A CCA and ICA-Based Mixture Model for Identifying Major Depression Disorder

Area of Science:

• Neuroimaging
• Biomedical Engineering
• Data Science

Background:

• fMRI signal processing typically involves filtering, which can lead to information loss in low-frequency fluctuations.
• Classical methods like ICA and CCA have limitations: ICA components may lack physiological interpretability, and CCA components can be contaminated by noise.

Purpose of the Study:

• To develop a novel mixed model integrating ICA and CCA for fMRI analysis.
• To avoid data filtering, thereby preserving high-frequency information in low-frequency fluctuations.
• To achieve statistically independent and physiologically interpretable components.

Main Methods:

• A mixed model combining Independent Component Analysis (ICA) and Canonical Correlation Analysis (CCA) was proposed.
• The model operates without the need for data filtering, preserving signal integrity.
• The approach was validated using synthetic data and applied to a real-world fMRI dataset.

Main Results:

• The proposed mixed model yields statistically independent components.
• The method successfully preserves useful information within low-frequency fluctuations.
• Experiments demonstrated advantages over classical ICA and CCA methods.
• Application to discriminating major depression from controls showed encouraging results.

Conclusions:

• The novel mixed ICA-CCA model offers an effective alternative to traditional fMRI analysis techniques.
• This approach enhances the preservation of crucial signal information and improves component interpretability.
• The model shows promise for clinical applications, such as diagnosing major depressive disorder.