| Thesis proposals | Researchers |
|
Advanced MRI and machine learning for classification and tracking progression of frontotemporal dementia subtypes
Frontotemporal dementia (FTD) is a neurodegenerative dementia that primarily affects the frontal and temporal lobes of the brain. These regions are typically associated with personality, behaviour, language, and decision-making. FTD typically leads to changes in these functions. Thus, FTD presents various subtypes, such as behavioural FTD (bvFTD) or primary progressive aphasia (PPA). Diagnosing FTD, however, remains a challenge due to the overlap of symptoms with other neurodegenerative and psychiatric conditions. Achieving an early and accurate diagnosis is crucial to improve patient care. Magnetic resonance imaging (MRI) plays an important role in diagnosing and tracking the disease's evolution.
Our project aims to address this critical gap in medical research. Using advanced data from structural MRI and diffusion tensor imaging (DTI), we aim to classify FTD patients with the help of statistical methods and machine learning algorithms. The classification information, using explainable machine learning techniques, will allow us to identify potential biomarkers to help diagnose the various FTD subtypes. Moreover, by leveraging MRI and DTI data, we plan to create normative models to track disease progression on an individual basis and to explore the different progression according to the FTD subtype.
This innovative research will be conducted in collaboration with the Alzheimer's Disease and Cognitive Disorders Group at Barcelona's IDIBAPS-Hospital Clínic, a world-renowned clinical institution. Together, we aim to make significant strides in understanding and diagnosing FTD, ultimately contributing to better outcomes for patients.
This research will be carried out in close collaboration with Dr Raquel Sánchez-Valle at Barcelona’s IDIBAPS-Hospital Clínic, a world-renowned clinical institution.
|
|
|
Advanced MRI, Sleep Parameters and Machine Learning for Tracking Progression of Alzheimer’s Disease
Alzheimer’s disease (AD) is the most common neurodegenerative dementia, characterized primarily by progressive deficits in memory, cognition and daily functioning. The disease manifests in stages, ranging from mild cognitive impairment to severe dementia. Studying AD progression remains challenging, particularly in the early stages; however, early, accurate diagnosis and progression monitoring are essential for optimal patient management. Magnetic resonance imaging (MRI) plays a vital role in this by detecting characteristic patterns of brain atrophy. Additionally, sleep parameters are increasingly recognized as both potential biomarkers and contributing factors in AD progression.
Our project aims to address a critical gap in medical research. Using advanced MRI data, we will study AD progression and its interaction with sleep parameters. Furthermore, we will investigate the utility of sleep parameters for early diagnosis. We will employ explainable machine learning techniques and normative models to track disease progression at the individual level and explore varying progression rates.
This innovative research will be conducted in close collaboration with Dr. Neus Falgàs and the Alzheimer's Disease and Cognitive Disorders Group at Barcelona's IDIBAPS-Hospital Clínic. Together, we aim to make significant strides in understanding AD progression to improve patient outcomes.
This innovative research will be conducted in close collaboration with Dr. Neus Falgàs and the Alzheimer's Disease and Cognitive Disorders Group at Barcelona's IDIBAPS-Hospital Clínic. Together, we aim to make significant strides in understanding AD progression to improve patient outcomes.
|
|
|
Application of high performance computing in bioinformatics
This research line focuses on the use of high performance computing (HPC) techniques for optimizing and developing new bioinformatics tools and algorithms based on advanced computer architectures. The aims are to effectively use environments like supercomputing, HPC clusters, grids, and cloud computing in the field of bioinformatics and to explore graphics processing units (GPUs) and other computing accelerators to enhance the performance of bioinformatics tools and algorithms.
|
|
|
CBCT imaging in dental and maxillofacial health
In this research line, you'll work at the forefront of dental and maxillofacial health, developing innovative tools and techniques to analyse essential structures like teeth, the maxilla, and the mandibular nerve.
Our aim is to expand the possibilities of cone beam computed tomography (CBCT) imaging, enhancing diagnostic accuracy and treatment planning for dental and maxillofacial applications. You'll have the chance to design and implement advanced imaging and bioinformatics methods, contributing directly to improvements in patient care and clinical decision-making.
With access to cutting-edge CBCT imaging technology and guidance from experienced mentors, you'll gain valuable expertise. A background in image processing, 3D modelling or craniofacial anatomy is helpful, but the most important requirement is a genuine passion for advancing this field.
|
|
|
Medical image abnormality detection
Medical image screening is tedious and time-consuming work. Clinicians can spend hours examining magnetic resonance (MR), ultrasound or computed tomography (CT) images for abnormalities. We are now capable of obtaining a wider variety of image modalities with much better quality, but with these improvements comes more time having to be spent on screening them due to the increase in information. Multimodality screening is a more advantageous option for detecting abnormalities, but it is a difficult process that requires medical training and specialization. Differences in the levels of expertise between raters can lead to varying diagnostic criteria with a significant impact on our healthcare system.
This project aims to deploy a tool that, based on the latest advances in deep learning techniques, will be able to decide whether a multimodal scan set is susceptible to having abnormalities or not. Moreover, in order to assist the specialist's assessment, it will output a colour map suggesting where the abnormal areas are. This tool will help clinicians to reduce the screening time per subject and will make the intra-observer decision-making process more robust. This research will be carried out in close collaboration with the Multiple Sclerosis Group led by Dr Sara Llufriu at Barcelona’s IDIBAPS-Hospital Clínic, a world-renowned clinical institution.
|
|
|
Synthetic medical data generation to automatically train neuronal networks
Deep learning refers to neural networks with many layers that extract a hierarchy of features from raw data. Deep learning models are now achieving impressive results and generalizability by training on a large amount of data. Thanks to these big datasets, we are able to train deep learning algorithms or general machine learning algorithms with an enormous amount of instances that provide robustness to variations and better generalization properties.
