Clara Lisazo^1,2, Natalia Noemí Massaccesi Bove^1,2, Daniela Zanchi³, María Paula Del Pópolo^1,2, Daniel Fino^1,2,4, Federico Julián González Nicolini^1,4,5, Sergio Mosconi³, Pedro Pablo Ariza^1,2, Trinidad Gonzalez Padín^1,2, Raúl Otoya⁶, Roberto Isoardi^3,5, Thomas Martin Doring⁷, and Sebastián Moguilner^5,8
1MRI Department, Fundación Escuela de Medicina Nuclear, Mendoza, Argentina, ²MRI Department, Fundación Argentina para el Desarrollo en Salud, Mendoza, Argentina, ³Nuclear Medicine Department, Fundación Escuela de Medicina Nuclear, Mendoza, Argentina, ⁴Instituto Balseiro, Universidad Nacional de Cuyo, San Carlos de Bariloche, Argentina, ⁵GQNYCS, Comisión Nacional de Energía Atómica, Ciudad Autónoma de Buenos Aires, Argentina, ⁶Neuromed, Mendoza, Argentina, ⁷GE Healthcare, Rio de Janeiro, Brazil, ⁸Department of Neurology, Massachusetts General Hospital and Harvard Medical School, Boston, MA, United States

Detecting epileptogenic zones (EZ) in subjects with non-lesional epilepsy can represent a challenge, and it involves the analysis of data from various diagnostic procedures. In this study, seed-based structural and functional connectivity matrices were obtained from DWI and resting-state fMRI for each patient and compared with a control group. Then, we assessed whether significantly different regions (p<0.001) coincided with foci of abnormal metabolism on PET images, and if these ROIs corresponded with the subject’s EZ. This reduced the false positive rate and improved specificity of the connectivity analysis, optimizing EZ localization in patients with non-lesional epilepsy.

Cláudia Fonseca¹, Inês Esteves¹, Marta Xavier¹, Ana Fouto¹, Amparo Ruiz-Tagle¹, Nuno A. Silva², Rita G. Nunes¹, Raquel Gil-Gouveia³, Joana Cabral⁴, and Patrícia Figueiredo^1
1Institute for Systems and Robotics, Instituto Superior Técnico, University of Lisbon, Lisbon, Portugal, ²Learning Health, Hospital da Luz, Lisbon, Portugal, ³Neurology Department, Hospital da Luz, Lisbon, Portugal, ⁴Life and Health Sciences Research Institute, University of Minho, Braga, Portugal

Sliding window Pearson correlation (SW) is the most commonly used approach for estimating dynamic functional connectivity (dFC). However, instantaneous phase coherence (PC) has gained popularity as it yields frame-by-frame dFC estimates. This work aimed to compare both metrics by analysing the mean lifetime, probability of occurrence and spatial similarity of dFC states with the canonical resting-state networks (RSNs). We found that the state lifetimes increase in SW compared to PC and with window length, worsening the detection of RSNs for smaller datasets. These findings indicate that the temporal blurring induced by SW compromises the ability to detect faster network dynamics.

Quimby Lee¹, Jingyuan E. Chen^2,3, Gregory Wheeler⁴, and Audrey P. Fan^1,4
1Department of Neurology, University of California, Davis, Davis, CA, United States, ²Department of Radiology, Harvard Medical School, Boston, MA, United States, ³Massachusetts General Hospital, Athinoula A. Martinos Center for Biomedical Imaging, Boston, MA, United States, ⁴Department of Biomedical Engineering, University of California, Davis, Davis, CA, United States

Heart rate variability (HRV), which is reflective of autonomic regulation, induces vascular effects in low (0.05-0.15Hz) and high (0.15-0.4Hz) frequency bands that can influence the blood oxygen level dependent (BOLD) signal of resting-state functional magnetic resonance imaging (rsfMRI). In this study, we utilized a wavelet transform coherence analysis to identify spectral and temporal differences in phased-locked behavior between HRV and resting-state network (RSN) time courses. Subjects differed in the frequency band with greatest time-averaged coherence and percentage of time with significant coherence of HRV and RSN signals.

