MSc Projects 2012-13

Here is a list of projects being offered for MSc students in the Department.
For details of deadlines and for registering submissions, see the Moodle page. Full instructions are available here.

Measurement of temperature using terahertz radiation

Supervisors: Dr Adam Gibson and Dr Vince Wallace

A number of mechanisms have been proposed whereby terahertz (THz) radiation may be able to measure the temperature of tissue. These include measurement of passive THz emission and measurement of THz signal from gold nanoparticles heated by an infrared laser. This project will investigate the possibility of measuring temperature using THz radiation. It will begin with a literature review and theoretical examination of the potential mechanisms by which temperature could change the THz signal. Based on the outcome of this, you will develop a tissue-equivalent test phantom which allows the temperature to be controlled in a known way. You will measure the temperature change using THz and relate it to the known change in the phantom. This project is somewhat open-ended and will involve both theoretical and experimental work.

Student: (Project available)

Maximising sensitivity and specificity of x-ray mammography.

Supervisors: Dr Adam Gibson and (not allocated yet)

We have details of patient history, imaging and outcome on 350 women attending a breast screening clinic. You will complete the assembly and validation of this database and go on to analyse it. We would like to test the hypothesis that computer-assisted image analysis predicts outcome better when combined with patient history than when analysed alone. You will be expected to investigate a number of different maching-learning options and find which is most appropriate for this problem. This project is open-ended and you will be expected to write computer programs and understand the mathematical concepts behind the different algorithms.

Student: (Project available)

MR-based planning of ERCP procedures for the investigation and treatment of pancreato-biliary cancers

Supervisors: Dr. Dean Barratt and Yipeng Hu

Pancreato-biliary cancers – i.e. cancers involving the pancreas and pancreatic and biliary ducts – have amongst the highest mortality among cancers, and are most often detected at a late stage. A standard method for investigating ductal tumours is to insert a small endoscope into the ducts via the duodenum, a procedure known as endoscopic retrograde choliopancreatography (ERCP). Instruments can be passed through the endoscope, for example to remove tissue samples, or insert a stent to open up ducts narrowed by disease. ERCP procedures are guided using x-ray imaging. This requires a radio-opaque contrast agent to be injected into the ducts at frequent intervals to visualise the ductal tree. However, navigation through the ductal tree is complicated by the fact that ERCP images are two-dimensional projections of the ductal tree, which can make the procedure technically difficult, especially if a tumour is in an anatomically complex region. One approach to improving surgical navigation is to use magnetic resonance (MR) imaging to obtain a three-dimensional (3D) images of the ductal tree and surrounding anatomy- so-called MRCP. The aim of this project is to extend a simple software tool for planning ERCP procedures using pre-procedural MRCP images. This tool, developed in MATLAB, enables MRCP images of the ducts to be visualised and used as an interactive road map for ERCP planning. The project will involve MATLAB programming and basic image processing, particularly to develop interactive and visual functionality. The performance of these features will be assessed by collaborating clinicians.

Student: (Project available)

Evaluation of inverse source modelling of the cortical evoked response in the anaesthetised rat in comparison with Electrical Impedance Tomography of fast neural activity.

Supervisors: Prof. D. Holder and Brett Packham

EIT is a novel medical imaging method, with which images of the electrical impedance of the head can be produced with a box about the size of a paperback book, laptop and EEG electrodes on the head. It is portable, safe, fast and inexpensive. The supervisor’s research has been to develop its use in imaging functional activity in the brain. It could be used for imaging fast neural activity over milliseconds, which would constitute a major advance in neuroscience methods. Research into this is currently being undertaken in the anesthetized rat. The brain is exposed and recording is undertaken during repeated physiological stimulation of the sensory or visual systems. Recording is undertaken with a mat of 32 cortical electrodes about 6 mm square. EIT data is recorded at the same time as the brain’s own response as well as other data such as intrinsic optical imaging.

Inverse source modelling of the EEG is a method in which the origin of electrical signals is calculated from boundary voltage signals – in this case, the evoked response signals recorded from the epicortical electrode array. In this project, inverse source modelling software from the SPM suite developed at UCL will be adapted for use with the EIT rat data. The results of source modelling will be compared with EIT data.

Skills to be acquired: programming in Matlab, learning SPM, signal processing, and data analysis. All data will be provided by medical researchers.

The project is suitable for a student with a background in physics, engineering or computing, or a medical student with computing experience and an interest in programming.

Student: (Project available)

Development of materials for realistic head-shaped tanks for Electrical Impedance Tomography of the head using a 3D printer

Supervisors: Prof. D. Holder and Dr. Kirill Aristovich

EIT is a novel medical imaging method, with which images of the electrical impedance of the head can be produced with a box about the size of a paperback book, laptop and EEG electrodes on the head. It is portable, safe, fast and inexpensive. The supervisor’s research has been to develop its use in imaging functional activity in the brain. One possible use could be to image changes in the brain during acute stroke over time as the brain pathology evolves. EIT has the unique potential to provide a bedside imaging method for this purpose which would alert medical staff to a deterioration and so lead to improvements in treatment. It could also be used for imaging fast neural activity over milliseconds. Research into this is currently being undertaken in the anesthetized rat.

It is useful to use anatomically realistic tanks for testing and refining EIT systems. Until now, a real skull has been used for the human head tank. It would be desirable to be able to make tanks with the correct electrical properties in larger numbers for comparison across research groups. This could be achieved using 3D printers but it will be necessary to print porous materials which have the correct electrical properties.

The project will be learn about bioimpedance and tissue properties, and also the methods for using a 3D printer. Different porous materials will be designed and evaluated. These will then be constructed into realistic head shaped tanks, either for the rat or human. If time permits, studies of their validity using EIT systems will be undertaken.

Skill to be acquired : use of 3D printer; knowledge of relevant biophysics and material science; experimental design and data and statistical analysis.

The project would be suitable for a student with a background in physics, engineering, computing, or medicine. If medicine, familiarity with computing and some programming is desirable.

Themes: Physics and/or Engineering

Student: Taken up by undergraduate student

Evaluation of the optimum number of electrodes for imaging fast neural activity or acute stroke with Electrical Impedance Tomography : a simulation study.

Supervisors: Prof. D. Holder and Dr. Kirill Aristovich

EIT is a novel medical imaging method, with which images of the electrical impedance of the head can be produced with a box about the size of a paperback book, laptop and EEG electrodes on the head. It is portable, safe, fast and inexpensive. The supervisor’s research has been to develop its use in imaging functional activity in the brain. One possible use could be to image changes in the brain during acute stroke over time as the brain pathology evolves. EIT has the unique potential to provide a bedside imaging method for this purpose which would alert medical staff to a deterioration and so lead to improvements in treatment. It could also be used for imaging fast neural activity over milliseconds. Research into this is currently being undertaken in the anesthetized rat.

Until now, imaging has been undertaken with 16 or 32 electrodes but a new system for use in the rat has the capability to use 128 electrodes and this may expanded to 256. However, it is not clear how many electrodes are optimal. Image resolution increases with increased electrodes but after a certain number additional electrodes confer no additional benefit because of noise.

The work will be to undertake computer simulation to estimate the optimal number of electrodes. The initial phase will be to become familiar with EIT imaging and the use of software for simulation of the imaging problem. Realistic values for noise and an accurate geometric model of the head will be used to determine the result. If time permits, this will be validated in studies in saline filled tanks.

Skill to be acquired : : medical image reconstruction and computer simulation, data and statistical analysis.

The project would be suitable for a student with a background in physics, engineering, computing, or medicine. If medicine, familiarity with computing and some programming is desirable.

Student: (Project available)

Real-time implementation of an algorithm for removing artefact from the EEG in Electrical Impedance Tomography (EIT) of epileptic activity.

Supervisors: Prof. D. Holder and Gustavo Santos

EIT is a novel medical imaging method, with which images of the electrical impedance of the head can be produced with a box about the size of a paperback book, laptop and EEG electrodes on the head. It is portable, safe, fast and inexpensive. The supervisor’s research has been to develop its use in imaging functional activity in the brain. One exciting application lies in its use to image changes in the brain due to epileptic activity. In epilepsy, abnormal activity may occur in the form of seizures in which there is continuous abnormal activity lasting a minute or so. EIT could be used to provide a uniquely new method for imaging brain activity in such seizures which could be used in surgery for epilepsy.

In order for this be realised, EIT needs to be recorded at the same time as EEG over several days in patients on a ward who have been specially brought in for observation. Both are recorded with about 20 electrodes glued to the scalp. Unfortunately, the EIT injects an artefact into the EEG signal. A method for removing this has been developed but it currently takes several minutes of post-processing off-line after the EEG has been acquired. As some clinicians need to see real-time EEG at the bedside as it is collected, it is desirable to run the cleaning algorithm in real time.

