A brain-machine interface for real-time functional network analysis

In our laboratory we are looking for biomarkers relevant to the early diagnosis of Alzheimer’s disease, and as part of this effort we are investigating factors that alter functional connectivity in the aging brain. We have developed a new, multivariate implementation of Granger causality analysis with an analytical solution that allows rapid estimation and visualisation of network structure within EEG data (e.g. Gao et al., 2015, Scientific Reports, 5, 10399). So far, this technique has been used ‘offline’ to estimate networks from EEG data after it has been collected. But now we are developing this technique to work in real-time. To help us achieve this goal, we have a PhD project available to generate a prototype brain-machine interface (BCI) that will compute and display network structures within hi-density, 64-channel EEG data in real-time while it is being collected.

The project will focus on functional networks involved in the encoding and retrieval of information from long-term memory (e.g. Allan and Allen, 2005, J Neuroscience; Tsivilis et al., 2015, Frontiers in Human Neuroscience), and will proceed in three stages. During the first stage, the student will have two lines of work. Our existing neuroimaging (EEG and MRI) data sets, which also include various cognitive, dietary and peripheral biomarker data, will be analysed to determine network parameters relevant to memory function that the BCI should target. In parallel, the student will work on the software implementation of the BCI that will convert time series data into functional networks for visualisation in real-time. During the second stage of the project, the operation of the prototype BCI will be tested via simulations using existing EEG data sets. During the third stage of the project, the operation of the prototype BCI will be tested during actual hi-density EEG recordings.

This project is cross disciplinary in nature, and is therefore most suitable for individuals with degrees either in psychology / neuroscience, or biomedical engineering and signal analysis, but regardless of the degree background a strong interest in statistics, brain-machine interfaces and machine learning is essential. The project will be supervised by Dr Allan, a cognitive neuroscientist based in the School of Psychology, and Professor Schelter, a physicist based in the Institute for Complex Systems and Mathematical Biology.

Candidates must hold (or be expected to obtain) an Upper Second Class honours degree in Psychology or cognate discipline as a minimum requirement. Additional research experience would be an advantage. Applications must be made through the University of Aberdeen applicant portal: http://www.abdn.ac.uk/study/postgraduate/apply.php. Applications must include: 1) An on-line form completed through the applicant portal. 2) A project proposal, maximum length 1 x A4 page including background, aims and suggestions for 2-3 studies to address the aims. A reference list can be added on a second page if you wish. Applicants can liaise with the supervisor when developing their proposal. 3) Two academic references – please attach the references to the application or include full referee contact details. 4) A CV outlining your academic qualifications and research experience to date. If possible, it would be useful to add an academic transcript.