Tamas Spisak¹, Balint Kincses^2,3, Raviteja Kotikalapudi¹, Matthias Zunhammer², Frederik Schlitt², Tobias Schmidt-Wilcke^4,5, Zsigmond Tamas Kincses³, and Ulrike Bingel^2
1Laboratory of Predictive Neuroimaging, Institute for Diagnostic and Interventional Radiology and Neuroradiology, University Hospital Essen, Essen, Germany, ²Klinik für Neurologie, University Hospital Essen, Essen, Germany, ³Department of Neurology, University of Szeged, Szeged, Hungary, ⁴Institut für Klinische Neurowissenschaften und Medizinische Psychologie, Heinrich Heine Universität, Düsseldorf, Germany, ⁵Neurozentrum, Bezirksklinikum Mainkofen, Deggendorf, Germany

Center effects significantly limit the generalizability of brain imaging-based biomarker candidates. Although our previously published resting state functional connectivity-based predictive signature for pain sensitivity (the RPN-signature) showed remarkable out-of-center generalizability, it remained unclear which connectivity features are the most generalizable across study centers.

Here, we re-trained the RPN-signature on multi-center data and found that it outperforms the single-center model in all three centers (explained variance: 26-38% vs. 16%-19%). Our results highlight that neurobiological interpretation of feature importance in predictive modelling is constrained both by center-specific artifacts and by certain characteristics (e.g. regularization) of the applied machine learning algorithm.

Kun Yue¹, Jason M Webster², Thomas J Grabowski², Ali Shojaei¹, and Hesamoddin Jahanian^2
1Department of Biostatistics, University of Washington, Seattle, WA, United States, ²Department of Radiology, University of Washington, Seattle, WA, United States

With advances in experimental therapeutics for Alzheimer's Disease (AD) the need for an accurate, non-invasive and widely available AD biomarker is more pressing than ever. Resting-state functional connectivity in default mode network is a candidate biomarker that is gaining traction in the field. However, the traditional stationary measurement of the default mode network connectivity cannot capture complicated dynamic patterns of functional connectivity that exist in the brain. Here, we have proposed a novel, reliable technique, based on Dynamic Condtional Correlation model, to quantify the dynamic functional connectivity in the brain and evaluated its sensitivity to cerebrospinal fluid biomarkers in AD.

Francesca Saviola¹, Stefano Tambalo¹, Dimitri Van De Ville^2,3, and Jorge Jovicich^1
1University of Trento, Center for Mind/Brain Sciences, Rovereto, Italy, ²Institute of Bioengineering, Center for Neuroprosthetics, Ecole Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland, ³Department of Radiology and Medical Informatics, University of Geneva (UNIGE), Geneva, Switzerland

The advent of fast fMRI acquisition techniques has enabled whole-brain acquisitions with sub-second temporal resolution enriching the information available throughout the time course. However, there is no consensus about how the application of frame-wise analysis,  performed to better understand functional brain fluctuations, could benefit from high temporal resolution fMRI. In this work, we demonstrate the potential need to: (i) gain a finer understanding of signal and noise signatures peculiar to fast acquisition and (ii) extend the models to better estimate dynamic functional connectivity in rapidly sampled fMRI time series.

Anita Monteverdi^1,2, Fulvia Palesi², Claudia AM Gandini Wheeler-Kingshott^1,2,3, and Egidio D'Angelo^1,2
1Brain Connectivity Center, IRCCS Mondino Foundation, Pavia, Italy, ²Brain and Behavioural Sciences, University of Pavia, Pavia, Italy, ³NMR Research Unit, Queen Square Multiple Sclerosis Centre, Department of Neuroinflammation, UCL Queen Square Institute of Neurology, UCL, London, United Kingdom

A comprehensive assessment of multiple resting-state networks is still lacking and the characterization of their excitatory/inhibitory (E/I) balance is an open field of research. In this exploratory work we performed a first characterization of the E/I balance in resting-state networks (visual, somatomotor, attention, limbic, frontoparietal, DMN) both in healthy and neurodegenerative conditions (Alzheimer’s Disease, Frontotemporal Dementia, Amyotrophic Lateral Sclerosis) taking advantage of a Virtual Brain based approach. Our results provided a new strategy to simultaneously characterize multiple networks at single-subject level, deepening their E/I balance and opening new perspectives for biomarkers research.

Anirban Sengupta¹, Arabinda Mishra¹, Feng Wang¹, Li Min Chen¹, and John C Gore^1
1Vanderbilt University Medical Center, Nashville, TN, United States

The goal of this study was to investigate the nature of spontaneous BOLD fluctuations in white-matter (WM) of spinal-cord and to use these to identify the intrinsic functional architecture of WM tracts and their correlations with the gray-matter (GM) hubs. Connectivity measures were obtained using resting state BOLD signals between WM-WM and WM-GM regions, followed by network analysis using graph-theory. We found WM and GM hubs on the dorsal side exhibit greater temporal correlation as exhibited by their stronger node strength in resting state. Also, within segment WM-WM and WM-GM correlations were found to be stronger than those between segments.