The purpose of the project will be to implement and test real-time implementation of the algorithm. Initially, the student will read relevant background literature and become familiar with the algorithm. They will then develop a way to run it in real time, initially on a PC running in parallel with commercial EEG software. If this is not sufficiently fast, then other methods to speed up processing will be investigated, such as the use of a parallel Graphical Processing Unit added to the PC.

Skills to be acquired : programming in Matlab, C or C++, signal processing, and biomedical instrumentation.

The project is suitable for a student with a background in physics, engineering or computing, or a medical student with experience and an interest in programming.

Student: (Project available)

Electrical Impedance Tomography (EIT) of evoked physiological activity.

Supervisors: Prof. D. Holder and Gustavo Santos

EIT is a novel medical imaging method, with which images of the electrical impedance of the head can be produced with a box about the size of a paperback book, laptop and EEG electrodes on the head. It is portable, safe, fast and inexpensive. The supervisor’s research has been to develop its use in imaging functional activity in the brain. One possible use could be to image increases in blood volume which occur over some tens of seconds during normal brain activity, such as during the standard clinical techniques of stimulation of the visual system by flashing lights or the somatosensory system by mild electrical stimulation at the wrist. Such imaging can already be performed by fMRI (functional MRI); the advantages of EIT are that similar images could be acquired with portable much less expensive technology which would increase its availability. EIT data has been collected in these situations before and led to a landmark publication in which reliable single channel data were observed but, unfortunately, the data was too noisy to form into reliable images. Since then, the electronics and imaging software have been improved – for example, we can now collect images at multiple frequencies whereas before they were only collected at one. This gives greater opportunities to reduce noise.

To start with, the student will evaluate a new EIT imaging system, the Kyung Hee Mk 2 system, in a saline filled tank and compare its performance with an older UCLH EIT system. If time permits, then students will work together to collect EIT data during repeated evoked activity in about 10 healthy volunteers, and then will help produce images using Matlab code written for this purpose. Digital photos will be taken around the head, and then photogrammetric software will be used to localise their positions. Images will be reconstructed using an MRI of the patient’s head, which needs to be converted to a Finite Element model with software for segmenting medical images and meshing them. The accuracy of these images will be compared with similar studies using fMRI.

Skills to be acquired: Students will spend time in the lab in Medical Physics at UCL learning relevant methods and analysing the data, and some time in Prof Holder’s department at UCH, learning how to collect evoked responses using scalp electrodes. Skills to be acquired will include one or more of: medical image reconstruction; photogrammetric software use; medical image segmentation and meshing software; EEG electrode placement and use; experimental design and data analysis.

The project would be suitable for a single student or a team of 2 or 3, with a backgrounds in physics, engineering, computing, or medicine.

Student: Taken up by undergraduate student

Canonical Correlation Analysis in the Study of Cerebral Interrelations with Systemic Variables: Application in Acute Brain Injured Patients.

Supervisors: Dr Ilias Tachtsidis and Dr. David Highton

Brain functional near-infrared spectroscopy (or fNIRS) is a technique that uses non-invasive optical reflection measurements to monitor brain tissue haemodynamics and oxygenation by resolving the concentrations of oxygenated haemoglobin (HbO2), deoxygenated haemoglobin (HHb) and brain tissue oxygen saturation (or tissue oxygenation index (TOI)). For several years we have been using this technique to monitor brain tissue physiology and pathophysiology in acute brain injured patients in the neuro-intensive care unit. In addition to our fNIRS measurements we monitor intracranial pressure, blood pressure, arterial saturation and other systemic physiological variables. One of our main interests has been deciphering the relationship between the brain fNIRS measurements with: (1) other brain physiological measurements such as intracranial pressure and; (2) the systemic physiology such as blood pressure. The aim of this project is to use canonical correlation analysis (or CCA) to investigate the above interrelationships.
CCA is a statistical method that analyzes the interrelation between variables in multi-dimensional datasets. CCA can be seen as an extension to normal correlation analysis, in which the proximity between two multidimensional datasets, instead of vectors, is analyzed by means of canonical angles. CCA determines how strongly the variables in both datasets are related. It is also possible to determine which and how many of the independent variables explain most of the variation in the dependent dataset.
The student will be using a CCA toolbox that was developed in our lab with multidimensional data set that collected in the neuro-critical care unit from acute brain injured patients. The project will involve some development on the analysis methodology and will require from the student to analyse data and do some statistics. This project is mainly computational and will be suitable for a student with a general interest in monitoring brain physiology, fair knowledge of MatLab and signal processing methods.

Student: (Project available)

Explore the optical measurements of brain and muscle tissue oxygenation and haemodynamics during a hypoxic ischaemic insult.

Supervisors: Dr Ilias Tachtsidis and Dr Aaron Oliver-Taylor

Monitoring the tight balance of brain blood flow, oxygen delivery and brain tissue metabolic rate is a major aim in patient diagnosis and care. A patient’s health is in great danger when there is a prolonged lack of oxygen delivery to meet the metabolic demand of the tissue; for example in neonatal encephalopathy secondary to birth asphyxia. The Perinatal-Brain Magnetic-Resonance Group at UCLH has a well established programme of research characterising and monitoring neonatal brain injury. Recently as part of collaboration with the Biomedical Optics Research Lab (BORL) they have been measuring brain tissue oxygenation and haemodynamics in a brain injury animal model following hypoxic-ischaemia (reduction of brain blood flow and oxygenation) using an in-house developed near-infrared spectrometer. The main objectives for this project will be: (1) develop and use a low cost optical spectrometer to monitor muscle oxygenation during hypoxic-ischaemia; (2) analyse data from both brain and muscle obtained before, during and after brain injury. The student will need to put together a basic optical spectrometer using available components, develop software for data collection, collect data during experiments and use novel software tools that will allow quantification of the near-infrared measurements and then will focus on analysing those in conjunction with systemic signals and possible magnetic resonance spectroscopy measurements. The larger scope of the analysis is to investigate the biophysical and biochemical changes that happen in birth asphyxiated infants. This project would be suitable for a student with an interest in physiology/pathophysiology, brain tissue biochemistry; and will involve data processing and statistical analysis.

Student: Jessica Johnson

Analysing simultaneous intracranial encephalography (iEEG) and laser Doppler flow measurements to understand the relationship between electrical activity and haemodynamic changes in human epilepsy.

Supervisors: Dr David Carmichael and Dr Ilias Tachtsidis

Epilepsy is the most common serious chronic neurological condition that affects people of all ages and is characterised by recurrent uncontrolled brain activity that is manifested by seizures. In up to 70% seizures are controlled by medication, however for the remainder surgery is an option provided the brain region responsible for seizure onset can be identified. Furthermore, the adverse effects of anti-epileptic drugs can be debilitating. Recently evidence has been growing that haemodynamic changes can precede epileptic events identified on scalp EEG and/or clinical manifestations. This suggests that the electro-clinical hallmarks of epilepsy may lag behind metabolic changes. This knowledge is important to correctly interpret the results obtained from wide range of non-invasive investigations reliant on both electrophysiological and haemodynamic changes such as EEG-fMRI, MEG, SPECT/PET and NIRS. Further, as treatment strategies move towards interventions such as deep brain stimulation and ‘closed loop’ drug delivery to stop or prevent seizures, the best and earliest predictors of possible seizures are required. In this context the information contained within the early haemodynamic changes may be important.
iEEG allows the recording of the brains electrical activity with exquisite sensitivity by directly measuring local electrical field potentials. These recordings show activity across a wide frequency range (dc-kHz) from small regions of cortex. This has made iEEG the gold-standard method for localising the origin of epileptic seizures in individuals with severe, drug-resistant epilepsy. Laser Doppler measurements allow us to measure cerebral blood flow without interfering with the brains ongoing electrical activity or measurements because it is an optical technique.
Outline
We have obtained simultaneous recordings from several patients of these two measures. This project entails two parts; 1. developing a pipeline for data analysis and 2. examining the relationship between the two measurements. We hypothesis that haemodynamic responses will precede electrical discharges.
This project will require writing/modifying programs in matlab to read in the data, and then creating general linear models of the expected haemodynamic responses based on features in the iEEG data and sources of noise. By fitting these models to the data we aim to obtain the average response of the haemodynamic signals to epileptic events visible on iEEG.

