Self-Funded PhD Opportunities

Supervisors

Primary: Dr Erwin Alles - e.alles@ucl.ac.uk

Subsidiary: Prof Ben Cox - b.cox@ucl.ac.uk

Project Proposal and Research Aims

Optical Ultrasound (OpUS) imaging is an emerging modality that uses light, rather than conventional piezoelectricity, to generate and detect acoustic waves. OpUS probes can be fabricated from just optical fibres, adhesives and minimal substrates, and are hence readily miniaturised, cost-effective, and insensitive to electromagnetic interference. As such, OpUS imaging has been successfully demonstrated in in vivo interventional preclinical settings [1] and under concurrent MRI or CT imaging with minimal mutual interaction [2], and is currently being developed for applications during external beam radiotherapy.

Whilst clinically meaningful OpUS imaging has been demonstrated in a variety of settings, the OpUS image quality is presently sub-par due to a low channel count necessitated by the experimental complexity and the use of a rudimentary delay-and-sum algorithm. Alternative algorithms, such as model-based inversion [3], non-linear beamformers [4], or deep-learning based denoisers, have shown great promise but have not yet been thoroughly assessed, and in addition lack computationally efficient real-time implementations required for practical applications.

In this project, you will thoroughly assess the currently implemented beamformers (delay-and-sum, delay-multiply-and-sum, short-lag spatial coherence and model-based inversion), and develop versions of these algorithms that – through parallel or GPU implementations – achieve high frame rates in real-time. Importantly, you will also develop novel imaging algorithms specifically targeting the low channel count of OpUS probes. These algorithms will be tested on both pre-existing experimental data sets and synthetic data sets generated by you.

[1] https://doi.org/10.1117/1.JBO.30.3.036005

[2] https://doi.org/10.1063/5.0225554

[3] https://doi.org/10.1121/10.0034450

[4] https://doi.org/10.1109/IUS54386.2022.9957149

Required Skills or Qualifications

A strong affinity with programming (in your own language of choice), as well as a solid mathematical background. 

Prior experience with parallel computing (e.g. CUDA, MPI or OpenMP) would be helpful, but not critical.

Prior experience with ultrasound imaging, image reconstruction, and ultrasound modelling is recommended but not essential. 

Supervisor(s)

Prof Sergio Bertazzo - s.bertazzo@ucl.ac.uk

Project Proposal and Research Aims

Pathological calcification, or mineral formation in tissues affected by a disease, can play a major (albeit generally not well understood) role in diseases as diverse as: atherosclerosis, rheumatic fever, aortic valve stenosis, cancers, macular degeneration and dementia. Because these diseases are so common, especially among the elderly, practically all humankind will be affected by calcification during their lifetime, either caused by or associated to a disease. 

This project introduces a research approach based on the creation and design of new methods for the full characterization of minerals formed in tissues affected by Alzheimer’s disease. The focus on minerals and their biological microenvironment has the potential to uncover the nature and the mechanism of formation of the calcification, which appear to be an early event in disease progression, and possible key players in Alzheimer’s disease.

The main objective of this project is to establish a new field of research in the medical sciences. This objective will be achieved through a research strategy that involves advanced nano-analytical techniques for the full and holistic characterization of minerals and their and their biological microenvironment in Alzheimer’s disease 

This project has the following primary objectives

O1 - Determine the location and exact nature of mineralisation in samples obtained from sufferers of Alzheimer’s disease.

O2 - Identify the cells and cell mechanisms responsible for the formation of the calcification.

O3 - Determine the effect of calcification on brain cells in culture.

Required Skills or Qualifications

This is a project involving experimental research, where laboratory experience is a plus. Initially, the project will make extensive use of cell culture and electron microscopes, with new techniques added as the research work progresses. The PhD candidate will be expected to collect and process animal and human tissues.

