- Automatic Seizure Detection in Clinical Electroencephalography (EEG) Records: The use of algorithms to assist diagnostic imaging has received considerable research interest due to the opportunities to improve accuracy and reduce costs with the advances in large data centres, cloud computing resources, and machine learning. I am particularly interested in the use of Bayesian optimisation to optimise and compare machine learning pipeline/model configurations and hyperparameters for the detection of seizures in real-world patient EEG records. This can provide practical guidance to inform future machine learning pipeline development which could be implemented into healthcare practice.
- Development of Portable Bio-sensing Platforms: As information and communication technologies improve, the prominence of research and delivery of portable hardware and software solutions for ambulatory patient monitoring has increased. The proliferation of portable health devices means there are increasing specialities in the medical applications for mobile health devices; with one such application being the registration of epileptic seizures. These aim to meet the need expressed by neurologists of more detailed patient movement information during seizures and more accurate seizure counts over time. This can be achived through movement (e.g. Electromyography), physiology (e.g. Electrocardiology), or cerebral (e.g. Electroencephalography) measurement. I'm particularly interested in multi-modal systems which could be used to select measurements tailored to an individuals particular epilepsy presentation for diagnosis and treatment evaluation.
Recent Conference and Workshop Involvement
- North West Epilepsy Interest Group (Invited Speaker)
- PyData Lancaster Meeting (Speaker)
- Lancaster University Data Science Group (Speaker)
- MobiESense: Co-design workshop for clinical research studies using bio-potentials in ‘in-the-wild’ settings (Co-Organiser)
- Lancashire Health Research Showcase (Speaker)
Having been awarded a 1st Class BSc in Psychology, I gained an ESRC (NWDTC) studentship, via the Advanced Quantitative Methods pathway, to continue in my MSc and PhD studies at Lancaster University. My early research focused on the neural activations of infants perceiving actions (BSc thesis) or social stimuli (MSc thesis). During my PhD, I began teaching statistical hypothesis testing and research design to undergraduate and master’s students and was appointed to the Faculty of Science and Technology Research Ethics Committee. My PhD research focused on developing hardware, software, and quantitative methods for portable neurological monitoring of childhood patients with epilepsy. I created project teams which encompassed a range of disciplines and organisations (Statistics, Computing, Engineering, Medicine, Psychology), and gained a number of successful grants from organisations such as Google, MRC, NIHR, and the EPSRC. This led to co-supervising undergraduate and PhD student research projects to develop the software and hardware for a portable electroencephalogram (EEG) device. After piloting our system in two NHS organisations, I focused on developing machine learning algorithms to address the current bottleneck on diagnosis speed and accuracy resulting from a fully manual marking procedure for patient EEG records. To complement my research in this field I am also developing a series of online Python tutorials for developing machine learning pipelines to detect epileptic seizures in brain data (https://github.com/Eldave93/Seizure-Detection-Tutorials).
In 2020 I joined the University of Edinburgh as a University Teacher in Mathematical Sciences Computing. Alike to many data scientists from a neuroscience background, I appreciate how human neurology and behaviour has inspired many modern machine learning advancements. I am passionate about the practical applications of machine learning and aim to educate and inspire others to develop research and models to tackle real-world challenges.
Semester 2 (20/21):
- MATH11008: Probability and Statistics (Course Organizer)
- MATH11205: Machine Learning in Python (Lecturer/Teaching Team)
Semester 1 (20/21):
- MATH08077: Introduction to Data Science (Co-organiser/Teaching Team)
- MATH11176: Statistical Programming (Tutor)
- MATH11203: Introductory Probability and Statistics (Course Organiser)
University of Lancaster
Oct 2016 - Present (Minor Corrections): PhD Applied Statistics
Oct 2015 - Oct 2016: MSc Developmental Disorders (Distinction)
Oct 2012 - Jul 2015: BSc Psychology (1st Class)
Teaching Associate at Lancaster University (2017/19)
- PSYC401: Analyzing and Interpreting Data,
- PSYC214: Investigating Psychology (Statistics),
- PSYC207: Personality and Individual Differences,
- Completed a Higher Education Academy (HEA) accredited Associate Teacher Programme (ATP).
Applied Behavioral Analysis (ABA) Services (2014 – 2015)