The fast without the spurious: developing a system for robust and rapid simultaneous EEG-fMRI measurements
Lead Research Organisation:
University College London
Department Name: Institute of Child Health
Abstract
To increase our understanding of how the brain works and how it goes wrong in people with neurological conditions such as epilepsy we need to develop better systems for making measurements of brain activity.
Electroencephalography (EEG) is an important modality in clinical and experimental neuroscience capable of measuring electrical changes occurring with sub-millisecond temporal resolution and a spatial resolution of centimetres while functional Magnetic Resonance Imaging (fMRI) can map haemodynamic changes over the entire the brain at a time-scale of seconds and spatial resolution of a few millimetres or better. Therefore together they can perform measurements across a greater range of brain activity occurring either at faster temporal or smaller spatial scales. However, during simultaneous EEG-fMRI acquisitions these both methods signal are degraded by noise related to motion, thereby significantly limiting the sensitivity of this type of study so far.
The purpose of this application is to build a robust system that can perform these simultaneous EEG and fMRI measurements of brain activity. To achieve this goal we will integrate new fast fMRI pulse sequences because images obtained in shorter time intervals are intrinsically less motion sensitive (like a faster shutter speed on a camera). Also better modelling of motion will be possible if we obtain more images per unit time because we can better separate and model the different sources of signal and noise that occur in different frequency ranges. In addition, we will optimise motion detection and prospective motion correction (PMC) using a camera system that tracks the subject's motion and updates the image acquisition process so that patient motion is supressed. When using these improvements to fMRI data acquisition we will need to develop novel EEG artefact correction methods for simultaneous in-scanner EEG recording which currently rely on the repetitive nature of the artefact in time. Both motion and PMC are likely to make the artefact more variable and so will require the development of novel correction methods.
Once we have developed this system we will apply it to ten patients with hard to treat epilepsy from Great Ormond Street Hospital who are being assessed for epilepsy surgery. This assessment aims to identify the epileptic brain regions and EEG-fMRI is a tool to help obtain this information. We will compare our current standard EEG-fMRI protocol which often suffers from degradation due to motion (particularly in young children) to our new robust EEG-fMRI incorporating PMC, fast fMRI and improved EEG artefact correction.
Electroencephalography (EEG) is an important modality in clinical and experimental neuroscience capable of measuring electrical changes occurring with sub-millisecond temporal resolution and a spatial resolution of centimetres while functional Magnetic Resonance Imaging (fMRI) can map haemodynamic changes over the entire the brain at a time-scale of seconds and spatial resolution of a few millimetres or better. Therefore together they can perform measurements across a greater range of brain activity occurring either at faster temporal or smaller spatial scales. However, during simultaneous EEG-fMRI acquisitions these both methods signal are degraded by noise related to motion, thereby significantly limiting the sensitivity of this type of study so far.
The purpose of this application is to build a robust system that can perform these simultaneous EEG and fMRI measurements of brain activity. To achieve this goal we will integrate new fast fMRI pulse sequences because images obtained in shorter time intervals are intrinsically less motion sensitive (like a faster shutter speed on a camera). Also better modelling of motion will be possible if we obtain more images per unit time because we can better separate and model the different sources of signal and noise that occur in different frequency ranges. In addition, we will optimise motion detection and prospective motion correction (PMC) using a camera system that tracks the subject's motion and updates the image acquisition process so that patient motion is supressed. When using these improvements to fMRI data acquisition we will need to develop novel EEG artefact correction methods for simultaneous in-scanner EEG recording which currently rely on the repetitive nature of the artefact in time. Both motion and PMC are likely to make the artefact more variable and so will require the development of novel correction methods.
Once we have developed this system we will apply it to ten patients with hard to treat epilepsy from Great Ormond Street Hospital who are being assessed for epilepsy surgery. This assessment aims to identify the epileptic brain regions and EEG-fMRI is a tool to help obtain this information. We will compare our current standard EEG-fMRI protocol which often suffers from degradation due to motion (particularly in young children) to our new robust EEG-fMRI incorporating PMC, fast fMRI and improved EEG artefact correction.
