DARTH group mission
Ever-increasing healthcare provision demands strain national health systems globally, which struggle to meet patients' needs. We passionately believe we can design and provide effective solutions which will revolutionize contemporary healthcare delivery through capitalizing technology and harnessing data.
We are a diverse research group based at the Usher Institute, University of Edinburgh working at the interface of engineering, mathematics, informatics, and medicine. Members in the DARTH group come from very different academic backgrounds, including engineering, mathematics, statistics, computer science, psychology, and medicine, and work closely together.
We develop and apply signal processing and statistical machine learning algorithms to explore data and decipher complicated concealed statistical relationships. Our algorithms are directly driven by and validated on complicated real-world problems, aiming to facilitate interpretation of the underlying key mechanisms of the modelled system. Our work is inherently multi-disciplinary, and we collaborate with industrial partners and researchers worldwide. We tackle challenging problems in healthcare domains from neurodegenerative disorders and mental health, to asthma, cardiovascular disease, and neonatal monitoring.
1 September 2019
Three new conference papers accepted for publication in IEEE BIBE
Holly, Minhong, and Thanasis will be presenting their research findings in the next IEEE BioInformatics and BioEngineering conference in Athens, Greece at the end of October 2019. Thanasis is co-organising a special session at the conference entitled Intelligent Digital Health Interventions towads Prevention, Self-Management, and Treatment of Pathologies.
16 August 2019
New journal article accepted for publication in Circulation
We developed a statistical machine learning algorithm which uses changes in cardiac troponin concentrations, controlling for age and gender, and demonstrated we can very accurately predict myocardial infarction. The findings have been validated on a very large cohort of approximately 8,000 people. The algorithm can be used to identify low-risk and high-risk patients and lead to better-informed clinical decisions.
1 August 2019
New conference papers accepted and keynote lecture - MAVEBA
Andres will be presenting his work on: "A neuromechanical model of jaw-tongue articulation in Parkinson's disease speech", and Thanasis will be presenting results from the Parkinson's Voice Initiative study. Thanasis will also be delivering this year's keynote lecture in MAVEBA, entitled: Developing new speech signal processing algorithms for biomedical and life sciences applications: principles, findings, challenges, and a view to the future.
4 June 2019
New journal article accepted for publication in BMJ Open
Asthma is a long-term condition with rapid onset worsening of symptoms (‘attacks’) which can be unpredictable and may prove fatal. Models predicting asthma attacks require high sensitivity to minimise mortality risk, and high specificity to avoid unnecessary prescribing of preventative medications that carry an associated risk of adverse events. This protocol paper describes the planned work to create a risk score to predict asthma attacks in primary care using a statistical learning approach trained on routinely collected electronic health record (EHR) data.
14 May 2019
New journal article accepted for publication in Scientific Reports
"Quantifying ultrasonic mouse vocalizations using acoustic analysis in a supervised statistical machine learning framework" Most mouse communication is produced in ultrasonic frequencies beyond human hearing. These ultrasonic vocalizations (USVs) are typically described by experts using nine call types. In this study, we replicated the expert-defined call types of communicative vocal behavior with 85% accuracy by using acoustic analysis to characterize USVs and a principled supervised learning setup. This has important implications in understanding animal behaviors.
11 April 2019
New conference article accepted for publication in EMBC
"New insights into Parkinson’s disease through statistical analysis of standard clinical scales quantifying symptom severity" In clinical practice, neurologists do not have time to use the full span of clinical scales and anecdotally rely on their experience to use a subset of the available items in fully established clinical metrics assessing Parkinson's disease symptom severity. This study aimed to investigate the practical implications of standard clinical assessment tools in monitoring Parkinson's disease symptom severity.
10 April 2019
New journal article accepted for publication in JASA
"Developing a large scale population screening tool for the assessment of Parkinson’s disease using telephone-quality speech". We collected the largest speech database in the world through the Parkinson's Voice Initiative and aimed to demonstrate that speech recordings collected under highly uncontrolled conditions may lead to acceptable accuracy in differentiating healthy controls from people diagnosed with Parkinson's disease simply processing sustained vowels collected over the standard telephone network.
22 March 2019
New article published in Urology News, highlighted in the frontpage of the magazine
"Rise of the machines: will artificial intelligence replace the urologist?" In this article we review the state of the art in the field of Artificial Intelligence (AI) focusing in the field of urology with a view to contemporary developments and future directions. We argued that data scientists and urologists need to work in close collaboration to harness opportunities: "rather than replace urologists, AI will mainly be used to inform, enhance, and complement their practice."
17 February 2019
New PhD student joining the DARTH group
Congratulations to Andres Gomez-Rodellar for successfully interviewing at the highly competitive DTP in Precision Medicine programme and being offered a position to join our research group. Andres will be starting with us in September 2019.
20 January 2019
New journal article accepted for publication in JMIR
Machine learning has attracted considerable research interest towards developing smart digital health interventions. These interventions have the potential to revolutionize healthcare and lead to substantial outcomes for patients and medical professionals. Unfortunately, research findings are rarely translated in clinical interventions. This study provides a literature review to contextualize the current state of the art in this setting and highlight contemporary limitations.