Synchronized feature patterns of local field potentials in hippocampal-prefrontal cortex during working memory based on time-varying spectral coherence / 国际生物医学工程杂志

ABSTRACT

Objective:To investigate the synchronized feature patterns of local field potentials in the hippocampus (HPC) and prefrontal cortex (PFC) during working memory based on time-varying spectral coherence so as to support the study of information processing mechanisms in working memory.Methods:The local field potentials (LFPs) signals of the ventral hippocampus (vHPC) and medial prefrontal cortex (mPFC) were collected from six SD rats during the performance of a spatial working memory task in the Y-maze, and the time-frequency distributions of vHPC and mPFC LFPs were calculated by applying the short-time Fourier transform (STFT) to determine the characteristic frequency bands of the working memory and then to investigate the synchronized patterns of vHPC and mPFC LFPs based on the coherent of the time-varying frequency spectrum. Finally, support vector machines were applied to explore the feasibility of applying spectral coherence values to predict working memory.Results:When rats performed working memory tasks correctly, the energy of the theta band (4 - 12 Hz) of the HPC and PFC increased (all P < 0.01), and the spectral coherence value of the theta band of the HPC-PFC increased ( P < 0.05). Support vector machine training and prediction using the average peak spectral coherence and the difference between the peak and the onset when correctly and incorrectly executing the working memory as features resulted in 89% accuracy, 90% precision, 88% recall, and 88% F1 scores, all of which were statistically significant differences compared to the results of the randomly disrupted labeled data rearranging (all P < 0.05). Conclusions:Synchronized synergy in the HPC-PFC theta band is one of the potential mechanisms for correctly performing information processing in working memory.