Stimulating language recovery after stroke: Tailored non-invasive electrical stimulation of the domain-general frontoparietal network.

Context
There are 1.2 million stroke survivors in the UK with a societal cost of £26 billion a year. Language impairment (aphasia) is a major cause of disability after stroke affecting a third of all patients. Current therapeutic strategies for treating aphasia have either had little success or demand prolonged and intensive speech and language therapy which is not feasible within the NHS. Most of these therapeutic strategies have largely focused on improving the residual language-specific processes or brain regions that support them. However, I have shown that residual language function after stroke is also dependent on support from the so-called domain-general brain regions that mediate general cognitive processes such as attention, working memory, and learning (or reacquisition) of a skill. Importantly, increased activity within the domain-general regions of a well-established bilateral frontoparietal brain network (FPN) has been shown to be associated with better recovery of language function after stroke, building a case for its potential as a prognostic biomarker and therapeutic target. The overall aim of this proposal is to investigate and to optimise the compensatory role of this domain-general brain network in recovery of language function after stroke.

Objectives
1. To use a rapid real-time brain imaging protocol to measure the function of the domain-general FPN across a wide range of language and non-language tasks in patients after stroke. Based on preliminary results, it is expected that patients who have a similar FPN function compared to the controls, will have the best language outcomes following their stroke. Conversely, those patients with a dissimilar FPN functional profile to the controls will have the worse outcomes.
2. To combine real-time brain imaging and non-invasive brain stimulation to rapidly identify personalised brain stimulation settings that maximise the function of the domain-general FPN for each individual patient.
3. To investigate whether the use of personalised brain stimulation of the domain-general FPN improves behavioural performances in patients after stroke.

Method
I will use advanced magnetic resonance imaging (MRI) to study brain structure and function. Importantly I will use a novel technique that analyses brain function in real-time whilst participants perform tasks in the scanner. This will allow for a rapid assessment of the function of domain-general brain regions of the FPN across a wide range of experimental conditions. I have used this technique in a pilot study of patients to show that the pattern of activity of the FPN across a range of language and non-language tasks, is related to the degree of language impairment after stroke. I will use the same technique on a group of 50 patients following a stroke, and relate their imaging measures to their language function after the stroke. In a second study, I will use the same real-time brain imaging technique combined with non-invasive brain stimulation, to identify for each individual patient the optimal brain stimulation settings that maximally activate their domain-general FPN. Here, the optimal phase and frequency of the brain stimulation will be identified for each individual patient. In the third study, I will apply brain stimulation using the personalised stimulation settings for each patient, to improve behavioural performances.

Potential applications and benefits
Language impairment after stroke is common, and it is unclear how best to rehabilitate individuals. Identifying individuals who will benefit from rehabilitation of domain-general brain regions that support their residual language function will provide a new therapeutic avenue. Demonstration of modulation of these regions using a personalised non-invasive brain simulation will not only benefit patients with language impairment, but will have far reaching implications for rehabilitation across a wide range of brain deficits and injuries.

Stroke-related aphasia is a major cause of disability with a UK incidence of ~30,000 per annum. A sound understanding of the mechanisms that underpin recovery of language function is an unmet need. Most therapeutic strategies for aphasia have largely focused on language-specific processes and brain systems with limited success. I have previously shown that domain-general brain networks are also critical in recovery of language function after stroke, building the case for investigating their potential as a prognostic biomarker and therapeutic target. The overarching aim of this proposal is to investigate the compensatory role of the well-established bilateral frontoparietal domain-general network (FPN) in recovery after aphasic stroke. Using real-time functional MRI combined with machine learning in patients after stroke, I will rapidly assess the activity of the FPN across a wide range of language and non-language tasks, with the aim of identifying a novel prognostic biomarker for longitudinal language recovery. I have shown in a pilot study that this is feasible in patients with stroke. In a second study, I will use a second real-time machine learning MRI combined with transcranial Alternating Current Stimulation (tACS) to rapidly tailor tACS parameters (phase and frequency) to maximise the function of the FPN in each individual patient. In the third behavioural study, the patients will undergo tACS stimulation using their individualised stimulation parameters identified in the previous step to investigate whether tACS induced increases in activity of FPN will improve behavioural responses during a language (speech production) and non-language (calculation) task. By using optimal stimulation parameters, this proposal will definitively address the question of modulating brain function using tACS. This approach will have far-reaching implications for cognitive neuroscience and rehabilitation in multiple clinical domains beyond aphasic stroke.

Most therapeutic strategies for language impairment (aphasia) after stroke have focused on residual language-specific processes and brain regions supporting them. This is despite an understanding that residual language function after stroke is also dependent on domain-general brain regions that normally mediate attentional and general cognitive control functions. This work will: 1) use neuroimaging to rapidly measure the function of domain-general brain regions in patients after stroke, and to explore whether knowledge of their function improves our ability to predict the degree of language recovery after stroke; 2) combine neuroimaging and non-invasive brain stimulation to individually optimise the function of domain-general brain regions for each patient; and 3) investigate whether the use of individually tailored stimulation of domain-general brain regions results in improved performances in patients. This would have impact across a number of areas:

Healthcare professionals (clinicians and speech and language therapists).
Despite the high prevalence of stroke survivors with persisting aphasia, there remains no consensus about either how to deliver language rehabilitation or its efficacy. This work will help with stratifying patients on the basis of imaging markers, to identify those likely to show the best recovery potential. The results will add to the body of knowledge that helps selection of patients into future clinical trials. The proposed research will also demonstrate why only a proportion of patients respond to therapeutic strategies. Further, identifying those patients most likely to benefit - from interventions targeting both domain-specific and domain-general processes - may help formulate long-term treatment programmes and a more rational approach than a policy of 'a few hours for most', which is the current norm within the NHS.

Researchers interested in modifying brain function with non-invasive brain stimulation.
There is considerable interest in the use of non-invasive brain stimulation techniques to modify brain function, with the technique having been used in ~2000 publications per annum over the last three years. The precise stimulation technique and parameters to use in each experiment and each individual remain open questions. The second study in this proposal will be very interesting to researchers using this technique across the wide spectrum of cognitive neuroscience as it allows identification of an individually tailored stimulation parameter to optimise a desired brain state.

Commercial-private sector.
Private sector companies supplying non-invasive brain stimulation will be interested in the outcome of this research. One such company is Rogue resolution. The company holds regular workshops and an annual conference (BrainBox initiative) which I have attended over the last two years. I intend to present my findings at their conference which will be attended by academics and commercial companies interested in non-invasive brain stimulation.

Charities.
The outcome of the research will be of interest to related Charites such as the Stroke Association and Dyscover.

Public.
Stroke is the leading cause of disability in the UK and the fourth cause of death. A third of all stroke result in language impairment. Therefore, results of this research are likely to be relevant to a large sector of UK population who will either be directly affected by stroke over the course of their life time, or will know someone close who is affected by it.

Economic and societal impacts.
The proposed work has the potential to make a large societal impact. There are over 1.2 million stroke survivors in the UK with a societal cost of around £26 billion a year. Therefore, even a small improvement in identifying biomarkers that could eventually help with better designed therapeutic strategies could have dramatic cost benefits.