Exploring the Potential of Multimodal and Multiphasic Brain-Computer Interfaces for Neurorehabilitation

Abstract

Stroke is a significant contributor to disability-adjusted life years in Europe, and its incidence is expected to rise due to demographic changes highlighting the need for effective post-stroke motor rehabilitation. Early intervention may be crucial, but many existing therapies require a minimum functional movement precluding their use early after stroke. Brain-Computer Interfaces (BCIs) are an attractive option for neurorehabilitation following a stroke due to their unique ability to be used even when the patient is experiencing motor paralysis.

In this thesis I report findings from investigations into the current state of the art in BCI methodology, with a view to addressing current challenges and improving future implementations of BCI for stroke. I describe how priming using Transcranial Magnetic Stimulation (TMS) - Neurofeedback (NF), can be used to accelerate and improve performance on standard Electroencephalography (EEG) BCIs, by providing real-time muscle specific feedback on how motor imagery excites or inhibits brain-muscle pathways. I introduce a two-phase multimodal BCI approach, where patients could use TMS-NF at the bedside in the early weeks following stroke to guide the development of optimal motor imagery strategies, followed by an extended rehabilitation phase practising guided motor imagery at home using a wireless, wearable EEG system. I also explore the potential BCI applications in Body Integrity Dysphoria and summarise my approach to medical BCI design. Additionally, I report results from investigating a research gap between white matter structural integrity in motor inhibition networks and falls in elderly individuals, using data from the Irish Longitudinal Study of Aging.

In conclusion this thesis offers a novel approach to BCI development aiming to offer a bridge from basic research to translational therapies in the field of medical BCIs.