Mofid Children Hospital

Network based connectivity analysis as applied on EEG data for an integral brain network analysis of autistic patients subject to eye to eye contact syndrome 

Abstract  

This research is designed for a focused study on a narrow class of autistic patients that are categorized as eye to eye contact type autism having characteristic marker dominantly observed in these subjects.  We will confine our study to high functional category of subjects that exhibit the designated characteristic behavioral marker in a dominant manner. The extent of high functionality will be determined by neurologist. The study also covers carefully chosen control healthy groups of similar age and gender for reference purposes and for comparative analysis. Noting that characteristic feature of autism is abnormalities in coordinated functioning of brain networks [Jun Wan et al, 2013].[Coben et al. 2008], [Belmonte M. K.et.al] the focus of this study is on brain connectivity analysis for which we will consider application of concepts and methods of graph theory as well as multivariate connectivity approaches on EEG data. We intend to consider both global and local properties of brain’s wiring diagram (networks and subnetworks) in patients with the designated autism spectrum disorder (ASD). We will also consider application of suitable time-frequency based signal processing methods using wavelets as applied on EEG data.

As part of study, we will consider connectivity of brain networks reported in literature that are affected by autism. These include fronto-posterior cortical section as well as salient brain network.  Salience network is reported to exhibit consistently atypical pattern of connectivity in autistic patients [1,3]. Therefore, changes in connectivity characteristics of salience network can be of significance and is used for understanding sensory over-responsivity (SOR). SOR is a common impairing condition in ASD subjects. On this line, we will focus on intrinsic network designated in literature to model attention to internal and external stimuli that is reported to carry useful information for characterizing autistic patients, namely "over-responsibility”. Overresnsivity is considered as a “difficulty in down regulating brain to sensory stimuli and failure to habituate.[Green SA, Hernandez L, Tottenham N]

Our analysis of brain networks will also include fronto-parietal brain network characterizing functional connectivity. On this line, we will consider connectivity analysis under set of suitably chosen primary tasks. Examples of these tasks include visionary task as well as those of attention.   Our connectivity analysis will include investigation of local, global, short and long distance connectivity analysis to be applied on EEG data. For investigation of causality, we will apply wavelet transform on EEG data as well as usage of Granger causality criteria.  We plan to apply wavelet-based coherence analysis in which causality will be investigated within transformed coefficient domain.

Explanatory Remarks

In this project, testing and data collection from control and autistic subjects is to be carried out in two stages with a time gap of 6-9 months .Identical tests procedures and protocols will be used for the two phases. This will allow to examine the changes in connectivity results due to growth and aging on both groups and evaluate changes in both groups. Differences from the earlier test results and connectivity measures will be highly informative as to which measures remain dominant in autistic subjects adding to the accuracy of the results. It can also be used to guide us which measures in both healthy and autistic subjects are reflective of brain growth or lack of it when autistic subjects are compared with healthy subjects.

 This project is designed as an interdisciplinary project with an active participation of medical doctors and neurologists. It is a project with real world application in contrast to a simulation type projects. For implementation and tests, carefully selected subjects – in this case children- from both control and autistic children will be tested under a preplanned test protocols. Details of test and project implementation including selection of children for tests and test standards are given in part 2 of this proposal. For data analysis and signal processing, we will utilize some of the open source software packages discussed in item No.4 that section “Further Details” of this proposal. They will be modified or expanded in accordance with the requirements of data analysis in the project.