Collaborators / tertiary supervisors: Prof. H. Cross (GOSH paediatric epileptologist); Alan Worley (GOSH Clinical scientist and biomedical engineer).

Student: (Project available)

Mathematical modelling of water absorption by superabsorbent hydrogels.

Supervisors: Prof. Alan Cottenden and Dr Becky Shipley

Superabsorbent polymer (SAP) hydrogels are remarkable materials that will absorb 100 times their own dry mass in water, a property exploited, for example, in absorbing body fluids in medical and hygiene products such as diapers and wound dressings, and in agriculture to retain water around germinating seeds and the roots of plants. However, the process of water flow between the SAP particles and diffusion into them is as yet poorly understood. We have a growing body of experimental data describing the kinetics of absorption under a variety of conditions and the aim of this project will be to develop mathematical models that capture the observed absorption behaviour. The student taking this project may have a physics, engineering or computing background, but will need to have strong maths. The work will be jointly supervised by Dr Becky Shipley (mathematician, Dept Mech Eng) and Prof Alan Cottenden (materials scientist, Dept Med Phys).

Student: (Project available)

An investigation of Fourier Transform Infrared spectroscopy for measuring the fibre footprint of fabrics

Supervisors: Alan Cottenden and (to be confirmed)

Urinary incontinence affects some 10% of women and about half as many men in the western world, leading to embarrassment, discomfort and degradation of quality of life. Much incontinence can be cured – or, at least, alleviated – but many people rely on incontinence products to manage their symptoms, of which absorbent pads are by far the most common. Although modern pads are increasingly effective at containing leakage, they often cause skin soreness due to abrasion caused by friction. In Dept Medical Physics we have an established program of work to measure friction between skin and pad materials (nonwoven fabrics) and to understand the mechanism of interaction in order to pave the way for the design and development of fabrics which are less damaging. We have developed an optical microscopy technique which enables us to image and measure the fibre footprint of fabrics on the skin (the interface which mediates friction) but it is laborious and time-consuming. However, preliminary work with Prof Peter Rich (UCL Dept Structural and Molecular Biology) suggests that it should be possible to use Fourier Transform Infrared Spectroscopy (FTIRS) – a technique routinely used by Peter for examining biological specimens – to quantify the contact area between fibres and a support surface, and to do so quickly and accurately. The proposed project aims to investigate this potential by developing the necessary methodologies and comparing data with results from our existing, laborious method. The project will be supervised by Alan Cottenden and Sabrina Falloon.

Student: Manpreet Padwal

Ergometer for therapy after Spinal Cord Injury: Display for Virtual Races

Supervisors: Prof Nick Donaldson and Chin-Wei Liu

A special cycling exercise machine (ergometer) is being built for a clinical experiment in which patients with spinal cord injury exercise while being electrically stimulated. Tacx commercial software is being used to provide the virtual cycle races that will encourage the subjects to exert themselves. At present, the software runs on a laptop and uses a hardware interface to the ergometer itself. However, the subjects of the experiment must use this system at home, where they may not have much space, and some may not be good with computers. The first aim of this project is to select a single-board computer that can run the game and set it up so that the game is very easy to use. The board will be mounted in the ergometer to save space. Instructions must be written for the patients and a manual for the clinical team so that they can use the board in future.
Information about the exercise, such as the time a session started, the torque produced at the crankshaft during the session, the cadence, etc, must be collected. There are commercial web-connected devices that can be programmed to collect measurements and place them in a spreadsheet. These devices can be connected to the internet using broadband from the patients’ homes so that the experimenters can upload data from time to time but they are very expensive. In the second phase of the project, the student will explore ways to do this more cheaply using the single-board computer and perhaps other hardware.
Skills: computers, internet and programming

Student: (Project available)

Models of dispersion in capillary networks

Supervisors: Prof. Daniel Alexander and Rebecca Shipley

The project will construct and evaluate mathematical and computational models of the MRI signal from water molecules in blood flowing through capillary networks. Diffusion-weighted MRI holds great promise as a non-invasive diagnostic probe in cancer imaging. Blood flow in capillaries (the intra-voxel incoherent motion or IVIM effect) is a significant confound that complicates its practical application. However, modelling the effect potentially allows us to exploit diffusion-weighted imaging in cancer more effectively and may even provide additional information that is currently discarded. The project takes the first steps towards this by constructing candidate mathematical models and validating them against computer simulations.

Skills required: theoretical (mathematics); computational (matlab and maybe C/java).

Student: (Project available)

Overcoming Susceptibility Artifacts in Mouse Brain MRI at 9.4 Tesla

Supervisors: Dr Karin Shmueli and Dr Simon Walker-Samuel

Magnetic Resonance Imaging (MRI) at high magnetic field strengths (e.g. 9.4 Tesla) is advantageous as it provides greater signal. However, image artifacts due to air-tissue magnetic susceptibility differences worsen as the field strength increases. Geometric distortion and signal drop-out artifacts caused by large magnetic field inhomogeneities around the ear-canals are a particular problem for gradient-echo MRI of the mouse brain at 9.4 Tesla.

This project will focus on developing and applying practical methods to overcome these susceptibility artifacts to improve preclinical MRI in mouse models. Potential artifact reduction methods include both hardware- and MRI pulse-sequence-based techniques. For example, in passive shimming a strongly diamagnetic substance (with negative susceptibility) can be positioned so as to cancel some of the large magnetic field inhomogeneities that lead to artifacts. Alternatively, the sequence of gradient and radio-frequency pulses used to acquire MR images can be modified to reduce the effects of field inhomogeneities. Strategies you could try include accelerating image acquisition by parallel imaging using multi-channel radio-frequency coils to reduce the time during which field inhomogeneities can cause artifacts. Other sequence-based techniques (Z-shimming or RF-shimming) involve measuring the magnetic field distribution and designing slice-selective gradients and/or radio-frequency pulses tailored to compensate for field non-uniformities.

Experimenting on the 9.4 Tesla MRI system at the Centre for Advanced Biomedical Imaging (CABI), you will evaluate, implement and compare several of these strategies to determine which is the most effective in reducing susceptibility artifacts in the mouse brain. This project will involve MRI experiments, image-processing (using Matlab and other software) and, potentially, MRI pulse sequence programming.

Student: Alim Yucel-Finn

Performance evaluation of elite cyclists with optical monitoring

Supervisors: Dr Terence Leung and James Hopker

Winning or losing a cycling competition is often determined during the final stages of a race. Cyclists with sufficient reserve may be able to overtake rivals in the last few miles. A key determinant of such endurance performance has been suggested to be oxygen consumption of the leg muscles. The aim of this project is to develop a new technique to calculate leg muscle oxygen consumption based on the data collected by an optical oximeter attached to the thighs of elite cyclists during training. This project is in collaboration with the Endurance Research Group (http://www.kent.ac.uk/sportsciences/index.html) in the School of Sport and Exercise Sciences at the University of Kent, where the student will have the opportunity to participate in the performance tests of elite cyclists. This project will suit a student with an interest in sport science and will involve mathematical modelling, signal processing and data analysis.

Student: Christy M C Moen

MRI Susceptibility Mapping of Oxygenation in Head and Neck Tumours

Supervisors: Dr Karin Shmueli and Dr Shonit Punwani

In patients with head and neck squamous cell carcinoma, the cancer may spread by metastasising to nearby cervical lymph nodes. There is also a high tumour recurrence rate after radiotherapy treatment. In the process of radiotherapy treatment planning, patients are scanned with various MRI sequences. The overall aim is to detect metastatic cervical nodes, differentiate them from benign nodes and predict tumour recurrence after treatment with radiotherapy.

The aim of this retrospective study will be to compare MRI phase images with the patient breathing air v. oxygen and to calculate tissue magnetic susceptibility maps from those images. These maps are expected to give an indication of the tissue oxygenation and may highlight metastatic tissue. MRI phase images are interesting because they provide complementary information to conventional MRI images based on the signal magnitude. Phase images arise primarily from the underlying magnetic susceptibility distribution of tissues. However, phase information is non-local and orientation-dependent, making it difficult to interpret. Susceptibility maps calculated from phase images overcome these disadvantages and reflect the tissue composition much more closely. Oxygen is paramagnetic (has a positive susceptibility), so we expect highly oxygenated tissue to show up bright in susceptibility maps. Tumours and metastatic lymph nodes have an abnormal blood supply with respect to healthy tissue and are therefore expected to have different levels of oxygenation, especially when patients are breathing oxygen v. air. You will calculate and compare phase and susceptibility images of patients with head and neck tumours and healthy controls and look for differences in susceptibility between metastatic and benign lymph nodes and between tumours and healthy tissue. Grouping patients by treatment outcome will also allow you to investigate whether values derived from phase and susceptibility images help predict tumour recurrence after radiotherapy.

