New Methods to Deliver Longitudinal Optical Images of Infant Brain Development
Lead Research Organisation:
University College London
Department Name: Medical Physics and Biomedical Eng
Abstract
The aim of my PhD is to develop novel methods to model the effects of head and brain growth on light migration through the heads of infants aged two years and under; through the use these methods another aim is to reconstruct images of typical and atypical functional brain activation of the first two year of life.
Functional near infrared spectroscopy (fNIRS) is an optical monitoring technique which uses sources and detectors of light placed on the surface of the head to illuminate the brain and measure changes in light attenuation which indicate changes in the concentration of haemoglobin in the brain.
fNIRS has been used to measure functional activation in order to investigate the development of the infant brain. The BRIGHT (BRain Imaging for Global HealTh) project is a longitudinal study acquiring data from 200 infants in the Gambia and 60 infants in the UK at six time points between birth and their second birthday. The BRIGHT project seeks to investigate the effect of malnutrition and early adversity on infant brain development. As part of BRIGHT, fNIRS is used to measure localised haemoglobin concentration changes across the cerebral cortex of the brain in response to a range of cognitive stimuli: a localised increase in haemoglobin concentration is a marker of functional brain activation.
Functional images mapping haemoglobin concentration changes in the cerebral cortex can be obtained, or reconstructed, from fNIRS data; this requires prior knowledge of head structural anatomy, and it is this structural anatomy that is used to obtain a model of photon migration through a given subject's head. Such structural anatomical information can be obtained from an MRI scan of an individual - however this would undermine some of the key advantages of fNIRS as a technique which include its portability and its use to investigate subjects for whom being placed in a MRI scanner isn't routine.
My project aims to produce an age-specific atlas, using infant structural MRI data from existing databases, which maps the spatial distribution of fundamental head tissues (grey matter, white matter, cerebrospinal fluid and extracerebral (skull and scalp) tissue) of infants at each age being studied. This atlas can then be used as a substitute for individual structural information and thus will enable the reconstruction of images of functional activation in infants without needing an individual's own MRI scan. The ultimate goal of this aspect of my PhD project is to develop a pipeline in order to reconstruct longitudinal images of brain function to chart the developmental trajectory of both Gambian and UK infants using fNIRS data collected as part of BRIGHT. This will first need work to develop protocols to more accurately determine positions of sources and detectors on an infant's head, another aspect of the PhD project.
The focus of my project will also include investigating image reconstruction techniques for neonatal and preterm-born infants. Such work will entail confronting similar challenges as above, such as producing and validating an age-specific atlas for neonatal infants (whose brain and head structure vary significantly on timescales of weeks). Previous work has been completed regarding image reconstruction using fNIRS data from neonatal and preterm-born infants as part of a collaboration between King's College London and University College London; I will therefore be working with King's College for that phase of my project. It is anticipated that this work will contribute to an ongoing project regarding which is investigating techniques to map sensorimotor disruption in infants at risk of cerebral palsy.
Functional near infrared spectroscopy (fNIRS) is an optical monitoring technique which uses sources and detectors of light placed on the surface of the head to illuminate the brain and measure changes in light attenuation which indicate changes in the concentration of haemoglobin in the brain.
fNIRS has been used to measure functional activation in order to investigate the development of the infant brain. The BRIGHT (BRain Imaging for Global HealTh) project is a longitudinal study acquiring data from 200 infants in the Gambia and 60 infants in the UK at six time points between birth and their second birthday. The BRIGHT project seeks to investigate the effect of malnutrition and early adversity on infant brain development. As part of BRIGHT, fNIRS is used to measure localised haemoglobin concentration changes across the cerebral cortex of the brain in response to a range of cognitive stimuli: a localised increase in haemoglobin concentration is a marker of functional brain activation.
Functional images mapping haemoglobin concentration changes in the cerebral cortex can be obtained, or reconstructed, from fNIRS data; this requires prior knowledge of head structural anatomy, and it is this structural anatomy that is used to obtain a model of photon migration through a given subject's head. Such structural anatomical information can be obtained from an MRI scan of an individual - however this would undermine some of the key advantages of fNIRS as a technique which include its portability and its use to investigate subjects for whom being placed in a MRI scanner isn't routine.