Planned Impact
EEG-fMRI is used for the pre-surgical evaluation of focal epilepsy. The aim of this grant is to achieve motion insensitive, low noise, high temporal resolution, and simultaneous EEG-fMRI data acquisition. This will enable in the timescale of the grant:
- The improvement of EEG-fMRI as a clinical tool in terms of sensitivity specificity. This means better information for clinicians and parents to make life changing decisions.
- The study of seizures which represents an important clinical goal.
- Improvement of fMRI as a clinical tool by a reduction in movement sensitivity.
- Locally we aim to implementation EEG-fMRI as a clinical service at GOSH (which has the largest paediatric epilepsy surgery program in the UK) and this service will benefit from these developments.
- A reduction in the need for sedation and the amount of lost data as greater movement will be tolerated with acceptable image quality with significant cost implications.
Two strands of research are likely to coincide in the next 10 years. Firstly there is a great deal of research towards using different kinds of treatments for epilepsy that rely on local and acute delivery of therapy rather than the current global and chronic approaches available now (oral drug therapies and surgery). At the same time the understanding of epilepsy in terms as a disease involving a network of brain regions has increased in part due to imaging brain function. We therefore need to characterise and understand brain networks and their transition to epileptic states. The work in this grant aims to develop a better tool for this wider goal.
Recently there has been a large interest in so called 'resting state' fMRI where signals are recorded without a task and brain networks investigated by extracting regions that are temporally correlated and this technique has been applied to many patient groups such as dementia and has significant advantages in terms of simplicity and it can be applied across ages and intellectual abilities. However, there are a number of areas of uncertainty:
1. There is a poor definition of a 'resting state' and its consistency between centres.
2. There is an increased recognition that there are dynamic changes in 'resting state' activity which can measured can be seen at different temporal scales.
3. Physiological noise and motion cause uncertainty in the results which rely purely on correlations and so are sensitive to these noise sources.
The proposed developments in this grant aim to make a system where brain state can be characterised using EEG and so between subject differences can be accurately modelled. The combination of fast fMRI and EEG allows for the analysis of dynamic changes in brain networks over a range of temporal scales because of increased temporal sampling. The use of PMC and fast fMRI minimises the effects of motion and physiological noise therefore robustly revealing correlations due to brain activity. This EPSRC first grant can contribute to resting state fMRI connectivity with implications across a broad range of diseases for diagnosis and treatment monitoring and testing.
These developments might also have some commercial value for the manufacturers of MRI compatible EEG equipment and MRI scanners. The implementation of new EEG correction strategies maybe commercialised such as in brain products analyser software and / or as a free toolbox (e.g. as an additional tool interfaced with SPM www.fil.ion.ucl.ac.uk/spm). The development of fMRI protocols and their application within clinical populations can be of commercial interest (e.g. to Siemens medical).
- The improvement of EEG-fMRI as a clinical tool in terms of sensitivity specificity. This means better information for clinicians and parents to make life changing decisions.
- The study of seizures which represents an important clinical goal.
- Improvement of fMRI as a clinical tool by a reduction in movement sensitivity.
- Locally we aim to implementation EEG-fMRI as a clinical service at GOSH (which has the largest paediatric epilepsy surgery program in the UK) and this service will benefit from these developments.
- A reduction in the need for sedation and the amount of lost data as greater movement will be tolerated with acceptable image quality with significant cost implications.
Two strands of research are likely to coincide in the next 10 years. Firstly there is a great deal of research towards using different kinds of treatments for epilepsy that rely on local and acute delivery of therapy rather than the current global and chronic approaches available now (oral drug therapies and surgery). At the same time the understanding of epilepsy in terms as a disease involving a network of brain regions has increased in part due to imaging brain function. We therefore need to characterise and understand brain networks and their transition to epileptic states. The work in this grant aims to develop a better tool for this wider goal.
