Modelling influence of structural brain connectivity on functional brain connectivity and its application for early diagnosis of cognitive impairment
Lead Research Organisation:
University of Southampton
Department Name: Electronics and Computer Science
Abstract
The neural function of the brain is characterized by activated brain regions and the connectivity among them. There are three types of connectivity between brain regions - structural, functional and effective connectivity. The relationship between anatomical (structural), functional and effective connectivity is still a grey area and hence, working towards exploring and understanding this area is a major interest in theoretical neuroscience. Functional connectivity is temporally dynamic, task dependent and changes rapidly in temporal scale of a millisecond. Whereas, structural connectivity of brain is relatively static and does not change over days and month. It is still unknown, how a static structural connectivity network affects the occurrence of task-dependent dynamic functional connectivity or why two structurally connected brain regions, are not functionally connected and vice-versa. Studies have shown, the underlying cause for many neuro-degenerative diseases is the disruptions in neural connections. So understanding the relationship between structural and functional connectivity is important for understanding the impairments characteristics in the brain networks. The purpose of this work is to characterize the structural connectivity and its influence on functional connectivity of brain by applying circuit theory based modelling approach. Modelling structural connection using circuit theory will allow the analysis of signal propagation in both time and frequency domains. So far the studies on correlation between structural and functional connectivity were done from time domain perspective of signal processing. But measurement of phase correlation for functional connectivity signifies that underlying physical connection of functional connectivity has filter like properties and holds the frequency-phase characteristics. In this work, we will (1) define the brain areas (ROIs) by a non-anatomical equal area parcellation process from structural MRI data, (2) extract white matter tracts from diffusion MRI data, (3) extracted geometrical properties of white matter tracts, (4) design the circuit model for single axon (5) define the transfer function for single axon from the circuit model and analyse its frequency response. In future work, we will model the coupling effects between two myelinated axons during signal propagation in circuit design. The goal is to use this coupling equation to join the transfer functions of a single axon, to model a dynamic system for white matter tracts. Analysing signal propagation characteristics of this system will give the frequency-dependent phase relationship between the adjacent ROIs from which is basically the functional connectivity. We will validate the same with the fMRI/EEG experimental data.
Organisations
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509747/1 | 01/10/2016 | 30/09/2021 | |||
| 1807814 | Studentship | EP/N509747/1 | 01/10/2016 | 30/10/2019 | Sarbani Das |
| Description | 1. A new parcellation methodology has been developed to parcellate the cortical surface in non-anatomical equal size areas from structural MRI image. 2. A structural connectome has been build by processing the di?usional MRI data, the white matter tracts were obtained among the non-anatomical parcellated areas. 3. A circuit model for the single axon to simulate Saltatory conduction through myelinated axon has been designed. 4. Define cut-off frequency of signal propagation through the myelinated section of a single axon. 5. Simulated coupling e?ects between two myelinated axons during signal propagation by considering low resistance value for the extra-cellular ?uid using the circuit model. 6. Determine the dominant firing frequency of coupled neuron cells and establish the relationship between the synchronized dominant firing frequency and external factors such as input current and coupling strength of coupled neuron cells based on Hodgkin-Huxley neuron cell model. 7. Simulate signal propagation through a bunch of axons using Pspice and determine the frequency response of the same. |
| Exploitation Route | 1. In the form of published paper in conference and journals. 2. By desgining software tool. |
| Sectors | Healthcare |
| Title | Equal Parcelletion of cortical surface into N number of equal areas |
| Description | We parcellated the cortical surface into any N equal sized areas using Freesurfer where each of the hemispheres were parcellated into N/2equal regions. Each of the parcellated areas were created as label files first and then they were put together to create the annotation files for each hemisphere. Each of the parcellated areas were assigned a colour code. We arranged the parcellated areas of each hemispheres in such a way that the symmetry of left and right hemispheres is maintained. |
| Type Of Material | Improvements to research infrastructure |
| Year Produced | 2017 |
| Provided To Others? | No |
| Impact | To enable the abstract treatment of neural systems as networks, the brain of the studied subject is registered to a generic brain with defined Brodmann areas. Brodmann areas were originally defined using cytoarchitectural differences between brain regions. However, the borders of the areas, with a few exceptions, do not match the sulci and gyri of the cortical surface or any other external morphological features [5,6]. The question arises regarding the accuracy of mapping those areas onto the brain in each individual case, especially in the context of brain maturation or abnormal anatomy. Cases of astounding cerebral reorganization after brain damage (neuroplasticity) have been reported bbppiifrfkkfjfjxkf[7]. To take an extreme example, in young children undergoing hemispherectomy for the treatment of intractable epilepsy, cortical plasticity and change of connectivity allow the contralateral hemisphere to assume the functions of the lost hemisphere without significant neurologic deficits [8]. It is questionable whether the registration of such a reorganized brain to a standard atlas based on the Brodmann areas will provide valuable insights into the newly adjusted brain network. Similarly, albeit not as dramatic, biases can be introduced when studying the developing brain of neonates, who have immature sulcation, if one partially or completely relies on adult brain atlases. Even in the case of normal anatomy of the adult brain, different subjects can have, for example, different dominant hemispheres. Is there a way to analyze and compare brain networks without the constraints of standardized anatomy. This cortical parcellation method will 1. enable the abstract treatment of neural systems as networks, 2. be helpful in the cases of astounding cerebral reorganization after brain damage (neuroplasticity) , 3. help where registration of a reorganized brain to a standard atlas based on the Brodmann areas does not provide valuable insights into the newly adjusted brain network such as young children undergoing hemispherectomy for the treatment of intractable epilepsy, cortical plasticity. |