Functional Connectivity

Tuesday 13 May 2014

Dany D'Souza¹, Andreas Bruns², Basil Künnecke², Daniel Alberati³, Edilio Borroni⁴, Markus von Kienlin², Annemie Van der Linden⁵, and Thomas Mueggler^2
1Pharma Research & Early Development Informatics, Disease & Translational Informatics, F. Hoffmann-La Roche Ltd., Basel, Basel Stadt, Switzerland,²Pharma Research & Early Development, DTA Neuroscience, Behaviour Pharmacology & Preclinical Imaging, F. Hoffmann-La Roche Ltd., Basel Stadt, Switzerland, ³Pharma Research & Early Development, DTA Neuroscience, Functional Neuroscience, F. Hoffmann-La Roche Ltd., Basel Stadt, Switzerland,⁴Pharma Research & Early Development, DTA Neuroscience, Biomarkers and Clinical Imaging, F. Hoffmann-La Roche Ltd., Basel Stadt, Switzerland, ⁵Bio-Imaging Lab, University of Antwerp, Antwerp, Belgium

In the present resting-state fMRI study we investigated whether an acute dose of a first or second generation antipsychotic (haloperidol, clozapine, & risperidone) or an mGlu2/3 agonist (LY354740) can reverse the hyper functional connectivity specifically across cortical areas in the rat brain elicited by acute treatment with ketamine. While acute antipsychotic dosing failed to block the effect of the NMDA antagonist (ketamine) we identified a set of brain regions, i.e. substantia nigra and motor cortex, commonly modulated by haloperidol, clozapine, or LY354740 likely reflecting the direct or indirect modulation of dopamine transmission originating in the nigrostriatal pathway, respectively.

Jingyuan Chen¹ and Gary Glover^1
1Stanford University, Stanford, CA, United States

In a prior study, we have demonstrated less variability of Pearson correlations with respect to the default-mode network (DMN) during working memory (WM) state compared to rest. Here, we attempt to employ co-activation patterns (CAPs) analysis to examine the fundamental changes in brain repertoires that underlie the macroscopic decrease in correlation variations during WM task, as shown by sliding-window analysis.

Sreenath Pruthviraj Kyathanahally^1,2, Karthik R Sreenivasan¹, Daniele Marinazzo³, Guorong Wu^3,4, and Gopikrishna Deshpande^1,5
1AU MRI Research Center, Department of Electrical and Computer Engineering, Auburn University, Auburn, Alabama, United States, ²Department of Clinical Research, Unit for MR Spectroscopy and Methodology, University of Bern, Bern, Switzerland, ³Department of Data Analysis, Ghent University, Ghent, Belgium, ⁴School of Life Science and Technology, University of Electronic Science and Technology of China, Chengdu, China, ⁵Department of Psychology, Auburn University, Auburn, Alabama, United States

Since the fMRI time series at each voxel is the convolution of an underlying neural signal with the hemodynamic response, there is a debate on whether the Default mode network(DMN) has a neural origin or is at least in part (or at most fully) a consequence of hemodynamic processes and physiological noise arising due to cardiac pulsation and respiration. In order to investigate this, we performed blind hemodynamic deconvolution of resting state fMRI data that was acquired with different TR and magnetic field strength. Subsequently functional connectivity maps were found using seed based correlation analysis on latent neuronal signals with a posterior cingulate seed in order to identify the DMN.

Bing Ji^1,2, Zhihao Li¹, Kaiming Li¹, Longchuan Li^1,3, and Xiaoping Hu^1
1Wallace H. Coulter Dept. of Biomedical Engineering, Emory University School of Medicine, Atlanta, GA, United States, ²University of Shanghai for Science & Technology, 200093, Shanghai, China, ³Department of Pediatrics, Marcus Autism Center, Children's Healthcare of Altanta, Emory University, Atlanta, GA, United States

In my study, we exploit the dynamic nature of FC in the parcellation of thalamus based its connectivities with the cortex and identify two dominant states, then derive two similar but distint parcellations results depending on these states. These parcellations, examined separately or combined, provide better correspondence with anatomic landmarks.

Molly G Bright¹ and Kevin Murphy^1
1CUBRIC, School of Psychology, Cardiff University, Cardiff, Cardiff, United Kingdom

Independent component analysis (ICA) can identify network structure in resting-state fMRI data. However, it is challenging to determine the origin of signal fluctuations in these networks. We apply breath-hold tasks that drive a global BOLD signal increase via the vascular response to hypercapnia. Using ICA, we demonstrate that neural and vascular networks can be identified and isolated, including those that share a similar network structure (e.g., default mode network). Our results suggest that vascular networks may be organised to support functional networks.

