested in translating signal processing and machine learning methods into practical intelligent healthcare and neurotechnology systems that can operate beyond conventional offline analysis.
My work spans physiological sensing, signal processing, machine learning and deployable intelligent systems.
Processing and analysis of physiological and neurophysiological signals, including EEG, fNIRS, EDA, PPG/BVP, ECG, EMG and PCG. My work considers signal preprocessing, temporal dynamics, feature representation and machine learning for extracting clinically and physiologically meaningful information.
Development of machine and deep learning methods for combining complementary physiological modalities. This includes multimodal fusion, representation learning, CNN-based modelling, recurrent architectures and attention and Transformer-based approaches for physiological time-series analysis.
Decoding information from brain and peripheral physiological activity using EEG, fNIRS and other biosignals. Current interests include temporal representation learning, subject-independent modelling and hybrid EEG-fNIRS approaches for brain-computer interface and neurotechnology applications.
Translation of physiological signal processing and machine learning pipelines from offline analysis to real-time operation. This includes lightweight model design, wearable sensor integration, continuous inference and deployment on embedded computing platforms.
These research methods are applied to problems in intelligent healthcare, physiological computing and neurotechnology.
Physiological and neurophysiological approaches for objective assessment of pain using signals such as EDA, PPG/BVP and fNIRS, including unimodal and multimodal machine learning methods.
Investigation of EEG, fNIRS and hybrid EEG-fNIRS signals for decoding brain activity, with an emphasis on machine learning methods for motor imagery and related BCI paradigms.
Integration of wearable physiological sensing with signal processing and machine learning to support continuous, practical and real-time intelligent healthcare applications.
Design and evaluation of end-to-end systems that combine biosignal acquisition, preprocessing, machine learning inference and embedded deployment under real-time operating constraints.
Core methods, signal modalities and technologies used across my research.
EEG · fNIRS · EDA · PPG/BVP · ECG · EMG · PCG
Deep Learning · CNNs · LSTMs · Transformers · Multimodal Learning · Representation Learning
Filtering · Time-Series Analysis · Feature Extraction · Spectral Analysis · Physiological Signal Processing
BIOPAC · EmotiBit · Artinis fNIRS · PLUX Biosignals · Raspberry Pi · STM32
Python · MATLAB · PyTorch · TensorFlow · C/C++ · Git
Embedded AI · Real-Time Inference · Wearable Sensing · Edge Computing · Sensor Integration
Further details of my research outputs and ongoing work are available through my publications and project pages.