This project will involve analysing MR images using Matlab. Image processing will range from basic image arithmetic and region-of-interest analysis through to image registration and susceptibility map calculation.

Student: Thomas Durnall

Optimization of the experimental set-up in x-ray phase contrast imaging through comparison between simulated and experimental data

Supervisors: Dr. Alessandro Olivo and Dr. Marco Endrizzi

X-ray phase contrast imaging is a new imaging modality which exploits interference and refraction effects instead of x-ray absorption. As a consequence, it has the potential to transform all applications of x-ray imaging (first and foremost diagnostic radiology), as it increases the visibility of all details in an image and it enables the detection of features classically considered x-ray invisible. This modality was considered to be restricted to very specialized facilities called synchrotrons for many years, but our group has recently developed a method which makes it work with conventional x-ray sources, thus potentially enabling its clinical translation.

This project deals with the optimization of the experimental set-up. In particular, the combination of source-to-sample and sample-to-detector distances resulting in maximum image contrast will have to be determined. This will firstly be done by means of (already existing) simulation code, followed by the experimental validation of the results. The student will learn how to use the simulation code, acquire images using the experimental set-up, analyze the data, and ultimately compare simulated and experimental results – hence gaining experience on all links in the chain.

Some preliminary experience with running computer simulations and performing data analysis with simple programs like excel and image J would be required.

Student: Eftychia Nafti

Exploring the validity limits of a quantitative phase retrieval method working with incoherent x-ray sources

Supervisors: Dr. Paul Diemoz and Dr. Alessandro Olivo

X-ray phase contrast imaging is a new imaging modality which exploits interference and refraction effects instead of x-ray absorption. As a consequence, it has the potential to transform all applications of x-ray imaging (first and foremost diagnostic radiology), as it increases the visibility of all details in an image and it enables the detection of features classically considered x-ray invisible. This modality was considered to be restricted to very specialized facilities called synchrotrons for many years, but our group has recently developed a method which makes it work with conventional x-ray sources, thus potentially enabling its clinical translation.

We have recently achieved a significant step in that we have developed a “quantitative” phase imaging method i.e. a way to separate absorption and phase information, thus obtaining two separate images containing complementary information. It is the first time this has been done with a completely incoherent source (which part of the community thought would have been impossible), and this has resulted in a recent publication in a high profile journal (PNAS). The student will be asked to explore the validity conditions of this “phase retrieval” approach using both existing simulation software and experimental data. This is a complex project encompassing all aspects of data simulation, acquisition and analysis thus leading to a wide range of acquired skills.

Preliminary experience with running computer simulations and performing data analysis is required; some familiarity with the theory of wave propagation highly advisable.

Student: (Project available)

Characterization of detector performance for x-ray phase contrast imaging

Supervisors:Dr. Marco Endrizzi and Dr. Alessandro Olivo

X-ray phase contrast imaging is a new imaging modality which exploits interference and refraction effects instead of x-ray absorption. As a consequence, it has the potential to radically transform all applications of x-ray imaging (first and foremost diagnostic radiology), as it increases the visibility of all details in an image and it enables the detection of features classically considered x-ray invisible. This modality was considered to be restricted to very specialized facilities called synchrotrons for many years, but our group has recently developed a method which makes it work with conventional x-ray sources, thus potentially enabling its clinical translation.

This method however is highly affected by the performance of the used x-ray detector. Not only does this cover classic parameters like noise and spatial resolution, but also more sophisticated and method-specific ones like signal spill-out between adjacent pixels. The student will be required to thoroughly characterize the available x-ray detectors (especially of a state-of-the-art direct conversion flat panel detector based on amorphous selenium), and assess the impact of the extracted parameters on the phase contrast performance of the devices. The skills the student will gain go beyond phase contrast imaging and also cover many aspects of x-ray detector technology and characterization, which are relevant to diagnostic radiology in general.

Some familiarity with the basics of data acquisition and analysis are necessary.

Student: (Project available)

Investigation into the use of a laser to provide positioning information for naso-gastric feeding in neonates.

Supervisors: Paul Ganney and Sandy Mosse

An NHS “never event” is one that should never happen. The misplacement of a naso-gastric feeding tube (i.e. into the lungs instead of the stomach) is one such. This project intends to pass laser light down the catheter via a fibre-optic cable in order to illuminate the tip and thereby provide a reliable visual indication as to correct (or otherwise) placement. Experiments will involve phantoms and organic material.
Skills required: Electronics, Physics, Experimental.

Student: (Project Available)

A real-time bed-state monitor

Supervisors: Paul Ganney and Paul Ostro

Hospital bed management is heavily reliant upon timely data. This project will construct a traffic-light based bedside monitor which communicates to a web server, providing instant information. Whilst the concept is simple, the execution will be more complex in eliminating false readings.
Skills required: Electronics, Computing.

Student: (Project available)

Creation of a contact network for a Hospital for infection control purposes.

Supervisors: Paul Ganney and Cliff Ruff

There has been much work done in recent years on using network models to simulate the transmission of infection. Some of this (Ancel Meyers, Ganney) has looked at mapping a particular setting, rather than using randomly generated networks. The aim of this project is to take the methods described in Ganney's PhD Thesis (2011, Hull) and apply them in two ways to the UCLH. The first is questionnaire-based, replicating the work undertaken by Ganney. The second is using UCLH incidence data to generate a contact network. The final part of the project will be a comparison of the two models and an attempt to explain any discrepancies that may exist.
Skills: Computing is necessary. A background knowledge of infection propagation would be useful, but is not essential.

Student: (Project available)

Measurement of absolute concentration of N-acetyl-aspartate in the brain

Supervisors: Alan Bainbridge and David Thomas

Neonatal encephalopathy (NE) often occurs unexpectedly following an otherwise uneventful pregnancy; 20 to 25% of the affected infants die in the first few days after birth and up to 75% of the survivors may develop significant life-long disabilities. There is a continuing need to develop novel treatment strategies for these infants. Our group at UCL runs an experimental model of NE which is used to test potential treatments prior to running clinical trials in humans. This model played an important role in the development of hypothermia as a clinical standard of care. The use of inhaled Xenon gas for neuro-protection is currently undergoing clinical trials following development work in this model.

Magnetic resonance biomarkers can serve as surrogate end points in clinical trials and studying the same biomarkers in humans and in our experimental model improves our ability to translate results into the clinic. A potentially very important biomarker is the absolute concentration of N-acetyl-aspartate (NAA). NAA is the second most abundant amino acid in the nervous system and is almost exclusively neuronal; NAA concentration is therefore a potential surrogate index of neuronal survival following injury. The aim of this project is to develop a protocol for measuring the absolute concentration of NAA in the brain using magnetic resonance spectroscopy. The student will have the opportunity to run experiments on the 9.4 Tesla research scanner installed at the Institute of Neurology and will develop software tools for processing the data. The project will run alongside an ongoing study developing this technique for use in human infants.

This project is part experimental and part computational. Some programming skills (in Matlab or other suitable language) would be useful but these skills could be acquired during the course of the project.

Student: (Project available)

Deformable registration of Cine MR images of the lungs

Supervisors: Jamie McClelland and Marc Modat

Breathing causes motion and deformation within the lungs, which can be problematic for radiotherapy treatments and other image guided interventions and image reconstruction. Although there has been considerable interest in respiratory motion in the past few years, most studies look at data acquired over only one or a few breath cycles, despite it being well known that respiratory motion can vary considerably from one breath to another.
Using MR imaging we have repeatedly acquired a 2D sagittal slice through the lungs for approx 3 minutes, acquiring about 4 slices every second, in order to study the respiratory motion and its variation. The first step in analysing the data is to use deformable registration to recover the motion from one image to the next. In order to do this a 2D B-spline registration will be extended to a 3D B-spline registration, where the 3rd dimension is time. This can then be used to recover the motion from one image to the next, while constraining the motion between successive frames to change smoothly. If this work is successful there will be the possibility of extending the concept to perform 4D registrations, where the 4th dimension is time and the images are 3D volumes. This will be implemented using NiftyReg, part of the NifTK open source software package developed in the Centre for Medical Image Computing.
This project will require very strong computing skills with experience of coding in C/C++ and an interest in medical image processing and analysis.