Recently there has been a large interest in so called 'resting state' fMRI where signals are recorded without a task and brain networks investigated by extracting regions that are temporally correlated and this technique has been applied to many patient groups such as dementia and has significant advantages in terms of simplicity and it can be applied across ages and intellectual abilities. However, there are a number of areas of uncertainty:
1. There is a poor definition of a 'resting state' and its consistency between centres.
2. There is an increased recognition that there are dynamic changes in 'resting state' activity which can measured can be seen at different temporal scales.
3. Physiological noise and motion cause uncertainty in the results which rely purely on correlations and so are sensitive to these noise sources.
The proposed developments in this grant aim to make a system where brain state can be characterised using EEG and so between subject differences can be accurately modelled. The combination of fast fMRI and EEG allows for the analysis of dynamic changes in brain networks over a range of temporal scales because of increased temporal sampling. The use of PMC and fast fMRI minimises the effects of motion and physiological noise therefore robustly revealing correlations due to brain activity. This EPSRC first grant can contribute to resting state fMRI connectivity with implications across a broad range of diseases for diagnosis and treatment monitoring and testing.
These developments might also have some commercial value for the manufacturers of MRI compatible EEG equipment and MRI scanners. The implementation of new EEG correction strategies maybe commercialised such as in brain products analyser software and / or as a free toolbox (e.g. as an additional tool interfaced with SPM www.fil.ion.ucl.ac.uk/spm). The development of fMRI protocols and their application within clinical populations can be of commercial interest (e.g. to Siemens medical).
Organisations
- University College London (Lead Research Organisation)
- Maastricht University (UM) (Collaboration)
- Universidade de São Paulo (Collaboration)
- Max Planck Society (Collaboration)
- University College London (Collaboration)
- Swiss Federal Institute of Technology in Lausanne (EPFL) (Collaboration)
- Brain Products (Collaboration)
- Massachusetts General Hospital (Collaboration)
- IMPERIAL COLLEGE LONDON (Collaboration)
- Brain Products GmbH (Project Partner)
People |
ORCID iD |
David Carmichael (Principal Investigator) |
Publications
Centeno M
(2017)
Combined electroencephalography-functional magnetic resonance imaging and electrical source imaging improves localization of pediatric focal epilepsy.
in Annals of neurology
Centeno M
(2016)
Optimising EEG-fMRI for Localisation of Focal Epilepsy in Children.
in PloS one
Maziero D
(2021)
Unified Retrospective EEG Motion Educated Artefact Suppression for EEG-fMRI to Suppress Magnetic Field Gradient Artefacts During Motion.
in Brain topography
McDowell AR
(2019)
Optimal repetition time reduction for single subject event-related functional magnetic resonance imaging.
in Magnetic resonance in medicine
Perani S
(2018)
Thalamic volume reduction in drug-naive patients with new-onset genetic generalized epilepsy.
in Epilepsia
Steinbrenner M
(2023)
Camera-based Prospective Motion Correction in Paediatric Epilepsy Patients Enables EEG-fMRI Localization Even in High-motion States.
in Brain topography
Tangwiriyasakul C
(2018)
Dynamic brain network states in human generalized spike-wave discharges.
