AUTHOR=Chen Chao , Meng Jiayuan , Belkacem Abdelkader Nasreddine , Lu Lin , Liu Fengyue , Yi Weibo , Li Penghai , Liang Jun , Huang Zhaoyang , Ming Dong 

TITLE=Hierarchical fusion detection algorithm for sleep spindle detection

JOURNAL=Frontiers in Neuroscience

VOLUME=Volume 17 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2023.1105696

DOI=10.3389/fnins.2023.1105696

ISSN=1662-453X

ABSTRACT=The spindle is a vital sign implying that human beings enter the second stage of sleep. It can effectively reflect a person's learning and memory ability, and clinical research shows that its quantity and density are crucial markers of brain function. The ‘gold standard’ of spindle detection is based on expert experience, the detection cost is high, and the detection time is longer. The accuracy of detection is influenced by subjectivity.
Method：
To improve the detection accuracy and speed, reduce the cost and improve the efficiency, this paper proposes a layered spindle detection algorithm. The first layer uses the Morlet wavelet and RMS method to detect spindles, and the second layer employs an improved k-means algorithm to improve spindle detection efficiency. The fusion algorithm was compared with other spindle detection algorithms to prove its effectiveness.
Results：
The hierarchical fusion spindle detection algorithm shows better performance stability, and the fluctuation range of detection accuracy is minimal. The average value of Precision is 91.6%, at least five percentage points higher than other methods. The average value of Recall can reach 89.1%, and the average value of Specificity is close to 95%. Compared with other methods, it has reached a peak. The performance on accuracy and F1-score is also better than the other three algorithms. The mean values in the subject sample data are 90.4% and 90.3%, respectively, as close as the 100% index of the ideal state.
Conclusion：
A spindle detection method with high steady-state accuracy and fast detection speed is proposed by combining the Morlet wavelet with window RMS and merging with the improved k-means algorithm. This method provides a powerful tool for spindle automatic detection technology and improves the efficiency of spindle detection. Through simulation experiments, the sampled data are analysed and verified to prove the feasibility and effectiveness of this method.