Projects
Brain-computer interfaces and neurofeedback
Neurofeedback principles
Real-time functional magnetic resonance imaging (rt-fMRI) can be used to feedback changes in brain activity to participants, using an indirect measure of neural activity such as the blood oxygenation level-dependent (BOLD) signal. This noninvasive approach, known as neurofeedback (NF), allows the participants to self-regulate the brain activity of a target brain region or to modulate the functional connectivity between regions.
There are several open questions unanswered in the research community.
E.g. the feedback interface is traditionally visual (as a “thermometer display” or continuous curve estimated based on the BOLD activation signal) or auditory. However, the choice of the NF interface depends on the functions to be regulated and the target region at hand.
Our research team was involved in different rt-fMRI neurofeedback experiments/analysis that aimed e.g. to assess general feasibility of the setup, evaluate novel neurofeedback target regions, assess connectivity patterns associated to neurofeedback training, feasibility of functional connectivity based rt-fMRI neurofeedback, assess the parametric control of brain activity.
Challenges and unanswered questions in neurofeedback research
BRAINTRAIN and Autism spectrum disorder (ASD)
The EU FP7 BRAINTRAIN project's main idea is to assess the feasibility of neurofeedback training in different diseases evaluate the clinical output of the intervention.
In our particular case (Coimbra's research team), the clinical population is ASD. ASD is a neurodevelopmental disorder with different clinical manifestations, from mild to severe. The most prominent symptom is social interaction impairment (but verbal and non- verbal communication deficits and stereotyped behaviors also have to be present).
EU FP7 BRAINTRAIN project, neurofeedback and ASD - technology transfer
My specific tasks involved the design, submission, management and results analysis of a clinical trial to assess the feasibility of rt-fMRI interventions in ASD.
As a postdoc, I was also involved in the supervision of two MSc students theses. The topics involved rt-fMRI neurofeedback and connectivity patterns.
Epileptic seizure prediction - development of real-time classification methods
The scientific community has made enormous efforts to understand the basic mechanisms underlying the generation of epileptic seizures. The analysis of the pre-ictal dynamics among different brain regions has been shown as an important source of information towards the understanding of the spatio-temporal mechanisms.
Real-time classification framework for epileptic seizure prediction
This study, partially a contribution to the EU FP7 EPILEPSIAE project, aims the prediction of unforeseeable and uncontrollable epileptic seizures.
The first part of this study aims the development of a patient-specific seizure prediction algorithm based on machine learning with high sensitivity and low false positives rate. The dynamical changes of the brain activity are analyzed using a high dimensional feature set obtained from both scalp and intracranial multichannel electroencephalogram (EEG). The features represent low complexity measures, implementable in real-time scenarios and the classification was performed using cost sensitive support vector machines (SVM).
We also addressed the characterization of the EEG spatio-temporal patterns and the classification of specific brain states. The method proposed, based on the segmentation of topographic maps and on a statistical framework (hidden Markov models), shows promising results for the identification of a pre-ictal stage.
Spatio-temporal analysis
Implementation of a novel Electronic Health Record documentation system
Clinical documentation is often unstructured, free-form text which tends to create difficulties in the re-use of information and data transmission, difficulties in the development of clinical decision support systems and data sharing with research partners.
A collaboration between IBILI (research center of the University of Coimbra, Portugal) and Pediatrics Hospital of Coimbra (Coimbra, Portugal) motivated the development of a novel approach.
The key aspects are the integration with currently in-use EHR systems, flexibility, storage of structured information, re-usable information model and semantic interoperability.
Key aspects of the new approach
Information model and SNOMED-CT terminology - the backbones of the project.
New web-based documentation platform
The platform enables the development of re-usable “coded” templates based on structured library of pre-established elements specified in the C-CDA information model. This structure provides the backbone for logically organized groups of variables. The filling of the templates during clinical encounters and/or procedures, facilitates real-time encoding and structuring of the clinical data.