Student: (Project available)

Deformable registration of Cine MR images of the lungs

Supervisors: Jamie McClelland and James Martin

Breathing causes motion and deformation within the lungs, which can be problematic for radiotherapy treatments and other image guided interventions and image reconstruction. During many procedures it can be difficult to directly measure the respiratory motion of the internal anatomy. Therefore motion models which relate the motion to an easily measured respiratory surrogate signal, such as the displacement of the chest/abdomen skin surface, have been proposed. These can then estimate the internal motion from the surrogate signal during the procedure. There have been many such models proposed over the last 5-10 years, but as yet very few studies comparing different models on the same dataset.
Using MR imaging we have repeatedly acquired a 2D sagittal slice through the lungs for approx 3 minutes, acquiring about 4 slices every second. In these dataset both the internal structures (airways and blood vessels) and the external skin surface can be seen, and they can be used to quantitatively compare the performance of different respiratory motion models. This project will compare the performance of several of the most promising motion models proposed in the literature, but will also give the opportunity for the student to develop their own novel motion model and to compare it against those from the literature. If successful this may lead to a conference or journal paper publication.
This project will require strong computing and mathematical skills. Experience with Matlab is essential.

Student: (Project available)

Analysis of Monitoring Data to Maximise Understanding of patient condition in Intensive Care

Supervisors: Julian.Henty and Nick Everdell

Bedside monitoring data is routinely acquired from a number of pieces of completely independent monitoring equipment in Intensive Care. Collectively these data could provide a much bigger picture regarding the patient’s condition. By acquiring these data using one, central computer we can lump everything together, and move our understanding much closer to that bigger picture.
We already know that by combining these routinely acquired data it is possible to provide additional physiological information, but there is currently no facility to achieve this. Using the system described here we can provide new physiological information that is proven but currently unavailable, and thereby improve our understanding of the patient’s condition. This will be achieved using a single computer system with hardware to collect data and software to analyse these data in real-time.
Using a PC with a number of serial ports, data will be collected from the independent machines such as monitors and ventilators. These data will then be stored and processed using LabVIEW software. Less routinely used equipment, such as a near infrared spectrometer, will also be connected to the PC. Online information regarding the setting up of such equipment can be provided for the operator (nurse/doctor) in order to achieve the highest level of data quality possible, and this could be used in conjunction with a training record database. There will be opportunities to visit intensive care at the Royal London Hospital for learning and testing purposes.
The theme for this project is a combination of Engineering and Computing.

Student: (Project available)

Evaluation of the Use of Near Infrared Spectroscopy in Cerebral Oxygenation of Left and Right Brain Hemispheres and in Detection of Compartment Syndrome in Critically Ill Intensive Care Patients

Supervisors: Julian.Henty and Nick Everdell

The Adult Critical Care Unit at the Royal London Hospital has recently acquired a Near Infrared Spectrometer (NIRO 200) which we have used to measure cerebral oxygenation in left and right brain hemispheres of patients with head injury. We wish to further evaluate use of this equipment for this purpose by setting up the machine routinely on suitable patients, and also begin testing for compartment syndrome in sedated patients. Compartment syndrome is a condition resulting in tissue death in limbs due to the compression of blood vessels (also nerves and muscle) within a closed space (a compartment). Since patients are sedated there is generally no indication that oxygen transport to tissue is compromised (pain is usually the best indication of this disease). A dedicated laptop will be used to acquire data from the spectrometer in both cases, and we will begin collecting data routinely very shortly.
In this project we require a student to analyse large amounts of data, using either dedicated software supplied by Hamamatsu and/or software written in LabVIEW, by the student. The project would suit a student who has some experience with large volume data analysis and if necessary computer programming, preferably in LabVIEW (or is willing to learn). There will be opportunities to visit intensive care at the Royal London Hospital for learning and data collection purposes.
The theme for this project is a combination of Physics and Computing.

Student: (Project available)

Continuous EEG Monitoring for Critically Ill Intensive Care Patients

Supervisors: Julian.Henty and Nick Everdell

The Adult Critical Care Unit at the Royal London Hospital routinely employs EEG for the detection of nonconvulsive seizures in sedated patients. Such seizures are associated with worst outcome, yet these clinically silent neurological events can otherwise go undetected. The Unit has only two EEG machines, each one with two channels. In this project it is proposed that a four channel EEG machine be constructed, consisting of amplifiers, filters, a PC with an analogue to digital converter and software written in LabVIEW. In addition, two extra channels can be used for EMG/EOG for determining sleep states.
Along with data collection and storage, the EEG signals would need to be processed in real time in order to provide spectral analysis (Delta, Theta, Alpha, Beta etc. bands) and display these data in a graphical form. There is also a requirement for software that is user friendly and would enable the user (nurse/doctor) to set the equipment up as simply as possible (i.e. online information regarding electrode placement and acquired signal quality). Finally, further signal processing could lead to some intelligent analysis of EEG signals, providing automatic seizure detection.
The project would suit a student who has some experience with computer programming, preferably in LabVIEW (or is willing to learn). Also, experience with FFT signal analysis in any programming language is ideal. EEG training will be provided and there will be opportunities to visit intensive care at the Royal London Hospital for learning and testing purposes.
The theme for this project is a combination of Engineering and Computing.

Student: (Project available)

Noninvasive Continuous Cardiac Output Monitoring

Supervisors: Julian.Henty and Nick Everdell

The Adult Critical Care Unit at the Royal London Hospital routinely measures cardiac output in order to assess the state of a patient’s blood circulation. Although the majority of patients are adequately monitored using heart rate and blood pressure, some patients with cardiovascular abnormalities require more detailed measurements. Two methods are used in the Unit: (1) lithium dilution, where a bolus of lithium is injected into the patient and a sensor detects the ejection of this bolus after a short delay and (2) transesophageal Doppler, which uses an ultrasound probe to measure the Doppler shift reflected by red blood cells in the descending aorta, and thereby measure blood flow velocity in this vessel. Disadvantages of method (1) are that it is time consuming to calibrate (and must be repeated every 24 hours) and costly in terms of associated consumables; disadvantages of method (2) is that a reasonably skilled operator is required, and the probe is single use and fairly costly. Neither of these methods can be defined truly continuous, and both are classified as “minimally invasive”.
It is proposed in this project to evaluate the concept of non-invasive continuous cardiac output monitoring. Since cardiac output is the product of heart rate and stroke volume (SV), an acoustic method of measuring SV, combined with a calibration procedure (and of course heart rate measurement) would theoretically meet the criteria. The acoustic signal would be collected using a transducer connected to an analogue to digital converter and a PC running LabVIEW software.
The project would suit a student who has some experience with computer programming, preferably in LabVIEW (or is willing to learn). Also, experience with transducers in any programming language is ideal. There will be opportunities to visit intensive care at the Royal London Hospital for learning purposes.
The theme for this project is a combination of Physics, Engineering and Computing.

Student: Ilektra Kompou

Mapping epileptic and cognitive networks in patients with epilepsy.

Supervisors: Dr David Carmichael and Dr Maria Centeno

Background
Epilepsy is a disease characterised by seizures which are abnormal synchronous electrical discharges over large regions of the brain leading to externally visible manifestations. Seizures are treated successfully with drugs in around 70% of patients; in the remaining 30% surgery can be an effective treatment. The aim of both surgery and drug therapy is to stop the occurrence of seizures with the least limiting impact on cognitive function. Abnormal synchronous electrical discharges are also seen between seizures. Both of these processes can potentially interrupt normal cognitive brain function.Mapping of the epileptic networks together with the identification of cognitive network dysfunction are crucial part of the pre-surgical evaluation of patients with epilepsy.Novel imaging techniques such as EEG-fMRI can produce maps of the networks involved is seizure generation and spread. Resting fMRI can be used to study connectivity of cognitive networks.
We have been collecting simultaneous measurements of EEG and fMRI in children with epilepsy as well as in healthy children.
(http://www.action.org.uk/our_research/epilepsy_improving_brain_scanning_surgery)
Aim
We aim to understand the correlation between those areas involved in seizure generation and spread and the connectivity of the different cognitive networks in the brain.
Outline
We hypothesize that resting state networks collocated with the epileptogenic network will show a greater disruption compared with those that does not share the same spatial location as the epileptic networks.
This will be tested by:
1. Network connectivity during resting state will be analysed using independent component analysis and compared with connectivity of networks in healthy controls.
2. Epileptic networks will be generated by modelling the epileptic brain activity captured with the EEG in a general lineal model using SPM.
3. Comparison between spatial location of epileptogenic network and the disruption of the cognitive networks obtain from resting state will be made intra subject.
Skills
This project will involve use existing software for processing neurioimaging data (e.g. SPM and FSL) and can involve writing code in matlab to perform more novel analysis to characterise connectivity and develop metrics for comparison of spatial similarity of networks.

