SPARClin Lab | Signal Processing and AI for Clinical Neuroscience

About

The SPARCliN Lab develops advanced signal processing and AI methods for clinical neuroscience. We focus on extracting meaningful patterns from complex brain data to improve diagnosis and treatment.

Based at Maastricht University, Department of Neurosurgery and MHeNs, and in close collaboration with TU Delft, Signal Processing Systems Group, we bridge engineering, neuroscience, and clinical practice.

Research

AI in Neuroimaging

Machine learning and signal processing for EEG, fMRI, and functional ultrasound data.

Epilepsy & Biomarkers

Identifying data-driven imaging biomarkers for neurological disorders with a special focus on epilepsy.

Tensor Methods

Advanced modelling of multi-dimensional data: decompositions, blind source separation and data fusion.

Clinical Translation

Our ambition is to bring AI tools into real-world clinical workflows.

Projects

NWO Vidi NeuroMark

Tensor-based network biomarker discovery

This project develops advanced tensor-based machine learning methods to identify network-level biomarkers in brain data, enabling improved diagnosis and treatment planning in neurological disorders.

DeTAIL

Training and Innovation in tensor-based AI methods for biomedical signals

A research and training initiative focused on developing next-generation tensor-based AI methods for analysing complex biomedical signals and neuroimaging data.

CONTACT Study

Effects of radio frequency thermoCOagulation on the brain NeTwork ACTivity in epilepsy patients

Investigating how radiofrequency thermocoagulation impacts brain network activity in epilepsy patients using advanced fMRI and invasive EEG recordings.

Publications

Selected publications highlighting our key research contributions:

Brain Connectivity: From network science to tensor models

Authors: B. Hunyadi, S. Aviyente

Journal: IEEE Signal Processing Magazine

Year: 2025

Networks are the hallmark of brain function and their temporal evolution is a key aspect. In this tutorial we deep-dive into two complementary techniques - based on network science and tensor decompositions - to model temporal networks.

A Comprehensive Guide to Multiset Canonical Correlation Analysis and Its Application to Joint Blind Source Separation

Authors: I. Lehmann, B. Gabrielson, T. Hasija, T. Adali

Journal: IEEE Transactions on Signal Processing

Year: 2025

Multiset Canonical Correlation Analysis (mCCA) identifies correlated variables across multiple datasets and can be a powerful tool in neuroscience.

The impact of radiofrequency thermocoagulation on brain connectivity in drug‐resistant epilepsy: Insights from stereo‐electroencephalography and cortico‐cortical evoked potentials

Authors: J. Gula, R.J.Slegers, R. Van Hoof, B. Krishnan et al

Journal: Epilepsia

Year: 2025

Will local lesions created by radiofrequency thermocoagulation in epilepsy patients affect distant brain connectivity and excitability?

View all publications on Google Scholar →

Team

Core Team

Dr Borbála Hunyadi

Principal Investigator

Associate professor, expert in signal processing for brain disorders.

Dr Isabell Lehmann

Senior Researcher

Postdoctoral fellow developing data fusion methods for neuroimaging.

Helena Puente Díaz

PhD candidate

Neuroengineer specializing in physiological signal processing.

Affiliated Researchers

Justyna Gula

PhD candidate

Maastricht University, Department of Radiology

Sofia-Eirini Kotti

PhD candidate

TU Delft, Signal Processing Systems Group

Contact

SPARCliN Lab
Maastricht University
Department of Neurosurgery
Faculty of Health, Medicine and Life Sciences
Mental Health and Neuroscience Research Institute (MHeNs)

Email: borbala.hunyadi [at] maastrichtuniversity [dot] nl