Mild cognitive impairment (MCI) represents the preliminary stage of Alzheimer's disease. Excavating effective electroencephalography (EEG) markers for the timely diagnosis of MCI remain an urgent scientific issue that is of great significance for preventing the deterioration of the disease. In this study, we proposed a novel method, multiscale dispersion recurrence plot (MDRP), to analyze the multiscale recursive information of signals. First, the Logistic model was used to generate stimulus signals. The performance of MDRP and multiscale order recurrence plot (MORP) algorithms were analyzed in data length, embedding dimension, signal complexity, signal-to-noise ratio, and scale factor. The results indicate the MDRP algorithm is less sensitive to the data length, embedding dimension, and noise, and can reflect the nonlinear characteristics of the chaotic model more accurately. Next, we implemented the MDRP algorithm to evaluate EEG data with normal cognition and mild cognitive impairment and calculated the determinism (DET) values on multiple time scales to explore the nonlinear dynamic characteristics of the EEG data. The results showed that the short-scale DET values of MCI patients were significantly lower than those of the control group. In addition, we performed the Pearson correlation analysis between cognitive scale score and EEG DET values. The results demonstrate the DET based on MDRP is an EEG marker associated with cognitive impairment. In conclusion, MDRP provides a new way to effectively characterize the nonlinear characteristics of MCI EEG signals at various scales.
Read full abstract