My project aims to produce an age-specific atlas, using infant structural MRI data from existing databases, which maps the spatial distribution of fundamental head tissues (grey matter, white matter, cerebrospinal fluid and extracerebral (skull and scalp) tissue) of infants at each age being studied. This atlas can then be used as a substitute for individual structural information and thus will enable the reconstruction of images of functional activation in infants without needing an individual's own MRI scan. The ultimate goal of this aspect of my PhD project is to develop a pipeline in order to reconstruct longitudinal images of brain function to chart the developmental trajectory of both Gambian and UK infants using fNIRS data collected as part of BRIGHT. This will first need work to develop protocols to more accurately determine positions of sources and detectors on an infant's head, another aspect of the PhD project.
The focus of my project will also include investigating image reconstruction techniques for neonatal and preterm-born infants. Such work will entail confronting similar challenges as above, such as producing and validating an age-specific atlas for neonatal infants (whose brain and head structure vary significantly on timescales of weeks). Previous work has been completed regarding image reconstruction using fNIRS data from neonatal and preterm-born infants as part of a collaboration between King's College London and University College London; I will therefore be working with King's College for that phase of my project. It is anticipated that this work will contribute to an ongoing project regarding which is investigating techniques to map sensorimotor disruption in infants at risk of cerebral palsy.
Organisations
People |
ORCID iD |
| Clare Elwell (Primary Supervisor) | |
| Liam Collins Jones (Student) |
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509577/1 | 01/10/2016 | 24/03/2022 | |||
| 1922799 | Studentship | EP/N509577/1 | 25/09/2017 | 25/09/2021 | Liam Collins Jones |
| Description | Functional near-infrared spectroscopy is a technique where sources and detectors of light are placed on the scalp, and is used to study brain activation. When a region of the brain is activated, more blood flows to that part of the brain; blood is a much stronger absorber of light than brain tissue, so a region of the brain that's activated absorbs more like than when it is at rest. The difference in light reaching the detectors between a resting and an activated state can tell us where in the brain has been activated. Unlike other imaging techniques that can be used to study brain activation, functional near-infrared spectroscopy is relatively tolerant of motion and so it is used to study activation in the baby brain. Skull-and-scalp thickness increases with age, affecting how light travels through the head and the amount of light that reaches the detectors, which means that measures of activation can be difficult to compare between babies at different ages. To take into account the different ways that light travels through the brain at different ages, structural models of the brain need to be constructed so that the movement of light through the head can be simulated computationally, and that model of light movement can then be combined with the data on changes in light intensity to locate where in the brain activation has happened. To get a scan of the structure of the head every baby under investigation could undergo an MRI scan, but this isn't feasible. Previous work has looked at combining MRI data from groups of previously scanned babies to produce a generic model of brain structure for babies at different ages post-birth, but such models do not include any non-brain tissue, which is crucial to consider when thinking of how light travels through the head from sources to detectors. So far, my work has been split between two projects: one project which involved constructing structural models of the newborn head, and another project investigating structural models of babies aged from 1 to 24 months. Using state-of-the-art MRI data, as part of my PhD I have produced a database of structural models of the head from over 200 newborn individuals at a wide range of ages post-conception. In this work, a method to derive non-brain tissue was devised and the accuracy of the non-brain tissue was also evaluated. It is envisaged that this database of structural models can be used to find a best-matching for a newborn patient or research subject based on measures that can be easily measured such as head size and age post-conception; as part of this work, I conducted an evaluation of a simple way of implementing this database to see what the error would be in locating an activated region of the brain using a matched structural models as opposed to a subject's own structural information. My PhD is linked to another project, the Brain Imaging in Global Health project, which is a study of two groups of babies - one in the Gambia and another in the UK - from birth to two years which aims to investigate the effect of malnutrition and other early adversity on brain development. Data is collected from these babies at 6 age points from 1 month to 2 years - as yet, no head models inclusive of brain tissue and non-brain tissue are freely available. After an extensive literature search, MRI data from groups of babies at these ages were found and used to produce models of the developing brain that can be used to simulate light movement from sources to detectors. At present, my PhD is focusing on seeing how best to compare images of brain activation between ages based on how much light reaches different parts of the brain, and anatomically labelling the brain regions where activation occurs. |
| Exploitation Route | The database of structural priors for newborns can be used to produce more spatially accurate images of brain activation, which it is hoped will improve the speed and accuracy of brain injury in newborns. The work completed for babies from 1 month to 2 years can be used as part of an effort to produce growth curves for brain development in high-income countries and low- and middle-income countries. |
| Sectors | Communities and Social Services/Policy,Healthcare |