summary: Researchers have developed a novel cortical surface template, “OpenNeuro Average” (onavg), that improves the accuracy and efficiency of analyzing neuroimaging data.
The template is based on 1,031 brains and provides a more uniform and less biased map compared to previous models, which allows for more efficient use of data, crucial for studies with limited datasets.
The onavg template is expected to have broad applications in cognitive and clinical neuroscience.
Key Facts:
- Uniform Sampling: onavg samples brain regions uniformly, reducing bias.
- Data Efficiency: Less data is needed for accurate analysis, which is useful for studies with limited data sets.
- Wide range of applications: It is useful for studying vision, language and neurodegenerative diseases.
sauce: Dartmouth College
The human brain is responsible for vital functions such as perception, memory, language, thought, consciousness and emotions.
To understand how the brain works, scientists often use neuroimaging techniques that record brain activity in subjects while they perform tasks and while they are at rest. Brain functions are organized in the outer layer of the human brain, the cerebral cortex.
Researchers often use what are called “cortical surface models” to analyze neuroimaging data and study the functional organization of the human brain.
Every brain is shaped differently. To analyze neuroimaging data from multiple individuals, researchers need to register the data to the same brain template. This allows them to identify the same anatomical locations on different brains, even if the brains are shaped differently. These locations are known as “vertices.”
Over the past 25 years, such templates have been refined many times, but the cortical surface template most commonly used today is based on data collected from 40 brains.
Now, researchers at Dartmouth College have created a new cortical surface template, called “OpenNeuro Average,” or “onavg” for short, that can improve the accuracy and efficiency of analyzing neuroimaging data.
The results of the study are: Nature Method.
“Our cortical surface template, onavg, is the first to uniformly sample different parts of the brain,” says lead author Feilong Ma, a postdoctoral researcher and member of the Haxby lab in Dartmouth’s Department of Psychological and Brain Sciences. “It’s a less biased and more computationally efficient map.”
The team built the template based on the cortical anatomy of 1,031 brains from 30 datasets on OpenNeuro, a free, open-source platform for sharing neuroimaging data. According to the co-authors, this is also the first cortical surface template based on the geometry of the brain.
In contrast, previous templates non-uniformly sampled different parts of the cortex and defined the locations of cortical vertices based on spherical shapes, resulting in biased distribution of vertices.
Using the onavg template reduces the data required for analysis.
“Neuroimaging data can be very expensive to obtain, and in some clinical populations, such as rare disease studies, it can be difficult or impossible to obtain large amounts of data, so the ability to get better results with less data is an asset,” Feiron says.
“Through more efficient data use, our template has the potential to increase the reproducibility and replicability of results in academic research.”
“We see ONAVG as a methodological advancement that has broad applications across all aspects of cognitive and clinical neuroscience,” says co-author James Haxby, a professor in Dartmouth’s Department of Psychological and Brain Sciences and former director of the Center for Cognitive Neuroscience.
He says their cortical surface templates could potentially be used to study vision, hearing, language and individual differences, as well as disorders such as autism and neurodegenerative diseases such as Alzheimer’s and Parkinson’s.
“We believe this work will have broad and profound impacts on the field,” Haxby said. Jiahui Guo, a former postdoctoral researcher in psychology and brain sciences and assistant professor in the Department of Behavioral and Brain Sciences at the University of Texas at Dallas, and Maria Ida Gobbini, associate professor in the Department of Medicine and Surgery at the University of Bologna, also contributed to the study.
About this brain mapping research news
author: Amy Olson
sauce: Dartmouth College
contact: Amy Olson – Dartmouth College
image: Image courtesy of Neuroscience News
Original Research: Open access.
“Cortical Surface Templates for Human Neuroscience” (Feilong Ma et al.) Nature Method
Abstract
Cortical Surface Templates for Human Neuroscience
Analysis of neuroimaging data relies on normalization to standard anatomical templates to resolve macro-anatomical variations across the brain. Existing human cortical surface templates sample locations non-uniformly due to distortions introduced by expanding the collapsed cortex to a standard shape.
Here we introduce the onavg template, which allows for uniform sampling of the cortex.
We created the onavg template based on 1,031 publicly available high-quality structural scans of brains, 25 times more than any existing cortical template. We optimized vertex locations based on cortical anatomy to ensure even distribution.
Between-participant correlations of multivariate pattern classification accuracy and representation geometry based on onavg were consistently higher than those based on other templates, and onavg required only three-quarters less data to achieve the same performance compared to other templates.
The optimized sampling reduces the variation in the number of vertices in each searchlight, and therefore reduces the CPU time of the entire algorithm by 1.3-22.4%.