Silvia Francesca Storti¹, Ilaria Boscolo Galazzo¹, Alessandra Del Felice¹, Francesca Pizzini², Chiara Arcaro¹, Emanuela Formaggio³, Roberto Mai⁴, Alberto Beltramello², and Paolo Manganotti^1,3
1Department of Neurological and Movement Sciences, University of Verona, Verona, Italy, ²Department of Neuroradiology, AOUI of Verona, University of Verona, Verona, Italy, ³Department of Neurophysiology, Foundation IRCCS San Camillo Hospital, Venice, Italy, ⁴Epilepsy Surgery Center, Niguarda Hospital, Milan, Italy

Arterial spin labeling can be very useful for the detection of perfusion changes in drug-resistant focal epilepsy. In order to identify regions related to the epileptic focus, quantification of CBF and a statistical analysis were computed in each patient and compared with a template of normal perfusion. A seed-driven connectivity was also used to identify networks regions that are differently organized in epileptic patients compared to healthy subjects. The investigation allowed us to correctly identify the epileptogenic zone in patients, in whom the results were confirmed by surgical resection and subsequent seizure freedom.

Gloria Castellazzi^1,2, Fulvia Palesi^2,3, Stefania Bruno⁴, Ahmed T. Toosy⁵, Egidio D'Angelo^2,6, and Claudia A.M. Wheeler-Kingshott^7
1Department of Industrial and Information Engineering, University of Pavia, Pavia, PV, Italy, ²Brain Connectivity Center, National Neurological Institute C.Mondino, Pavia, PV, Italy, ³Department of Physics, University of Pavia, Pavia, PV, Italy, ⁴Overdale Hospital, Jersey, United Kingdom, ⁵Department of Brain Repair and Rehabilitation, UCL Institute of Neurology, London, United Kingdom, ⁶Department of Public Health, Neuroscience, Experimental Medicine, University of Pavia, Pavia, PV, Italy, ⁷NMR Research Unit, Department of Neuroinflammation, Queen Square MS, UCL Institute of Neurology, London, Italy

During the execution of complex “continuous” cognitive tasks, the brain elaborates information over multiple domains and time scales, integrating it across space and over time. In literature, only few rs-fMRI works report how resting state networks (RSNs) change over space and time when stimulated by external inputs. We investigated the dynamic changes in brain activity occurring in subjects listening to a narrated story. Results show that RSNs respond to the stimulus with specific dynamics of alteration and suggest the existence of a spatiotemporal hierarchy of changes, the levels of which depend on the specific activity each network is involved in.

Mary Beth Nebel^1,2, Ani Eloyan³, Carrie Nettles¹, Kristie Sweeney¹, Katarina Ament¹, Rebecca Ward¹, Ann S Choe^1,2, Anita D Barber^1,2, Brian S Caffo³, James J Pekar^1,2, and Stewart H Mostofsky^1,2
1Kennedy Krieger Institute, Baltimore, MD, United States, ²Johns Hopkins School of Medicine, Baltimore, MD, United States, ³Johns Hopkins School of Public Health, Baltimore, MD, United States

One problem experienced by children with autism that is potentially critical for acquiring social skills is difficulty imitating others’ actions, which depends on visual-motor integration; however, it is unclear what brain mechanisms contribute to this deficit. Using resting state functional MRI, we show that children with autism exhibit significantly stronger anticorrelation between motor and visual areas compared to their typically developing (TD) peers, and the stronger the anticorrelation between motor and visual networks, the more severe their autistic traits. In TD children, motor-visual functional connectivity strength was correlated with imitation performance; children with stronger positive visual-motor coupling were better imitators.

Xiaojia Liu^1,2, Xiang Li², Fuqing Zhou², Jiaolong Cui², Adrian Tsang^1,3, Iris Y. Zhou^1,3, Ed X. Wu^1,3, and Yong Hu^2
1Laboratory of Biomedical Imaging and Signal Processing, The University of Hong Kong, Hong Kong, China, ²Department of Orthopaedics and Traumatology, The University of Hong Kong, Hong Kong, China, ³Department of Electrical and Electronic Engineering, The University of Hong Kong, Hong Kong, China

Resting state networks in the human cervical spinal cord has only been scarcely explored. In this study, we investigated the resting state network using a clinically relevant 3T whole body MRI scanner from 14 healthy subjects. The correlation coefficient computed from rsfMRI images between each ventral or dorsal horn of different segments was used to generate the correlation matrix. Segments C2 and C6 demonstrated stronger correlations with other segments. Segment C2 has a stronger inter-segment correlation than other segments. Functional connectivity distribution among segments was detected, which demonstrated the neural network in the human cervical spinal cord.

Giulio Ferrazzi^1,2, Maria Kuklisova Murgasova^1,2, Tom Arichi³, Joanna Allsop^1,2, Christina Malamateniou^1,2, Mary Rutherford^1,2, Shaihan Malik^1,2, Paul Aljabar^1,2, and Joseph V. Hajnal^1,2
1Center for the Developing Brain, King's College London, London, United Kingdom, ²Division of Imaging Sciences & Biomedical Engineering, King's College London, London, United Kingdom, ³Department of Bioengineering, Imperial College London, London, United Kingdom

During typical fetal Resting State fMRI examinations, maternal respiration and fetal movement together result in large scale and unpredictable motion. Conventional fMRI processing methods, which assume that brain movements are infrequent or at least small, are not suitable. We seek to address the problem of fetal motion in fMRI using image registration to place the acquired data into a self-consistent anatomical space. Bias field and spin history corrections are also discussed, aiming at achieving a robust framework that allows as much as possible, ideally all acquired data, to be retained as part of the final network analysis.