System Identification and Signal Processing for Neuro Imaging Data Analysis

Lead Research Organisation: University of Sheffield
Department Name: Automatic Control and Systems Eng

Abstract

Functional Magnetic Resonance Imaging [fMRI] provides observations of the Blood Oxygen Level Dependent [BOLD] signal. fMRI does not measure neural activity directly and hence a core problem is how to interpret the BOLD signal to make inferences about the neural activity and interactions. The main objective of this research proposal is to apply nonlinear signal processing and system identification methods to identify nonlinear mesh models associated with the core mechanisms between activation and flow from data collected across a wide range of stimulus conditions. This is a highly complex and multi-disciplinary project involving neuroscientists, psychologists, and control and systems engineers.
 
Description We made 3 major findings during this grant. (1) the relationship between cerebral blood flow and volume can be modelled as a compliant balloon with model parameters related to the compliant properties of blood vessels. (2) The neural signals recorded by a microelectrode in the somatosensory cortex can be decomposed into subsystems (brain stem, thalamus and barrel cortex) via mathematical modelling. (3) The neurovascular coupling relationship can be largely modelled by a linear time-invariant system within the range of stimulation used. The structure of the model indicated that blood vessels dilate and constrict during the evoked response.
Exploitation Route All the findings from the EPSRC grant provided foundations for our current research work on balanced neural excitation and inhibition funded by the BBSRC. In the long term we aim to relate invasive neural and hemodynamic recordings to non-invasive recordings such as EEG and fMRI, thus providing algorithms to enable researchers and medical workers to interpret human EEG and fMRI signals to the underlying changes in neural activity.
Sectors Healthcare
 
Description Our findings obtained from this EPSRC grant are crucial to our current research in understanding the local field potential recordings in terms of neural excitation and inhibition through experimental design and mathematical modelling. The important application of this work is the interpretation of human EEG signals using the balance and interaction of neural excitation and inhibition.
First Year Of Impact 2004
Sector Healthcare
Impact Types Societal
 
Description Responsive mode
Amount £615,105 (GBP)
Funding ID BB/K010123/1 
Organisation Biotechnology and Biological Sciences Research Council (BBSRC) 
Sector Public
Country United Kingdom
Start 04/2013 
End 06/2016
 
Title Concurrent LFP and LDF data for different stim duration 
Description Stimulus: electrical stimulation of the whole whisker pad. 1.2mA, square pulse of 0.3ms. Stim frequency: 5Hz. Stim duration: 2s, 8s and 16s. Probing stim: 2s at inter-block-intervals 0.6s,1s,2s,3s,4s,6s,8s Each trial: 60s consisting of a conditioning block (2,8 or 16s) followed by a probing block (2s) at 7 different intervals Each run: 21 different trials interleaved Each experiment: 10 runs 
Type Of Material Database/Collection of data 
Provided To Others? No  
Impact A mathematical model of neurovascular coupling was derived. The data also implied blood vessel dilation and constriction during neural activity. This finding formed part of the investigation of my current grant from the BBSRC.