Saturday, Jul 22: 8:00 AM - 5:00 PM
Educational Course - Full Day (8 hours)
The brain does not exist in isolation. The field of neuroscience is increasingly interested in the intricate relationships between brain function and diverse concurrent physiological processes throughout the body. For many neuroscientists, this is not a new concern - strategies for removing physiological artifacts (e.g., cardiac, respiratory, myogenic) from neuroimaging data have been explored for decades. However, there is a resurgence of interest in physiologic signals, particularly in conjunction with fMRI data. Improvements in data quality facilitate the isolation and quantification of these physiologic signals, resulting in promising new biomarkers for brain health, physiology and function that can complement our understanding of neural activity, as well as in improved fMRI denoising. With advanced monitoring techniques, the diversity of physiological processes that can be characterized during neuroimaging experiments is expanding.
This course introduces the physiologic signals that influence typical fMRI datasets. We present methods for recording physiological processes during scanning and modeling the associated signal variance, to either study or remove these effects. Alternatively, data-driven methods for achieving similar outcomes will also be presented, which can bring new insight into existing data resources that lack physiological recordings. In addition, we will introduce cutting-edge techniques to monitor electrogastrography and explore how respiratory and cardiac information can provide further insights into cognition. To complement theoretical content, we will host hands-on sessions to show how to collect physiological data and incorporate them into neuroimaging analysis pipelines, either as confounding factors or signals of interest using available data and software resources.
- Learn how to set up physiological data collection in a neuroimaging setting.
- Learn how to utilize neuroimaging and physiological data to study brain physiology, and interactions of the brain with peripheral physiological processes..
- Learn how to model and remove physiological fluctuations from functional timeseries, with or without physiological recordings.
Researchers and clinicians working with functional neuroimaging (fMRI, M/EEG), from cognitive to clinical research, interested in knowing how to improve data denoising and how to add physiological-related data to their research repertoire.
Researchers and clinicians interested in exploring the use of BOLD fMRI to investigate cerebrovascular health, both with and without external physiological recordings.
Although fMRI is a powerful tool for studying neural activity, the signals we measure reflect a diverse range of physiologic processes that intrinsically couple the brain to the body. We do not scan a brain in isolation from the human being that hosts it, and it is nearly impossible to “turn off” a beating heart or breathing lungs when we probe brain function. In this educational session, I will discuss the physics and physiology of how cardiac and respiratory factors influence the fMRI timeseries, and demonstrate best practices for recording and interpreting physiologic signals during an fMRI scan. I will also introduce less well-established physiologic signals that can also be explored in fMRI, to better understand arousal, blood pressure, gut-brain interactions, hormonal fluctuations, and how all these aspects of human physiology may interact with our understanding of brain function when using neuroimaging.
This tutorial will provide attendees with a firsthand opportunity to learn how to collect good-quality physiological measurements and to prepare physiological data for use in neuroimaging workflows. Beginning with a live demonstration of photoplethysmography (PPG), respiratory belt, and skin conductance recordings, we will discuss best practices for equipment set-up and quality control during functional neuroimaging scans, with particular emphasis on techniques appropriate for the MRI scanner environment. In the course of this real-time physiological data acquisition, attendees will learn to assess and improve the quality of their measurements and to identify the presence of common artifacts in their data. After becoming familiar with the key features of physiological measurements, we will discuss how to integrate peripheral recordings with subsequent neuroimaging analyses. In an interactive coding demonstration, attendees will practice preparing their physiological data for sharing and processing using phys2bids, currently the most complete open-source toolbox for standardizing physiological data formats in python.
, McGill University Montreal, Quebec
As an indirect measure of neural activity, fMRI is inevitably sensitive to blood flow
and oxygenation changes regulated by the autonomic nervous system, including both faster-scale oscillations time-locked to the respiratory/cardiac cycles and slower-scale variations (< 0.1 Hz) that overlap with the spectrum of intrinsic brain activity. Isolating local functional dynamics from these confounding physiological factors is key to improving the sensitivity and neuronal specificity of fMRI for broad cognitive and clinical applications. In this educational talk, I will provide an overview of existing approaches that remove various physiological artifacts from fMRI time series, based on external physiological recordings. I will also discuss potential challenges in de-noising fMRI data when autonomic regulation tracks global neuronal changes in specific physiologic and cognitive states.
, Massachusetts General Hospital and Harvard Medical School Cambridge, MA
Modeling physiological effects in fMRI data can increase the precision of fMRI for measuring local neural activity, as well as open new avenues for studying brain physiology and brain-body interactions in health and disease. However, it is not always feasible to record physiological signals during fMRI, and many existing open-source datasets do not have concurrent physiological recordings. Even when these signals are acquired, they may suffer from poor quality or missing data. This talk will focus on techniques for extracting systemic physiological information directly from fMRI data. We will also highlight implications for understanding individualized brain function and physiology.
