Brain Fingerprinting

Brain fingerprinting aims to identify unique, individual-specific patterns of brain activity, often using functional connectivity (FC) derived from fMRI data. In this study, we employed a convolutional autoencoder to capture shared features in functional connectomes, creating a residual FC matrix that isolates subject-specific variations by subtracting common patterns likely related to task or network effects. This residual matrix, containing unique individual features, was further analyzed with sparse dictionary learning, allowing us to extract distinct connectivity patterns that enhance fingerprinting accuracy across subjects.