Student: (Project available)

Comparison of reproducibility of cerebral blood flow measurements made using arterial spin labelling MRI at 3T

Supervisors: Dr David Thomas and Dr Enrico De Vita

Introduction
Arterial spin labelling (ASL) is a non-invasive MRI method for measuring tissue perfusion. The method has been used most widely in the brain to measure cerebral blood flow (CBF), but its application has so far been mostly restricted to research studies. We are keen to implement ASL on our scanners at the National Hospital for selected clinical investigation and as a standard option for clinical research studies. The aim of this project will be to analyse test-retest ASL data from healthy volunteers, acquired with a range of protocols, to quantitatively assess the precision of the CBF estimates provided by ASL.
Work involved in the project
The student will have access to ASL data acquired from two MR scanners: a 3T Siemens Trio scanner at the High Field Lab, and a wide-bore 3T Siemens Skyra at the National Hospital (both at Queen Square). Two different ASL sequences have been used, one using a multi-slice 2D readout (EPI) and one using a 3D readout (3D GRASE). These data will allow assessment of within-subject, within-session and between-session reproducibility. Data will be processed using freely available software including FSL (http://www.fmrib.ox.ac.uk/fsl/) and SPM (www.fil.ion.ucl.ac.uk/spm).
The initial part of the project will aim to investigate the best method to realign/register/normalise the datasets that will enable the tests described above. Software developed within the UCL Centre for Medical Image Computing (CMIC) will also be available for this task. If necessary, acquisition of additional healthy volunteer data will be planned and performed.

Skills required for the project
In order to carry out this project, the student will be expected to:
• Understand the concept of using magnetically labelled blood water to measure tissue perfusion with MRI
• Understand and apply the theoretical models required to convert signal in the ASL images to quantitative CBF maps
• Understand the concepts of spatial realignment, registration and normalisation and apply available software to ensure data from different acquisitions can be compared on a voxel-by-voxel basis
• Test and select the most promising/sensitive statistical methods for assessing reproducibility and compare ASL data with different protocols

Student: (Project available)

Investigation of the optimal acquisition approach for generating isotropic 3D double inversion recovery (DIR) MR images

Supervisors: Dr David Thomas and Dr Olga Ciccarelli

Introduction
Double inversion recovery (DIR) images have been shown to be extremely useful for identification of the grey matter lesions that occur in multiple sclerosis (MS). As a result, DIR sequences are now being included in standard clinical MRI scanning protocols for MS patients. By using appropriate inversion times, DIR images selectively null signal from white matter and cerebrospinal fluid, leaving only signal in the grey matter. However, due to the contrast mechanisms involved, the acquisition of a 3D high resolution (1mm isotropic) whole brain DIR image requires a relatively long scan time (usually > 10 minutes), which is undesirable in the clinical context. This project will investigate the feasibility of generating synthetic DIR images using other images which are also acquired during routine clinical scanning. Image quality and contrast will be compared between the acquired and synthetically generated DIR images, with a view to assessing whether the DIR acquisitions can be removed from the scan protocol.
Work involved in the project
The student will use images which have been acquired as part of an on-going study in the Department of Brain Rehabilitation and Repair using a 3T Siemens Trio MRI scanner. As part of the scan protocol, proton density and T1-weighted images are acquired, from which quantitative T1 maps are generated. Using these maps, the student will generate artificial DIR images, simulating the pulse sequence parameters used in the actual DIR acquisition. The resulting images will be compared quantitatively to assess differences in contrast-to-noise ratio (CNR), and visually inspected to evaluate differences in image resolution and blurring.

Skills required for the project
In order to carry out this project, the student will be expected to:
• Understand the basic contrast mechanisms underlying T1-weighted MR imaging
• Have some experience using Matlab for MR image manipulation and processing
• Be keen to extend their Matlab programming skills to simulate MR signal behaviour and generate new images using the information contained in MR scans

Student: (Project available)

Advancing knowledge of kidney function using arterial spin labelling MRI

Supervisors: Dr David Thomas and Dr Isky Gordon

Background
MRI is non-invasive and does not use ionising radiation. It provides detailed anatomical images of the internal organs. Recent advances in MRI have been directed towards functional assessment of the brain. Some of these techniques are now being evaluated to assess kidney function. Our unit has been undertaking research using MRI into kidney function for over 5 years.
Description
Arterial spin labelling (ASL) is a non-invasive MR method for assessing the supply of blood (blood flow or perfusion) to an organ. ASL has been applied extensively to studies of the brain, but has so far had limited application in the body. Non-invasive repeatable measurements of renal perfusion could prove invaluable in early diagnosis and management of renal diseases. The aim of this project is to determine the robustness of ASL to measure renal perfusion in healthy kidneys, and to explore its potential use in the clinical environment. The technique for analysis of the perfusion MRI renal data has been developed recently and will now allow us to evaluate the potential of ASL in healthy volunteers before we apply it to patients with kidney disease.

From this project, the student will gain the following expertise:
• Detailed understanding of the structure and function of the kidney
• Working in a unit with a team of MRI experts (physicists, computer scientists, medical doctors and statisticians)
• Understanding of the physical and physiological basis of MRI
• Close supervision in the processing of the ASL data
• Ability to attend the regular seminars of the unit
Skills required for the project
In order to carry out this project, the student will be expected to:
• Use Matlab to generate quantitative perfusion maps from ASL data
• Analyse these maps to estimate the relative importance of acquisition parameters and post-processing steps (such as image realignment)
• Generate estimates of the accuracy and precision of ASL renal perfusion measurements, in order to assess the viability of clinical application of the technique.

Student: Singh, Maheep

Measuring axon permeability using DEXSY MR imaging.

Supervisors: Dr Bernard Siow and Prof Daniel Alexander

The permeability of axons is abnormal in de- and dys- myelinating diseases (for example multiple sclerosis and schizophrenia), thus its measurement could serve as an biomarker for early detection and classification of these diseases. The project will investigate the sensitivity of a novel magnetic resonance imaging technique (MRI), Diffusion EXchange SpectroscopY (DEXSY) MRI, to axon permeability. This will be the first time that DEXSY will be used in an imaging application.

The first stage of the project will be to design and refine the imaging parameters of a DEXSY MRI protocol for sensitivity to axon permeability in white matter regions of a rat brain ex-vivo. Data will be acquired on a 9.4T preclinical scanner. A postprocessing pipeline will need to be adapted for DEXSY imaging. The interpretation of the resultant data would inform future projects that would lead to the use of DEXSY in clinical scenarios. Furthermore, the findings will help improve biophysical models of white matter.

Student: (Project available)

Determining the relationship between estimates of axon radius using diffusion MRI and scanning electron microscopy.

Supervisors: Dr Bernard Siow and Simon Richardson, and Prof Daniel Alexander

Axon radius has abnormal distribution in numerous white matter diseases such as motor neuron disease and autism. Reliable estimates of axon radius would serve as early detection and classification biomarkers for these diseases. Recently, there has been increasing efforts to estimate the radius of axons in white matter regions of the brain. Of particular interest are diffusion Magnetic Resonance Imaging (MRI) techniques that use biophysical models of tissue. Diffusion MRI is non-invasive but the biophysical models of tissue are simple compared to the complex microstructure of actual tissue. Scanning electron microscopy (SEM) can directly image the complex microstructure of tissue but would require highly invasive biopsy, which is highly inadvisable in clinical scenarios! This project will determine the relationship between estimates of axon radius from biophysical models of tissue using diffusion MRI protocols and those from SEM. The findings will aid in the efforts to quantify axon radius non-invasively and improve biophysical models of tissue.

Project outline:
1. Scan biological samples (ex-vivo rat brain) using a rich diffusion MRI protocol on a 9.4T pre-clinical scanner.
2. Scan white matter regions of the brain using SEM.
3. Estimate axon diameter using diffusion MRI data for various biophysical models using software provided.
4. Data analysis and interpretation.

Student: (Project available)

Erythrocyte ghost phantoms for pore shape imaging

Supervisors: Dr Bernard Siow and Prof Daniel Alexander

Pore shape imaging using diffusion Magnetic Resonance Imaging (MRI) has recently been the subject of several high impact publications. Pore shape imaging is of interest in numerous industrial and medical applications, for example porous carbonate rocks containing oil and the shape of cells within the body. MR phantoms comprised of erythrocyte ghosts are highly flexible, with cell size, density, orientation and shape being controllable.

The aim of this project is to set up a pipeline for the design and construction of erythrocyte ghost diffusion MRI phantoms for validation of a unique theoretical framework being developed at UCL. The findings will provide an essential piece of the puzzle in the on-going work that would potentially lead to applications clinical scenarios.

