Brain Connectivity Bias Correction with Traveling Subjects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing brain activity analysis methods face challenges in accurately predicting neurological/mental disorders due to small sample sizes and site-to-site differences in brain image data collection, leading to over-fitting and poor generalization of discriminators across facilities.
Innovation Solution
A method and system using a generalized linear mixed model to adjust brain functional connectivity correlation values, correcting measurement biases across multiple facilities by calculating measurement bias as a fixed effect, and applying machine learning with feature selection to harmonize brain activity classifiers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If brain image data is collected from multiple facilities to increase sample size, then the statistical power and generalizability of the analysis is improved, but site-to-site differences and measurement biases cause over-fitting and poor generalization of discriminators
Solution Approach 1:
The patent introduces traveling subjects as an intermediary element to bridge multiple measurement facilities. These subjects are measured at all facilities, creating a common reference frame that enables the estimation and correction of site-specific measurement biases, thereby allowing data integration while maintaining generalization capability
Solution Approach 2:
The patent applies statistical transformations to the functional connectivity data, including Fisher z-transformation of correlation coefficients and application of linear mixed models with site-specific random effects. These parameter changes enable the separation of true biological signals from site-specific measurement artifacts
2Quantity of substance
If brain functional connectivity data from multiple facilities is used without correction, then more data is available for analysis, but measurement biases at each facility lead to inaccurate results
Solution Approach 1:
The patent employs a feedback mechanism where measurement data from traveling subjects is used to estimate site-specific bias parameters, which then feed back into the correction process for all subjects at each facility. This iterative correction improves measurement precision while preserving data availability
Solution Approach 2:
Traveling subjects serve as an intermediary reference group that enables the quantification and subsequent correction of facility-specific measurement biases, allowing accurate integration of data from multiple facilities
3Measurement precision
If discriminators are trained on data from a single facility to achieve high accuracy, then the model performance is optimized for that facility, but the discriminator cannot be generalized to other facilities
Solution Approach 1:
The patent creates a universal preprocessing and bias correction framework that can be applied across multiple facilities. The linear mixed model structure with site-specific random effects and Fisher z-transformation provides a multi-functional approach that maintains discriminator accuracy while enabling generalization to new facilities
Solution Approach 2:
By applying statistical transformations (Fisher z-transformation) and using linear mixed models with site-specific parameters, the patent standardizes the data representation across facilities, enabling the development of universal discriminators that maintain accuracy across different measurement environments
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables accurate and generalized brain activity analysis across multiple facilities, allowing for objective determination of healthy or diseased states and effective brain activity biomarkers for neurological/mental disorders.
Implementation Method 1
In Magnetic Resonance Imaging (MRI), changes appearing in detected signals in accordance with changes in the blood stream make it possible to visualize an active portion of a brain activated in response to an external stimulus
Implementation Method 2
brain activities are measured by measuring increase in the nuclear magnetic resonance signal (MRI signal) of hydrogen atoms representing a phenomenon where the deoxygenated hemoglobin level in red blood cells decrease in minute veins or capillary vessels (BOLD effect)
Implementation Method 3
Oxygenated hemoglobin is diamagnetic and does not have any influence on relaxation time of hydrogen atoms in the surrounding water. In contrast, deoxygenated hemoglobin is paramagnetic and changes surrounding magnetic field
Implementation Method 4
deoxygenated hemoglobin is paramagnetic and changes surrounding magnetic field
Data Source
AI summary
A harmonization system for a brain activity classifier harmonizing brain measurement data obtained at a plurality of sites to realize a discrimination process based on brain functional imaging: obtains data, for a plurality of traveling subjects as common objects of measurements at each of the plurality of measurement sites, resulting from measurements of brain activities of a predetermined plurality of brain regions of each of the traveling subjects; calculates, for each of the traveling subjects, prescribed elements of a brain functional connectivity matrix representing the temporal correlation of brain activities of a set of the plurality of brain regions; using a generalized linear mixed model, calculates measurement bias data 3108 for each element of the brain functional connectivity matrix, as a fixed effect at each measurement site with respect to an average of the corresponding element across the plurality of measurement sites and across the plurality of traveling subjects; and thereby executes a harmonizing process.