However, in some domains large datasets may not be available, which is a significant problem in several medical areas because their training datasets are relatively small compared to large-scale image datasets (e.g., ImageNet), making it difficult to achieve generalization across datasets. Moreover, current deep learning architectures are based on supervised learning and require the generation of manual ground truth labels, which is tedious work with large-scale data (Akkus et al., Journal of Digital Imaging, 2017).
In this project, we aim to design and develop methods to generate synthetic data from real magnetic resonance imaging (MRI) data. The main objective is to expand or create new data that realistically mimic variations in MRI data and could alleviate the need for a large amount of data. For instance, autoencoders could be used to generate synthetic data (Bengio et al., NIPS, 2013), but it is necessary to consider the type of data and how to modify the data in order to produce variations that are as realistic as possible. Furthermore, methods to assess the data utility are critical and they need to be developed to ensure that synthetic data are realistic enough to train machine learning models. This work will be done in close collaboration with the Multiple Sclerosis Group led by Dr Sara Llufriu at Barcelona's IDIBAPS-Hospital Clínic, a world-renowned clinical institution.
|
|
|
Retinal analysis for eye and brain disease diagnosis and prognostication
The analysis of the back of the eye (fundus) is crucial to identify not only eye disease but also serious neurological conditions that can lead to blindness, brain injury and even death.
Determining these conditions is a common clinical challenge in ophthalmic, neurosurgical and neurological clinics, emergency departments, and intensive care units, and their accurate assessment sometimes involves invasive testing that carries moderate to severe risks of injury.
At present, eye screening data is assessed by an expert to identify lesions and clinical markers that may lead to further patient referral. This procedure is technically challenging and limits its usefulness in non-specialist environments. However, the substantial volume of available routinely collected digitized ocular images and clinical data presents a unique opportunity to develop novel deep learning systems that automate eye screening data assessment. The implementation of such systems in eye clinics can enable cost-effective, scalable and sustainable clinical pathways for the review and management of eye diseases and more accurate quantification of morphological features and pathological lesions for prognostication.
In this project we aim to develop deep learning-driven systems for the automated detection of ophthalmic and neurological conditions using multiple modalities of image and video retinal data. The study is conducted by an interdisciplinary team of ophthalmology, neurology and artificial intelligence experts.
|
|
|
Rethinking the Tree of Life: Quantifying Horizontal Evolution in Eukaryotes
Since Darwin's time, evolution has been represented as a tree. While powerful, this metaphor is insufficient as it only represents vertical inheritance – the divergence of species from a common ancestor via the parental line.
However, evolution has a horizontal dimension that is often neglected. As Griffith's experiments showed, species can acquire genetic material from others in a process now known as horizontal gene transfer (HGT). While historically thought to be specific to Bacteria, HGT also occurs in eukaryotes. The most paradigmatic example is the chimeric origin of the eukaryotic cell, which evolved by incorporating an alphaproteobacterial endosymbiont that became the mitochondrion.
Some scientists have proposed transitioning from a "tree" to a "net" to represent these horizontal processes. By exploring hundreds of genomes from modern eukaryotic species, the student will quantify how frequent horizontal processes are in eukaryotic evolution. This will help clarify to what extent the evolutionary tree is still the right metaphor for eukaryotes
|
|
|
Does Evolution Repeat Itself? Machine Learning to Unravel Genomic Signatures of Convergent Evolution in Eukaryotes
Is evolution predictable? If we rewound the tape of life, would we end up with the same biodiversity we see today? While we can't re-run this experiment, we can study convergent evolution – the fascinating process where different species independently evolve similar traits (phenotypes) to solve similar environmental problems.
This project tackles a key question: does this outward similarity (phenotypic convergence) mean the underlying genetic changes are also the same (genetic convergence)?
The student will have the opportunity to dive into the massive genomic datasets now available for thousands of eukaryotic species. Cutting-edge computational tools and machine learning will be employed to discover which gene content features are linked to specific convergent adaptations. Next, ancestral gene content reconstruction will allow pinpointing when these key genetic changes occurred in evolutionary history.
This work will allow the identification of the specific genetic events that enabled some of the most striking examples of convergent evolution in eukaryotes, offering new insights into the balance of chance and determinism at the molecular level.
|
|
|
Omics in Nutrition, Cancer and Cardiometabolic Diseases
This research line focuses on proteomic, lipidomic, metabolomic and metagenomic analyses from human biological fluids applied to the most prevalent causes of human death. The specific aims of this line are to elucidate potential mechanisms or novel biomarkers of disease onset, progression or treatment efficacy derived from epidemiological or interventional trials.
|
|
|
Interpretable Deep Learning Models for Oral Cancer Classification
This research line forms part of a project developed in collaboration with Universidad Complutense de Madrid and UNED. A dataset of more than 8,000 images of the oral cavity has been collected and categorized into 4 broad categories: healthy, benign, potentially malignant and malignant. The goal of this research is to propose new deep learning models and provide analysis that enables the interpretation of the implemented models.
|