in Brain : a journal of neurology
Description | We have found that we can use prospective motion tracking for EEG-fMRI to both limit the impact of motion on the fMRI but also to correct for motion artefacts in the simultaneously acquired EEG data. We found that we could use a model of the motion induced EEG artefacts from the motion tracking system to very effectively suppress motion artefacts. In addition we have proven that using prospective motion correction for fMRI does not limit the EEG data quality and may actually improve it. This work has been published. We have then been able to use the motion information to improve the removal of MRI signal artefact when subjects are moving, this work is being written up for publication. These methods have been piloted in a small patient group the results are being written up for publication. We have tested fast fMRI sequences for clinical use (in the individual). We have obtained results that have been published that show a moderate increase in fMRI speed is beneficial but using ROC analysis that faster sequences reduce detection efficiency. We also show the appropriate AR models required to perform this analysis. We have obtained data using motion correction in a pilot group of 10 children with focal epilepsy and recorded epileptic activity that has been identified and preliminary analyses performed. We have developed an integrated EEG data cleaning methodology and will evaluate the quality of the results obtained to date. |
Exploitation Route | As part of this project we are beginning to use this technology to obtain EEG-fMRI data in children with epilepsy where movement is a problem. We have shown that we can obtain far higher data quality for the EEG and fMRI using a camera tracking / prospective motion correction system. We have had initial discussions with commercial partners about our work. |
Sectors | Healthcare |
Description | We have now obtained data using the robust motion correction EEG-fMRI data in n=10 children with focal epilepsy. The results obtained have been discussed in several of these individuals within the context of their clinical management. We are evaluating the quality of the data obtained and its potential impact for recovering clinically useful data. |
First Year Of Impact | 2019 |
Sector | Healthcare |
Description | SLMS UCL Faculty equipment grant |
Amount | £120,000 (GBP) |
Organisation | University College London |
Sector | Academic/University |
Country | United Kingdom |
Start | 06/2011 |
End | 08/2011 |
Description | Science without borders |
Amount | £43,000 (GBP) |
Organisation | National Council for Scientific and Technological Development (CNPq) |
Sector | Public |
Country | Brazil |
Start | 04/2014 |
End | 05/2015 |
Description | TBC |
Amount | £85,000 (GBP) |
Funding ID | 2444300 |
Organisation | Engineering and Physical Sciences Research Council (EPSRC) |
Sector | Public |
Country | United Kingdom |
Start | 09/2020 |
End | 03/2024 |
Description | Brain products |
Organisation | Brain Products |
Country | Germany |
Sector | Private |
PI Contribution | We are developing motion correction software and hardware to reduce motion sensitivity of EEG recorded with an MRI scanner |
Collaborator Contribution | Providing software solutions and hardware interfacing for EEG recording and correction. |
Impact | EPSRC Grant |
Start Year | 2014 |
Description | Hammersmith |
Organisation | Imperial College London |
Department | Computational, Cognitive and Clinical Neuroimaging Laboratory (C3NL) |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | We have provided methodological expertise, in particular for the set up and analysis of EEG-fMRI data. |
Collaborator Contribution | We have been given methodological input regarding the analysis of resting state fMRI data in particular using ICA. We have been provided support to develop and test stimulation paradigms. |
Impact | Grant (Ines Violante) Paper (under consideration J Neuroscience) |
Start Year | 2012 |
Description | Herbst and Poser |
Organisation | Maastricht University (UM) |
Country | Netherlands |
Sector | Academic/University |
PI Contribution | We are installing and testing new fMRI sequences that can take data more quickly and make use of prospective motion tracking. |
Collaborator Contribution | OUr partners have developed fMRI sequences that can take data more quickly and make use of prospective motion tracking. |
Impact | Not as yet |
Start Year | 2015 |
Description | Lausanne |
Organisation | Swiss Federal Institute of Technology in Lausanne (EPFL) |