Supervisor(s)

Prof Adam Gibson - adam.gibson@ucl.ac.uk

Project Proposal and Research Aims

We have a hyperspectral imaging (HSI) system which can map the spectral content of a surface of a sample. However, it has limited depth information so overlying layers can obscure the layer of interest. Optical Coherence Tomography (OCT) is a three-dimensional optical imaging technique that can show the thickness of an overlying layer at specific points.

The objective of this work is to adapt a two-dimensional scanning HSI camera so that it can perform a large area, three-dimensional OCT scan that is co-registered with the HSI image, enabling the thickness of overlying layers to be identified, allowing improved chromophore identification. A combined system would have application in medical imaging and beyond, where a coloured layer of interest is partially obscured by a translucent overlying layer.

We are not aware of any applications where HSI and OCT have been applied together over extended areas. This will be the first time HSI and OCT have been combined in analysis. The additional information provided by OCT will require the development of new spectral unmixing algorithms. We anticipate publications in the areas on instrumentation (e.g. Reviews of Scientific Instruments), spectral unmixing (e.g. Biomedical Optics Express) and medical imaging (e.g. Physics in Medicine and Biology)

Required Skills or Qualifications

This open ended, cross-disciplinary project will require a combination of instrumentation development, computer control and development of a new spectral unmixing algorithm, which is likely to be based on deep learning.

Research aims

1. Adapt an existing HSI camera so that it also performs OCT
2. Integrate combined camera, light sources and scanning frame into a complete imaging system
3. Develop new spectral unmixing algorithm that takes account of additional information to remove effect of overlying layers
4. Demonstrate combined imaging device on controlled test samples and clinical specimens

Supervisor(s)

Prof Charlotte Hagen - charlotte.hagen.10@ucl.ac.uk

Project Proposal and Research Aims

Micro computed tomography (micro-CT) is an x-ray based technology that provides 3D images of cm scale samples at micrometric resolutions in a non-destructive manner. It has emerged as a key enabling technology in a breadth of disciplines, underpinning research and development in areas such as biomedicine and materials science. Traditionally, micro-CT images are formed from differences in the attenuation of x-ray photons as they pass through different materials in a sample, but new developments have enabled image formation from x-ray refraction, offering greater contrast between materials with similar attenuation properties. Alongside this, micro-CT allows probing samples at different resolution levels, from tens of micrometres to a few micrometres or less. 

UCL’s Advanced X-Ray Imaging (AXIm) group has pioneered new micro-CT technology such as the so-called edge illumination approach (Olivo Journal of Physics: Condensed Matter 2021), which allows detecting x-ray refraction alongside attenuation with conventional (lab-based) x-ray tubes, as well as innovative solutions for multi-resolution imaging (Zekavat et al Physics in Medicine & Biology 2024).

This PhD project will build on these developments to create disruptive solutions to the 3D high-contrast, high-resolution imaging of biological tissue and other materials, through the development of imaging workflows for existing micro-CT systems and/or the design of new micro-CT systems. The project further seeks to demonstrate benefits of the novel methods to new applications in biomedicine and/or materials science. The candidate will be based within the AXIm group, where they will receive training in the technical elements of the project.

Required Skills or Qualifications

We are looking for a curious and motivated individual with an interest in scientific research and technology development and a background in physics, engineering, mathematics or a comparable subject. Basic coding skills are essential (e.g. Matlab, Python). Experience with imaging technology is a bonus but not a strict requirement.

Supervisor(s)

Prof Ben Hall - b.hall@ucl.ac.uk

Project Proposal and Research Aims

As human tissues age, cells accrue mutations. Some of these mutations offer a selective advantage, enabling clonal expansion and takeover of the tissue. Recently, it has been shown that these clones with a selective advantage do not necessarily lead to tumour development, but in fact can compete with precancerous clones, presenting a barrier to cancer development. How different mutations combine, and what determines the transition from a clone in the healthy tissue to a lesion and onto a tumour is not known. This PhD will involve the study of new datasets and development of computational models of epistasis and tumour initiation.  

