Friday, Jun 27: 3:45 PM - 5:00 PM
2383
Symposium
Brisbane Convention & Exhibition Centre
Room: Great Hall (Mezzanine Level) Doors 5, 6 & 7
The field of brain connectivity has seen rapid growth, but validation of connectivity measures remains an important challenge. This timing is crucial as connectivity research increasingly influences our understanding of brain function, disease biomarkers, and treatment efficacy. As both analysis methods and research studies continue to accumulate, understanding what these measures actually tell us about brain organization requires validation across multiple domains. Preclinical research provides powerful experimental control for validating biological mechanisms, while careful examination of preprocessing choices and statistical assumptions helps establish reliable results. Finally, clinical validation allows to establish validity and feasibility of brain connectivity measures for clinical translation. In this symposium, we will show how validation frameworks across biological, statistical and clinical dimensions help ground our interpretation of connectivity measures, establishing more reliable tools for both basic neuroscience and clinical applications. After a short introduction on the concept of validation by the organizers, Alan Jasanoff will tackle biological validation, showing how a novel genetically encoded activity reporter in animal models can inform the biological basis of fMRI brain connectivity. Second, Andrea Hildebrandt will show how a systematic characterization of the brain connectivity multiverse can provide valuable insights into assessment of statistical robustness across analytical choices as a prerequisite for enhancing replicability. Third, Zhen-Qi Liu will provide an overview of the properties of MR, MEG and PET brain connectivity measures, highlighting which metrics possess desirable statistical and biological properties. Last, Matej Perovnik will show a real-life example of systematic clinical validation based on PET brain networks.
Through the symposium any attendee should:
1) Understand how different validation approaches (statistical, biological, and clinical) strengthen the interpretation of brain connectivity measures
2) Learn to evaluate the impact of methodological choices on connectivity findings, from preprocessing to statistical inference
3) Gain perspective on how different research fields (including MR, PET and MEG research on animal and human participants) can inform each other's connectivity findings
This symposium is designed for neuroimaging researchers and clinicians who use or plan to use connectivity measures in their work, particularly those interested in establishing robust and biologically meaningful connectivity findings. The symposium aims to discuss fundamental and timely questions in the field of MR, EEG/MEG and PET brain connectivity, and is therefore of interest to the broadest brain mapping community, including both human neuroimaging researchers and preclinical experts.
Presentations
The brain’s function depends intimately on its connectivity. We have recently developed a series of molecular tools that enable structure-function relationships to be studied on a brain-wide scale in rodents and nonhuman primates. Here we will describe how a novel genetically encoded activity reporter enables information flow in neural projections to be monitored by fMRI. Using this reporter, we show how the characteristics of efferent population activity from brain regions differ depending on the projection target. Conversely, we show how regional brain activity can be described specifically in terms of synaptically defined input sources. In further studies, we use a second genetic tool for MRI to show how variations in the physical connectivity of the brain relate to functional connectivity in mice. We find that differences in physical and functional connectivity can be dissociated both over individual subjects and over experience, with positive or negative correlations observed in different contexts. These results indicate how molecular imaging techniques enable mechanistic analyses of input/output and structure-function relationships throughout extended brain networks.
Presenter
Alan Jasanoff, PhD, Massachusetts Institute of Technology Boston, MA
United States
Improving replication rates in graph-based brain connectivity analysis requires that researchers disclose their degrees of freedom by making explicit the arbitrary but equally defensible choices in design, data processing, and statistical analysis. This approach, known as multiverse analysis, is gaining ground in cognitive neuroscience. However, knowledge of the full range of analytical options - the "garden of forking paths" - is essential to this approach. Through literature mining and expertise crowdsourcing, we have created a dynamic knowledge space of analytical decisions that can be used interactively by researchers. The space contains 61 different steps, 17 of which have controversial parameter choices as options. We show how this space can be used for the design and application of multiverse analysis in graph-based brain connectivity research, including an active learning-based sampling approach of potential pipelines to approximate an exhaustive multiverse analysis. We will discuss the implications and the next steps to be taken.
FC is a statistical construct and its operational definition is largely arbitrary. Inter-regional functional interactions in the brain extend far beyond commonly used zero-lag Pearson's correlations. In this talk, we will present a comprehensive investigation on the topological and geometric organization, neurobiological associations, and cognitive-behavioral relevance using 239 pairwise interaction statistics, based on MR, MEG and PET imaging. We highlight the importance of tailoring a pairwise statistic to a specific neurophysiological mechanism and research question. https://www.biorxiv.org/content/10.1101/2024.05.07.593018v1
Presenter
Zhen-Qi Liu, Montreal Neurological Institute Montreal, Quebec
Canada
In this talk, we will present thirty year of research endeavors to identify, validate and assess metabolic brain networks at a single-subject level scaled subprofile model/principal component analysis (SSM/PCA) and discuss the added clinical value for early and accurate diagnosis and prognosis of patients with neurodegenerative disorders. We will showcase the validity and reliability of the model across different disorders, sites, and disease stages.