Country | Switzerland |
Sector | Public |
PI Contribution | We are providing expertise in EEG-fMRI experimental set-up and also in the use of prospective motion correction technology. Data exchange. |
Collaborator Contribution | Provision of MRI pulse sequences and image processing expertise. Data exchange. |
Impact | None yet |
Start Year | 2014 |
Description | Lemieux |
Organisation | University College London |
Department | Institute of Neurology |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | We have provided methodological expertise and advice. Provided equipment and training, |
Collaborator Contribution | Provided equipment and training. Provided expertise and advice. |
Impact | Scientific papers e.g. Pugnaghi M, Carmichael DW, et al Generalized Spike and Waves: Effect of Discharge Duration on Brain Networks as Revealed by BOLD fMRI Brain Topogr. DOI 10.1007/s10548-013-0311-0. C Pedreira, AE Vaudano, RC Thornton, UJ Chaudhary, S Vulliemoz, H Laufs, R Rodionov, DW Carmichael, SD Lhatoo, M Guye, R Quian Quiroga, L Lemieux. Classification of EEG abnormalities in partial epilepsy with simultaneous EEG-fMRI recordings, Neuroimage. Volume 99, 1 October 2014, Pages 461-476. |
Start Year | 2011 |
Description | Nigel Hunt |
Organisation | University College London |
Department | Eastman Dental Institute |
Country | United Kingdom |
Sector | Academic/University |
PI Contribution | We have developed a partnership to use advanced motion correction technology in MRI. |
Collaborator Contribution | Prof. Hunts team have made oral inserts to allow us to accurately track skull position which are very safe and comfortable and so suitable for use in our patient populations. |
Impact | Paper currently in revision (Maziero et al, NeuroImage) |
Start Year | 2013 |
Description | University of Hawaii, formally University of Sao Paulo |
Organisation | Universidade de São Paulo |
Country | Brazil |
Sector | Academic/University |
PI Contribution | Danilo Maziero is a PhD student who I sponsored to spend 1 year working in my lab. He worked with me on the EEG correction work packages in my EPSRC first grant. This work has led to a significant paper that is in the advanced stages of review he is currently finalising a second publication. We are continuing this collaboration while he is in the University of Hawaii. |
Collaborator Contribution | Danilo has been continuing to develop in MRI EEG motion correction methods while at the University of Hawaii. Danilo spent a year working at our centre having obtained a grant of £43k from science without borders grant funded by conselho nacional de desenvolvimento científico e tecnológico www.cnpq.br, Brazil. Both Danilo and his supervisors at Sao Paulo Epilepsy centre / University Tonicarlo R. Velasco and Carlos E.G. Salmon contributed to writing the manuscript. |
Impact | Scientific publication (NeuroImage 2016) and abstracts at conferences (e.g. ISMRM Hawaii 2017, Human Brain Mapping, Hawaii, 2015) |
Start Year | 2014 |
Description | Weiskopf |
Organisation | Max Planck Society |
Department | Max Planck Society Leipzig |
Country | Germany |
Sector | Academic/University |
PI Contribution | This collaboration involves the application of MRI imaging technology. |
Collaborator Contribution | Our partners have supplied MRI imaging technology that they have developed. |
Impact | We obtained joint funding for motion correction technology and secured grant funding for projects. |
Start Year | 2014 |
Description | Yacoub |
Organisation | Massachusetts General Hospital |
Department | Martinos Center for Biomedical Imaging Massachusetts |
Country | United States |
Sector | Hospitals |
PI Contribution | We are going to test the MRI imaging developments for use in clinical populations |
Collaborator Contribution | They have supplied MRI pulse sequences |
Impact | Not yet |
Start Year | 2015 |
Title | REEGMAS |
Description | Motion correction for simultaneous EEG and fMRI data acquisition. |
Type Of Technology | New/Improved Technique/Technology |
Year Produced | 2015 |
Impact | Paper published (Neuroimage, 2016). |
URL | https://www.ncbi.nlm.nih.gov/pubmed/27157789 |
Description | Presentation workshop in Newcastle Australia |
Form Of Engagement Activity | A talk or presentation |
Part Of Official Scheme? | No |
Geographic Reach | National |
Primary Audience | Professional Practitioners |
Results and Impact | Satellite meeting to the Australian Cognitive Neuroscience Society |
Year(s) Of Engagement Activity | 2016,2017 |