Statistical Connectomics

Chung, Jaewon; Bridgeford, Eric; Arroyo, Jesús; Pedigo, Benjamin D.; Saad-Eldin, Ali; Gopalakrishnan, Vivek; Xiang, Lei; Priebe, Carey E.; Vogelstein, Joshua T.

doi:10.31219/osf.io/ek4n3

Cited by 9 publications

(10 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, correlation or coherence between each pair of EEG channels is computed and organized into an adjacency matrix, on which summary statistics are derived including degree distribution and variants of centrality [11, 12, 13, 10, 15]. Such network-based measures are not based on well formulated hypotheses of the role of the epileptic tissue in the iEEG network, and many different networks (adjacency matrices) can have identical summary statistics resulting in ambiguous interpretations of such measures [38].…”

Section: Discussionmentioning

confidence: 99%

Neural Fragility as an EEG Marker of the Seizure Onset Zone

Huynh

Fitzgerald

et al. 2019

Preprint

View full text Add to dashboard Cite

Epilepsy is a global epidemic and 30% of the 60 million patients do not respond to medication treatment. The only treatment options for patients with medically refractory epilepsy are surgical removal or electrical stimulation of the epileptogenic zone (EZ) i.e. the source of their seizures. Despite extensive evaluations with neuroimaging, visual EEG analysis and clinical testing, surgical success rates vary between 30-70%. Currently, no computational methods have been translated into the clinic to assist in localizing the EZ. Here, we applied a dynamical network model that quantifies the fragility of nodes within a patient's intracranial EEG (iEEG) brain network. Fragility is quantified as the minimal amount of perturbation that must to be applied to a node's influence on a "balanced" network to cause imbalance. Here, a balanced network is one in which the connectivity between excitatory and inhibitory nodes render a stable system, and an imbalanced network is unstable and hence can generate seizures. Using iEEG data from 91 patients treated across 5 epilepsy centers (44 successes, 47 failures), we demonstrated that nodal fragility is greater in electrodes within the EZ. In addition, we compared fragility of iEEG nodes to 7 frequency-based and 14 graph theoretic features of the EZ in both seizure (n=91) and non-seizure data (n=54). We calculated a confidence statistic, defined as the ratio of the value of a given feature averaged across electrodes in the clinically annotated seizure onset zone to its average across all other electrodes. Fragility has a significantly greater effect size difference between surgical outcomes when compared to other features. This novel feature, outperformed the most popular iEEG features when comparing across surgical outcomes, possibly defining a superior network-based EEG fingerprint for the EZ.1 Also reachable at adam2392 at gmail dot com. For more information, visit the NCSL website: https://sarmalab.icm.jhu.edu/.

show abstract

Section: Discussionmentioning

confidence: 99%

Neural Fragility as an EEG Marker of the Seizure Onset Zone

Huynh

Fitzgerald

et al. 2019

Preprint

View full text Add to dashboard Cite

show abstract

“…The homophilic effect is the tendency of connectivity to be stronger between regions in the same hemisphere than different hemispheres. To formalize this relationship, we leverage the Structured Independent Edge Model [3]. We use the Mann-Whitney U -statistic [8] to investigate the magnitude of the homotopic and homophilic effects.…”

Section: Control Numerical Experimentsmentioning

confidence: 99%

When no answer is better than a wrong answer: a causal perspective on batch effects

Bridgeford

Powell

Kiar

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Batch effects, undesirable sources of variance across multiple experiments, present a substantial hurdle for scientific and clinical discoveries. Specifically, the presence of batch effects can create both spurious discoveries and hide veridical signals, contributing to the ongoing reproducibility crisis. Typical approaches to dealing with batch effects conceptualize 'batches' as an associational effect, rather than a causal effect, despite the fact that the sources of variance that comprise the batch -- potentially including experimental design and population demographics -- causally impact downstream inferences. We therefore cast batch effects as a causal problem rather than an associational problem. This reformulation enables us to make explicit the assumptions and limitations of existing approaches for dealing with batch effects. We therefore develop causal batch effect strategies---Causal Dcorr for discovery of batch effects and Causal ComBat for mitigating batch effects -- which build upon existing statistical associational methods by incorporating modern causal inference techniques. We apply these strategies to a large mega-study of human connectomes assembled by the Consortium for Reliability and Reproducibility, consisting of 24 batches including over 1700 individuals to illustrate that existing approaches create more spurious discoveries (false positives) and miss more veridical signals (true positives) than our proposed approaches. Our work therefore introduces a conceptual framing, as well as open source code, for combining multiple distinct datasets to increase confidence in claims of scientific and clinical discoveries.

show abstract

“…One way to determine if a motif is significant within a real biological dataset is to compare its observed frequency to its expected frequency within a random graph model. There are many different random graph models, each with different properties and capturing a different aspect of real world data [43]. Here, we perform subgraph search on five different random graph models.…”

Section: Comparison To Random Graph Modelsmentioning

confidence: 99%

“…The Watts-Strogatz model addresses the edge dependence issue more directly, recognizing that many real world networks have high clustering coefficients, which is a measure of the extent to which two adjacent nodes have similar neighbors [29]. The Barabási-Albert model is a preferential-attachment model that ensures a power-law degree distribution, making it a scale free graph model [30,43].…”

Section: Comparison To Random Graph Modelsmentioning

confidence: 99%

DotMotif: An open-source tool for connectome subgraph isomorphism search and graph queries

Matelsky

Reilly

Johnson

et al. 2020

Preprint

View full text Add to dashboard Cite

As connectomics datasets continue to grow in size and complexity, methods to search for and identify interesting graph patterns offer a promising approach to quickly reduce data dimensionality and enable discovery. Recent advances in neuroscience have enabled brain structure exploration at the level of individual synaptic connections. These graphs are often too large to be analyzed manually, presenting significant barriers to searching for structure and testing hypotheses. We combine graph database and analysis libraries with an easy-to-use neuroscience grammar suitable for rapidly constructing queries and searching for subgraphs and patterns of interest. This abstracts many of the computer science and graph theory challenges associated with nanoscale brain network analysis and allows scientists to quickly achieve reproducible findings at scale. We demonstrate these tools to search for motifs on simulated data and real public connectomics datasets, and share simple and complex structures relevant to the neuroscience community. We contextualize these results and provide case studies to motivate future neuroscience questions. All of our tools are released open source to empower other scientists to use and extend these methods.

show abstract

Statistical Connectomics

Cited by 9 publications

References 20 publications

Neural Fragility as an EEG Marker of the Seizure Onset Zone

Neural Fragility as an EEG Marker of the Seizure Onset Zone

When no answer is better than a wrong answer: a causal perspective on batch effects

DotMotif: An open-source tool for connectome subgraph isomorphism search and graph queries

Contact Info

Product

Resources

About