Altered pattern analysis and identification of subjective cognitive decline based on morphological brain network

Area of Science:

• Neuroimaging
• Neurology
• Biophysics

Background:

• Subjective cognitive decline (SCD) represents the earliest clinical stage of Alzheimer's disease (AD), making its accurate diagnosis crucial for timely intervention.
• Understanding the specific alterations in brain network morphology in SCD is vital for developing targeted prevention strategies.
• Current knowledge regarding the precise morphological network patterns in SCD individuals remains limited.

Purpose of the Study:

• To identify and characterize altered morphological network patterns in individuals with SCD.
• To develop a method for differentiating SCD from normal controls (NCs) using brain connectome features.
• To gain insights into the neuroimaging mechanisms underlying SCD.

Main Methods:

• Utilized the Jensen-Shannon distance-based similarity (JSS) method to construct individual morphological brain networks from 36 SCD cases and 34 NCs.
• Applied t-tests for discriminating nodal graph metrics and leave-one-out cross-validation (LOOCV) for consensus connections.
• Employed a multiple kernel support vector machine (MK-SVM) to integrate multiple brain connectome features for classification.

Main Results:

• Discriminative connectome features, including consensus connections and nodal graph metrics, were predominantly located in the frontal, limbic, and parietal lobes, associated with the default mode network (DMN) and frontoparietal task control (FTC) network.
• SCD cases exhibited a trend towards enhanced modularity and local efficiency in their brain networks.
• The MK-SVM model achieved high classification performance (AUC: 0.9510, accuracy: 91.43%) in distinguishing SCD from NCs.

Conclusions:

• The combination of multiple connectome attributes derived from morphological brain networks provides a robust method for differentiating individuals with SCD from normal controls.
• The identified altered patterns in multidimensional connectome attributes offer valuable insights into the neuroimaging mechanisms of SCD.
• This approach holds promise for early detection and intervention strategies in individuals experiencing subjective cognitive decline.