, Vanderbilt University Nashville, TN
The most widespread use of functional BOLD MRI is to detect neural activity. However, such detection is only indirect, as it measures changes in blood driven by neurovascular coupling. For this reason, other physiological fluctuation can act as strong confounding factors. Complementing the theory on physiological denoising from previous talks, in this hands-on session we will show how to take into account these noise sources to be more confident in the cognitive interpretation of neuroimaging data. We will help attendees familiarize themselves with toolboxes to (1) process physiological signals and detect peaks, (2) prepare physiological fluctuation models to denoise functional BOLD MRI during preprocessing and data analysis, (3) model physiology-related noise fluctuations from fMRI. Attendees will be able to perform the tasks alongside the presenter, using a provided jupyter notebook or a container, as well as data for a demo.
, École polytechnique fédérale de Lausanne (EPFL)
In this educational talk, I will discuss the use of fMRI and simultaneously acquired physiological signals (related to cardiac and respiratory processes) to probe cerebrovascular health. I will outline the physiological basis of fMRI signals and their relationship to systemic physiology, explaining how they can be used to measure cerebrovascular function. Practicalities of measuring physiological signals in the MR environment and common difficulties will be discussed. I will summarise the advantages of measuring cerebrovascular function to complement other functional imaging measures.
, Cardiff University Brain Research Imaging Centre Cardiff, Wales
Physiological components in the BOLD fMRI signals are often thought to be nuisance signals and are discarded in the preprocessing of BOLD data. However, useful physiological parameters can be extracted from the BOLD fMRI signals. Such physiological parameters can not only help to improve the interpretation of BOLD fMRI results, but also be useful biomarkers by themselves. In this educational talk, I will describe how to obtain parametric maps of physiological parameters using typical task and resting-state fMRI data without external recordings. I will also discuss the utilization of these physiological parameters in clinical applications.
, University of Maryland Baltimore, MD
Whether you are specifically interested in designing experiments to measure cerebrovascular reactivity or you would like to learn how to obtain this important metric of brain health “for free” from your fMRI data, this tutorial will have something for you! From preparing physiological data for the modeling algorithm, to interpreting the resulting outputs, attendees will have the opportunity to learn what cerebrovascular reactivity mapping is all about. In addition, attendees will have the option to follow-along with the demonstration via shared data and coding environment. Attendees will learn how to prepare an external respiratory physiology recording for use as the “reference signal” in a cerebrovascular reactivity model. Then, we will demonstrate how to use this signal with phys2cvr, a python-based open-source toolbox for mapping cerebrovascular reactivity. Attendees will also learn how to obtain similar maps even without an external recording. Lastly, we will discuss the outputs from the modeling algorithm and recommended practices for visualizing and interpreting the cerebrovascular reactivity results.
, Northwestern University
The stomach continuously produces a 0.05 Hz electrical oscillation (i.e. 1 cycle every 20 seconds), which serves to pace the stomach contractions necessary for digestion. This so-called gastric rhythm can be measured non-invasively by placing electrodes over the abdomen, a technique known as the electrogastrogram. This electrophysiological technique can be safely incorporated during fMRI recordings by using MRI compatible electrodes and amplifiers of the same type used to measure electrocardiography or electromyography. While common in the field of gastroenterology, electrogastrography has yet to gain wider adoption in the field of cognitive neuroscience. The objective of this educational talk is thus to introduce the electrogastrography technique, as well as its potential applications for cognitive neuroscience. During the talk, I will review the physiological basis of the gastric rhythm, and provide an overview of brain-stomach anatomical pathways. I will then present the standard practices to acquire, analyze and interpret the electrogastrogram inside the fMRI scanner.
, German Institute of Human Nutrition Potsdam-Rehbrücke (DIfE) Nuthetal, Brandenburg
Physiological signals from the body (e.g., cardiac activity, respiration, or skin conductance) are tremendously rich in information in their own right, which can be used as an alternative and additional source of evidence to support the involvement of specific cognitive mechanisms and neural pathways. For instance, cardiac activity recorded during MRI scanning can be used to derive heart rate variability or cardiac phase information, both of which have interesting neuropsychological and behavioral correlates. Skin conductance can also be used as a proxy of emotion and saliency detection mechanisms. In this educational talk, we will see why and how to easily analyze such signals using free open-source software to boost the reliability and strength of your conclusions at a low cost.