Pattern classification using principal components of cortical thickness and its discriminative pattern in...

Area of Science:

• Neuroscience
• Medical Imaging
• Computational Biology

Background:

• Schizophrenia is a complex psychiatric disorder with known neuroanatomical alterations.
• Cortical thickness is a key neuroimaging biomarker affected in schizophrenia.
• Accurate classification of schizophrenia using neuroimaging data remains a challenge.

Purpose of the Study:

• To develop and validate a pattern classification method for schizophrenia using cortical thickness data.
• To identify discriminative patterns of brain abnormalities associated with schizophrenia.
• To assess the reliability of cortical thickness as a feature for diagnostic tools.

Main Methods:

• Principal component analysis (PCA) was applied to cortical thickness data from schizophrenic patients and healthy controls.
• Leave-one-out cross-validation and simulated validation sets were used for classification accuracy assessment.
• Support vector machines (SVM) were employed to analyze discriminative patterns in the feature space.

Main Results:

• Classification accuracy varied by lobe, with temporal lobe showing high accuracy (e.g., 93.60% left).
• The study identified specific brain regions with discriminative power, including precentral, postcentral, and cingulate gyri.
• Optimal classification was achieved using 40-70 principal components, ranked by effectiveness.

Conclusions:

• Cortical thickness is a reliable neuroimaging feature for pattern classification in schizophrenia.
• The proposed method demonstrates potential for developing objective diagnostic tools for schizophrenia.
• Findings are consistent with previous morphological studies of schizophrenia, enhancing confidence in the results.