Required Skills or Qualifications

Either experience in mathematical/physical/computational sciences with a demonstrable interest in biomedicine or a background in life sciences and experience of working with computational methods.

Supervisor(s)

Dr Yipeng Hu - yipeng.hu@ucl.ac.uk

Project Proposal and Research Aims

This project aims to develop machine learning algorithms to enhance the accuracy and reliability of cancer detection across multiple medical imaging modalities. Specifically, it will focus on creating deep learning models to identify correlations between imaging and cancerous features, using prostate and liver cancers as example applications across MR, CT, and ultrasound data. Beyond improving predictive performance, the project will investigate medical image analysis tasks, including image representation, generation and registration, to extract novel insights from the learned models and contribute to advancing data-driven understanding of cancer detection.

Required Skills or Qualifications

Experience in programming, e.g. Python C/C++.

Supervisor(s)

Dr Henry Lancashire -  h.lancashire@ucl.ac.uk

Project Proposal and Research Aims

Neural interfaces generate dramatic clinical benefits, from deep brain stimulation for Parkinson's patients, to enabling communication for patients with locked-in syndrome. To continue delivering long term improvements for patients, neural interfaces must be made increasingly smaller and more reliable. As devices are miniaturised new challenges arise in maintaining device function over patient lifetimes. Understanding device reliability requires understanding the biological and non-biological failure modes of miniaturised implants.

Research projects are available in the investigation of non-biological safety modes of neural implants including in:
Electrochemical performance of electrical neural stimulation;
Longevity of active integrated electronics in implanted devices;
Design and manufacture of novel and reliable neural interfaces.

You will gain skills in implant development from component to system, and in manufacturing integrated microelectronic and/or discrete component devices. You will learn to characterise appropriate material properties, to develop and apply methods of accelerated life testing, to interpret the results with respect to changes in underlying mechanisms of device failure, and to apply statistical approaches to estimate device safety.

Required Skills or Qualifications

Projects are suitable for students with backgrounds in Engineering, Physics, and/or Chemistry. Practical, hands-on, research experience and strong analytical and mathematical skills are desirable. 

Research experience in any of the following fields would be desirable but is not essential: analogue electronics, electrochemistry, implanted devices, materials science, microfabrication, microtechnology, and neural engineering.

Supervisor(s)

Prof Terence Leung - t.leung@ucl.ac.uk

Project Proposal and Research Aims

We believe that the purpose of a smartphone camera is more than just taking good quality photos. For over a decade, we have been researching the use of smartphone cameras as scientific instruments to measure physiological functions. Together with our clinical partners in Ghana, India and the UK, we have developed smartphone imaging techniques to investigate a range of clinical conditions, including screening newborn babies for jaundice, screening young children and pregnant women for anaemia, and monitoring liver patients for decompensated cirrhosis. 

At the heart of our research are techniques that can measure colour accurately. For example, the sclera colour in the eye is normally white but can turn yellow in a jaundiced patient indicating the bilirubin level in their blood is high, which forms the basis of our approach to screen and monitor jaundiced patients.

We are currently recruiting a PhD student to join us in developing new core techniques which can contribute to jaundice monitoring based on sclera colour. The first objective is to develop an AI segmentation technique to identify the sclera region from a face photo using open-source software such as Mediapipe. The AI model should also be optimised so that it can be implemented on a smartphone. The second objective is to develop a personalised bilirubin prediction model that can accurately convert sclera colour into bilirubin level. The successful completion of both objectives will greatly advance our on-going research in clinical monitoring.

For an example of our research, please see: www.detect-jaundice.org

Required Skills or Qualifications

A background in engineering, physical sciences or computer science is desirable. 

A passion in applying technology, especially AI, to healthcare problems is essential. The research will involve coding so any experience in Matlab, Python, Java, etc, would be an advantage. Applicants should also be comfortable to work in a multidisciplinary environment with interactions with both engineers and clinicians.

