Ordinal Pattern: A New Descriptor for Brain Connectivity Networks.

Area of Science:

• Neuroscience
• Network Science
• Machine Learning

Background:

• Brain connectivity networks from MRI/fMRI are crucial for understanding brain structure and function.
• Existing network descriptors often overlook valuable weight information or the ordinal relationships of weighted edges.
• There is a need for advanced methods to analyze complex brain connectivity patterns for disease diagnosis.

Purpose of the Study:

• To propose a novel network descriptor, ordinal patterns, for brain connectivity network analysis.
• To develop an ordinal pattern-based learning framework for automated brain disease diagnosis using resting-state fMRI data.
• To evaluate the proposed method's performance against state-of-the-art approaches.

Main Methods:

• Construction of brain functional connectivity networks from resting-state fMRI data.
• Development of an algorithm to identify frequently occurring ordinal patterns in patient and control groups.
• Feature extraction based on selected discriminative ordinal patterns for a learning model.
• Application of the framework to Alzheimer's Disease Neuroimaging Initiative and ADHD-200 datasets.

Main Results:

• The proposed ordinal patterns effectively capture weight information and ordinal relationships of weighted edges.
• The ordinal pattern-based learning framework achieved superior performance in disease classification and clinical score regression.
• The method demonstrated significant improvements over several state-of-the-art approaches on benchmark datasets.

Conclusions:

• Ordinal patterns offer a powerful new descriptor for brain connectivity networks, surpassing traditional methods.
• The developed framework provides an effective tool for automated brain disease diagnosis using fMRI data.
• This approach holds promise for advancing neuroimaging-based diagnostics and understanding brain disorders.