Student: (Project available)

Monte Carlo modelling of scattering experiments

Supervisors: Robert Speller and Nick Calvert

Monte Carlo modelling is frequently used to study problems in radiation physics (and many other areas). This project is to develop code that can study one of several problems in scattered radiation fields. It is purposely general in its nature as the particular problem would be decided in conjunction with the student.
This project requires someone with interests and skills in computing. There is little experimental work involved but some validation of the models will be required that could include limited experimental measurements.

Student: (Project available)

Characterisation of silicon photomultipliers

Supervisors: Robert Speller and Nick Calvert

Recent developments in photo detectors includes solid state photomultiplier tubes. These have many advantages and this project is to characterise some of these devices with a view to including them in the development of novel radiation detectors.
The project is experimental and hence requires a student who has good experiment skills.

Student:(Project available)

Basal ganglia parcellation and motor function in children with cerebral palsy

Supervisors: Dr Chris Clark and Dr Neil Wimalasundera

Cerebral palsy can be caused by a range of brain lesions, including malformations, periventricular white matter injury, or focal cortical infarcts and often the basal ganglia and thalamus are involved. Advanced magnetic resonance neuroimaging techniques capable of determining the structural integrity of white matter tracts have been used to reveal the abnormal connectivity of the brain in CP. A key question is to determine how the damage to specific white matter tracts is related to loss of motor function. If this is known at an early stage it will be possible to predict future motor deficits and better inform the appropriate course of rehabilitation. This project aims to address this question by using new imaging techniques we have recently developed for measuring the size of the portions of the thalamus that connect to the motor and sensory parts of the cortex of the brain in individual children. The thalamus fulfills a key function in motor control as a relay centre, subserving both motor and sensory mechanisms with connections to and from the cortex, the basal ganglia, and cerebellum. We hypothesise that thalamo-sensorimotor cortical connections are disrupted in CP and the degree of damage measured in terms of the volume of thalamus that projects to these cortical areas and measures of structural integrity in pathways projecting from these areas correlate with measures of motor function in CP.

Skills required: some computing either with C or matlab would be an advantage but not essential.

Broser P, Vargha-Khadem F, Clark CA. Robust sub-division of the thalamus in children based on probability distribution functions calculated from probabilistic tractography. Neuroimage 2011; 57(2): 403-415.

Draganski B, Kherif F, Kloppel S, Cook PA, Alexander DC, Parker GJ, Diechmann R, Frackowiak RS. Evidence for segregated and integrative connectivity patterns in the human basal ganglia. J Neurosci 2008; 28: 7143-52.

Student: (Project available)

Automated tractography for mapping white matter tracts in children with brain tumours

Supervisors: Dr Chris Clark and Dr Darren Hargrave

Diffusion MRI based tractography of white matter tracts is playing an increasingly important role in neurosurgical planning, allowing the mapping of eloquent pathways in the living human brain, thus helping the avoidance of damage to these pathways and the focal neurological deficits that would arise. However, current methods involve user interaction in defining seed regions of interest for tractography that is both time-consuming and subjective. We have developed new automated procedures for mapping white matter tracts, such as the arcuate fasciculus, as shown in the figure below. The proposed project will evaluate this method in a cohort of children with brain tumours for which we have diffusion tensor imaging available for tractography. This will be an important step in translating this technology into the clinical arena.

Skills required: some computing either with C or matlab would be an advantage but not essential.

Clark,C.A., Byrnes,T. (2011). DTI and tractography in neurosurgical planning. Chapter 36 in Jones,D.K. (ed.) Diffusion MRI: Theory, Methods and Applications. Oxford University Press, 589-607. ISBN: 9780195369779.

Student: (Project available)

Graph theoretical measures of brain disconnection in children with epilepsy and relation to cognitive dysfunction

Supervisors: Dr Chris Clark and Dr Jon Clayden and Professor Rod Scott

The brain can be thought of as a distributed network (like a plan of the railways). New approaches to the understanding of brain function and organisation have used a method called ‘graph theory’ applied to brain imaging data to describe these networks and how efficiently they work (like how long or how many stops it takes to get between destinations). These new methods have recently been successful in predicting IQ in individuals and have been shown to predict cognitive performance (e.g. how well people can do tests involving language, memory and spatial tasks). Abnormal connectivity of brain networks has been suggested as a way that brain function might be impaired in epilepsy. With the development of these new imaging tools it is now possible to investigate this idea. The purpose of this project is to measure brain networks in children with epilepsy and to determine if such connection measures relate to cognitive performance. If so, these new tools may be used in the future to predict the how a child will develop and to test the effectiveness of disease modifying therapies. Because the technology is available now the results of this study would be applicable to patients in the short term.

Skills required: some computing either with C or matlab would be an advantage but not essential.

Vaessen MJ et al. 2011. White matter network abnormalities are associated with cognitive decline in chronic epilepsy. Cereb Cortex. 2012 Sep;22(9):2139-47. Epub 2011 Oct 29.

Student: (Project available)

Catheter Design and Localisation

Supervisors: Dan Stoyanov and Adrien Desjardin

Localising the shape and position of guidewires and catheters within the vascular anatomy is critical for endovascular surgery and the precise delivery of interventional radiology treatments. This project will focus on the development of methods for catheter localisation using customized catheter design with integrated markers that can be used to infer the catheter shape from X-Ray fluoroscopy images. The project has two potential tracks depending on the skills and experience of the student A) dealing with catheter design and fabrication to produce a customised device B) dealing with processing fluoroscopy images to detect the device and infer its 3D geometry and position. The project is therefore suitable to students with a bioengineering or mechanical engineering background for track A) and students with a computer science, programming and image processing or computer vision background for track B. More information can be obtained at: http://www0.cs.ucl.ac.uk/staff/Dan.Stoyanov/teaching.html

Student: (Project available)

Analysing instrument motion for understanding and objectively measuring operator learning curves and skill in Transesophageal Echocardiography (TOE) procedures.

Supervisors: Dan Stoyanov and Adrien Desjardin

This project will focus on analysing instrument motion for understanding and objectively measuring operator learning curves and skill in Transesophageal Echocardiography (TOE) procedures. The tool motion will be recorded from an advanced simulation environment for developed by HeartWorks (http://www.heartworks.me.uk). TOE is now routine for monitoring cardiac function and surgical treatment in the operating theatre, as well as on the cardiac ITU. TOE requires the insertion of an ultrasonic transducer into the gastroesophageal tract and the correct manipulation of the transducer to visualise multiple planes through the heart. The HeartWorks simulator allows trainees to develop the core skills required for performing TOE examinations but currently the learning process and the relationship between the motion of the transducer and the skill of the operator is poorly understood. The aim of this project will be to investigate the rich information provided by the simulation environment in order to develop evaluation metrics that can provide real-time feedback during training and quantitative assessment methods. Development will initially be performed using Matlab and will involve a combination of signal processing techniques such as Dynamic Time Warping (DTW), dimensionality reduction algorithms like Principal Component Analysis (PCA) and machine learning techniques for classification. More information can be obtained at: http://www0.cs.ucl.ac.uk/staff/Dan.Stoyanov/teaching.html

Student: (Project available)

Computational approaches to correcting geometric distortion in MR: application to the mapping of brain microstructure using NODDI

Supervisors: Gary Hui Zhang and (not allocated yet)

- Theme: Physics; Computing
- Description: Geometric distortion is a common artifact plaguing images acquired using magnetic resonance imaging (MRI). Such artifact has to be removed from the affected images before they can be analyzed to assist clinical diagnosis of diseases. There are ongoing interests in identifying the technique that can produce the most artifact-free images. This project will compare a recently developed technique against the standard approach in the field. The assessment will be made in the context of brain microstructure mapping using MRI, which is playing a key role in image-assisted diagnosis, e.g., of dementia. The student will use FSL, a popular and well-documented software suite that implements both correction techniques, to process and analyze a set of existing MRI data for mapping brain microstructure. The aim of the project is to identify the correction technique that is most suitable for the mapping of microstructure.
- Requirements: Basic math and computing skills
- Skills to be acquired: In-depth understanding of MRI; MR data visualization and analysis with matlab and FSL

Student: (Project available)

Robust diffusion tensor MR imaging for clinical trials of novel therapeutics

Supervisors: Gary Hui Zhang and (not allocated yet)