Supervisor(s)

Prof Karin Shmueli - k.shmueli@ucl.ac.uk

Project Proposal and Research Aims

MRI quantitative magnetic susceptibility mapping (QSM) is a rapidly emerging technique with increasingly widespread clinical applications in the human brain. QSM uses the often discarded phase of the complex MRI signal to provide information related to tissue composition, including iron, deoxyhaemoglobin, myelin and calcification content, which change in diseases such as Parkinson’s and Multiple Sclerosis. Although many studies have shown group differences, there is an unmet clinical need to facilitate the use of QSM for individualized diagnosis. To enable diagnosis and staging of disease, large datasets in healthy subjects are needed to provide “normative” values for comparison with susceptibilities measured in individual patients. 

Barriers include: creating a robust harmonized QSM acquisition protocol; quality control and harmonization of QSM for datasets previously acquired across different acquisition protocols and centres; and characterisation of uncertainty, bias and variability due to age, sex, calcifications, head orientation etc. Therefore, this project aims to tackle these issues to create a normative QSM atlas. This will allow QSM to be used as a quantitative disease biomarker in individuals, and for radiological training. 

Specific aims include: developing a robust protocol to acquire QSM in healthy subjects based on recent consensus recommendations. Developing an automated quality control system to detect artifacts and reject data of insufficient quality. Use DL or AI-based techniques, e.g. transfer learning, to harmonize QSM across many existing datasets. This will allow comparison of QSM from previous clinical studies with the normative atlases developed to determine which QSM metrics are most clinically impactful in each disease.

Required Skills or Qualifications

A background in a relevant Physics, Engineering or Mathematics based subject is essential. You must show a clear interest in Magnetic Resonance Imaging (MRI) physics, particularly as applied to neuroscience and healthcare. Experience in numerical computing and programming in languages such as Matlab, Python or C/C++ will be advantageous. 

Other desirable skills include creative and critical thinking, excellent writing and communication skills, self- and time-management and a capability to work effectively both in a team and independently, and to take the initiative when appropriate.

Supervisor(s)

Dr Robert Moss - robert.moss@ucl.ac.uk

Project Proposal and Research Aims

Micro X-ray Computed Tomography (microCT) is a powerful technique that can reveal the internal structure of objects in 3D, while X-ray diffraction (XRD) is a material-specific technique which can provide a fingerprint to identify unknowns against a database or set of standards.

This project will combine microCT imaging with rapid energy-resolved XRD to provide an instrument that can track and identify materials of interest in an otherwise opaque 3D volume.  The idea here is to develop an XRD add-on that can supplement existing microCT systems to deliver a new capability.

Such a system would have the following example use cases:

• Identification of size and composition of calcifications in biological tissues in cardiac and vascular research.
• Identify crystal phase evolution on charging/discharging cycles of novel battery materials.
• Understand the effect of aggregates and inclusions in building materials for both new and heritage scenarios.

The aims of the project will be to:

1. Become an expert user of our microCT system and energy-resolving detectors necessary to capture XRD data.
2. Identify and develop a relationship with potential users to understand use cases in more detail.
3. Design an XRD add-on that can be installed in the microCT enclosure and maintain a suitable geometry.

Develop a software pipeline for the capture of a microCT dataset, identification of regions of interest, automatic alignment of those regions with the XRD instrument, acquisition of XRD data and data processing/spectral matching.

Required Skills or Qualifications

Interest in complete system development which will require aspects of physics, engineering and computer science.

Strong desire for independent research and willingness to take ownership of research direction.

Willingness to work with ionising radiation.

Willingness to work with and travel to external suppliers/manufacturers and potential uses.

Desirable:

Programming skills in Matlab and/or Python.

3D modelling and 3D printing.

Coding for PC-hardware interfacing (e.g. motion control and/or data acquisition via Arduino, USB, RS232, etc.).

Monte Crlo radiation transport modelling (e.g. GEANT4).