- Theme: Physics; Computing
- Description: Diffusion tensor MR imaging (DTI) is an advanced MRI technique with important applications in clinical trials of novel therapeutics for brain disorders, such as dementia. The technique provides unique probe of tissue microstructure non-invasively, making it an indispensable tool for detecting early pathological changes as well as for monitoring therapeutic effects. One key challenge facing DTI is its relatively long acquisition time, which makes it prone to data corruption due to intermittent subject motion. The inclusion of such data may compromise the accuracy of the resulting tissue microstructure estimates. However, its effect has not yet been examined in the context of therapeutic trials. This project will investigate this problem, in particular in the context of ongoing studies of neurodegenerative diseases at UCL. The student will use Camino, a well-documented software suite for analyzing DTI data, to study a variety of strategies to robustly process existing clinical data set. This study will identify the most robust strategy for use in clinical trials.
- Requirements: Basic computing and math skills
- Skills to be acquired: In-depth understanding of MRI and common neurodegenerative diseases; MR data visualization and analysis with matlab and Camino; statistical analysis

Student: (Project available)

SmartNail calibration: implant bending compensation

Supervisors: Stephen Taylor and (not allocated yet)

This project builds upon existing successful work to build a computer controlled calibration rig for applying known loads to an instrumented implant (the nail), and measuring the implant response in order to calibrate it. A LabView (v9) graphical user interface (GUI) is used to control a semi-automatic rig which calibrates the nail which houses force measuring sensors inside it. The GUI controls stepper motors to apply loads to the nail, and the developed strains are recorded on the GUI. The applied force is modified in response to the readings of a load cell which continuously monitors the applied loads. A feedback loop operates such that the input loads are applied; a succession of forces and torques are thus applied in sequence and the nail strains measured. The strains are related to the applied loads with a sensitivity matrix determined by multiple linear regression. The mechanics are already in place, and the LabView program currently operates to achieve the above.

The remaining task (this project) is to connect linear potentiometers to the nail such that the locus of bending of the nail is captured at each load step, in order to align the applied loads measured by an in-line load cell to the nail, to compensate for the bending. This will be achieved by logging the potentiometer length changes, triangulating to find the XYZ centre corresponding to the local nail centre, and correcting the load cell measurements to the local nail point. This will enable the sensitivity matrix to be corrected for the bend of the nail. In order to streamline the procedure for a succession of nails to be implanted in man, the project will also involve automating the procedure as far as possible; this could be in MATLAB, LabView, or Excel. In vivo implants will be built early next year, and should be available for calibration during the latter months of the project.

A good working knowledge of LabView is desirable, although almost all LV programming is done. The main task will be data analysis and load vector correction in software. It will also be necessary to calibrate two in-line 6 degree-of-freedom load cells initially. Some knowledge of electronic instrumentation, and a knowledge of vector algebra would be an advantage.

Student: (Project available)

Control logic for a stimulator IC

Supervisors: Dr. Anne Vanhoest and Prof. Nick Donaldson

This project is for one or two students with prior knowledge of electronic logic (and micro-controllers) and basic programming skills. There is currently a discrete control circuit for one of the IDG's stimulator IC that relies on a purpose-built FPGA. The students are to replace this circuit with a simpler and more user friendly logic controller, probably involving a micro-controller. Hence the pre-requisites of knowing some basic electronics, enough to understand the current circuit and design the new controller. It is open to either one student or a pair with good collaboration skills. For more information on the research carried on in the Implanted Devices Group

Student: (Project available)

An interactive e-learning tool for electrical stimulation theory

Supervisors: Dr. Anne Vanhoest and Mr Nathaniel Dahan

In a first stage the student will undertake a literature review to understand the basics of the theory of electrical stimulation, including notions of rheobase and chronaxie, with a focus on stimulation optimisation from a point of view of charges and energy. This is of importance for future development of electrical stimulators as minimising the energy per stimuli would improve the power requirements of the stimulator while the charge delivered may be linked to potential nerve damage if it is excessive. The theoretical findings shall be integrated in an interactive e-learning tool showing several stimulus waveforms with adjustable parameters and calculating the charge and energy requirements. This project does not involve any lab work and is therefore suitable for a student living further away. No prior knowledge of electrical stimulation principles is required. The student must be comfortable with integrations and derivations as the first phase of the project is theoretical. Some previous experience of designing web-interfaces, with user interactions and animations, may also be required. For more information on the research carried on in the Implanted Devices Group

Student: (Project available)

Investigating chromatic adaptation to monochromatic light in an immersive viewing sphere

Supervisors: Lindsay MacDonald and Dr Adam Gibson

Chromatic adaptation is the change in visual colour perception caused by a change in the prevailing scene illumination. It enables us to see an object as having the ‘same’ colour under variable lighting conditions, a phenomenon known as colour constancy. Under photopic (daylight) conditions the adaptation is generally believed to occur rapidly, within a few tens of seconds, whereas the transition to scotopic (moonlight) vision takes at least 20 minutes, as the retinal rod photoreceptors become active.
We are using a unique piece of new experimental apparatus: a 750 mm fibreglass sphere painted grey on the inside, into which near-monochromatic light is projected. By using a series of 20 nm bandpass filters from 400 to 700 nm, sixteen adapting wavelengths can be obtained. The experimental task for the observer is to adjust slider controls to achieve a visually neutral (i.e. ‘perceived grey’) field on a display screen viewed through a 4-degree circular aperture. The task is repeated for a series of adapting wavelengths and brightness levels.
Initial results suggest that the period of time required for complete chromatic adaptation is much longer than generally supposed, and indeed may be up to an hour. Also it seems that the observer’s previous viewing environment, for example in a lab with subdued lighting or in an office with bright fluorescent lighting or outside in daylight, has a strong influence on the course of adaptation until it finally settles in equilibrium. This has profound implications for all colour vision experiments, because the experimental data gathered in psychophysical experiments may depend more on where the observer was previously than on the ambient conditions within the experimental room.
The objective of this project is to gather more experimental data with a number of subjects, both male and female, to visualise the results in 3D graphical presentation, and to develop a model of the chromatic adaptation process as a bivariate function of wavelength and time. Some engineering skills will be needed for the configuration, measurement and calibration of lamps and displays. Familiarity with the human visual system and with Matlab programming and data analysis techniques would be an advantage.

Student: (Project available)

An fNIRS 4D sensitivity atlas

Supervisors: Lorenzo Fabrizi and Madeleine Verriotis and Rob Cooper

Functional Near Infrared Spectroscopy (fNIRS) is a non-invasive monitoring technique to measure the concentration of oxygenated and deoxygenated haemoglobin in the blood. This is often used to detect brain activation and metabolism at the bedside. One possible application of fNRIS is to monitor the oxygen consumption in the developing brain of premature and term infants. However, placement of the recording channels (the optodes) depends on the brain regions of interest. In neonates, this is complicated by the fact that brain anatomy changes rapidly across the first weeks of life.
This is a simulation-based project that aims at generating a 4D sensitivity atlas to inform optodes placement in neonates. In order to achieve this, the student will have to produce Finite Element Models of the infant brain at different gestational ages using existing MRI structural images; from these, it will be possible to choose the most appropriate positions to place the optodes in order to monitor the development of a pre-selected region of interest.

Student: (Project available)

Positron Emission Mammography for Breast Cancer

Supervisors: John Hipwell and Thomy Mertzanidou
Clinical Supervisor: Dr Mo Keshtgar (Consultant Breast Surgeon)
Nuclear Medicine Supervisor: Danny McCool (Head of the Nuclear Medicine Physics)

Breast cancer is the most common cancer in women in the UK with an incidence rate that was 20% higher than the most frequently occurring cancer in men (prostate) between 2007 and 2009. Although breast cancer is increasingly treatable it was the leading cause of death (3,665 or 12%) for females between the ages of 35 and 64 years in 2009 in England and Wales. The Department of Surgery at the Royal Free Hospital has recently acquired a new "Mammi" Breast Positron Emission Tomography (PET) scanner (http://www.gem-imaging.com/productos/mammi/mammi.php?idm=en) and is keen to evaluate its potential for breast cancer, in particular for staging and estimating response to chemotherapy. Dedicated PET imaging specifically for mammography is a relatively new application which offers considerably better spatial resolution and photon-detection sensitivity than conventional whole body PET scanners, however its clinical role in the context of other competing (or complementary) breast imaging modalities, such as Dynamic Contrast Enhanced MRI, has yet to be established. This project will investigate a number of aspects of this technology including (according to the student's interest and expertise) estimating the spatial resolution of the scanner (in the z-axis with bed position), registration and comparison to MRI and PET-CT, reconstruction issues, adaptive filtering to reduce dose, detection of DCIS and issues of introducing this technology into the clinical workflow.

Student: (Project available)