CN-118484634-B - M wave extraction method based on self-adaptive threshold

Abstract

The invention provides an M wave extraction method based on a self-adaptive threshold, which comprises the steps of S1, calculating the mean value, standard deviation and quantile of an electromyographic signal of an M wave to be extracted, self-adaptively setting a stimulation period amplitude threshold, obtaining a pulse sequence and a pulse width sequence according to the threshold and a signal gradient, denoising, dividing the stimulation period and a non-stimulation period according to the distance between pulses, S2, self-adaptively setting the pulse amplitude threshold based on the maximum value and the minimum value of the stimulation period signal, positioning a stimulation pulse waveform, adopting a FastICA method to carry out electric stimulation artifact filtering, S3, determining the starting point and the ending point of the M wave according to the positioned stimulation pulse waveform and the pulse width sequence, and extracting the M wave. The invention realizes self-adaptive setting of a series of thresholds, completes M-wave signal extraction, and can reduce the labor burden and improve the working efficiency in the occasion of real-time or large-scale processing of functional electric stimulation data in scientific research or clinic.

Inventors

• CHEN XIANG
• LIN PENGHUI
• ZHANG XU

Assignees

• 中国科学技术大学

Dates

Publication Date

20260512

Application Date

20240624

Claims (2)

1. 1. An M-wave extraction method based on an adaptive threshold, the method comprising the steps of: Step S1, adaptively setting a stimulation period amplitude threshold value based on the mean value, standard deviation and quantile of an electromyographic signal of M waves to be extracted, obtaining a pulse coordinate sequence based on the threshold value and a signal gradient, calculating a pulse width sequence based on the pulse coordinate sequence, denoising the pulse width sequence, and dividing a stimulation period and a non-stimulation period according to the inter-pulse distance; step S2, adaptively setting a pulse amplitude threshold based on the maximum value and the minimum value of the stimulation period signal, positioning the stimulation pulse waveform of the stimulation period signal, filtering the stimulation period signal and removing the electrical stimulation artifact; Step S3, determining a starting point and an ending point of the M wave in the stimulation period signal after removing the electric stimulation artifact according to the highest point coordinate sequence and the lowest point coordinate sequence of the stimulation pulse waveform of the positioned stimulation period signal and the denoised pulse width sequence, so as to extract the M wave; wherein, the step S1 includes: Step 1.1, calculating the mean value, standard deviation, quantile and signal gradient of an electromyographic signal of M wave to be extracted, adaptively setting a stimulation period amplitude threshold value based on the parameters, determining a stimulation point coordinate and obtaining a starting point and ending point coordinate sequence of each pulse; Step 1.2, adaptively setting a pulse width threshold value, and removing burst pulse noise with shorter duration time; step 1.3, adaptively setting a pulse distance threshold value to realize the segmentation of the stimulation period and the non-stimulation period.
2. 2. The method of M-wave extraction based on adaptive threshold according to claim 1, wherein said step S2 comprises: Step 2.1, adaptively setting a pulse amplitude threshold value, and determining a coordinate sequence of the highest point and the lowest point of a stimulation pulse waveform in a stimulation period; And 2.2, removing the electric stimulation pulse artifact in the stimulation period signal through a FastICA algorithm.

Description

M wave extraction method based on self-adaptive threshold Technical Field The invention belongs to the field of biomedical signal processing, and particularly relates to an M-wave extraction method based on a self-adaptive threshold value. Background The functional electrical stimulation (functional electrical stimulation, FES) belongs to the category of neuromuscular electrical stimulation (neuromuscular electrical stimulation, NES), and is characterized in that low-frequency (1-100 Hz) pulse current with certain intensity is utilized to stimulate one or more groups of muscles through a preset program to induce muscle movement or simulate normal autonomous movement so as to achieve the purpose of improving or recovering the functions of the stimulated muscles or muscle groups. The electromyographic signal is a bioelectric signal which is recorded by the electrode when nerve excitation impulse is transmitted to the muscle to cause the muscle to shrink, carries nerve-muscle control information and can reflect the activation condition of the muscle to a certain extent. M wave is a special electromyographic signal generated by directly inducing muscle contraction through an electric stimulation motor nerve, has specific morphology and characteristics, can be used for measuring the response of stimulated muscle to electric stimulation, the local fatigue degree, the control condition of a nervous system on the muscle and the like, and is one of the most common bases for researching the excitatory change of a myomembrane. The M wave unfolding analysis on the functional electric stimulation is an important means for researching the effect of the functional electric stimulation. The current vast majority of M wave research related works mainly adopt a fixed threshold segmentation method to extract stimulation period signals according to the characteristic that the electric stimulation signals and the common electromyographic signals are different in amplitude, then adopt a fixed threshold segmentation method to stimulate the period signals according to the characteristic that the electric stimulation signals and the M waves are different in amplitude in the stimulation period, or adopt an energy window method to segment according to the characteristic that the stimulation signals and the M waves are different in energy. However, due to differences in electro-stimulated muscle position, stimulation frequency, signal noise conditions, individual conditions, and effects of electro-stimulation artifacts, extracting M-waves often requires manual adjustment of the threshold size, determination of the length of the extraction interval, and the start and end points of each extraction of the M-wave interval, etc. In clinical analysis, more experienced physician manual screening is required. Disclosure of Invention In order to overcome the limitations of the existing M-wave extraction method, the invention provides an M-wave extraction method based on an adaptive threshold. Unlike M-wave extraction methods based on fixed thresholds, which require frequent changes in threshold to remove electro-stimulation artifacts and extract M-waves when faced with different situations, the present invention can adaptively extract M-waves in different subjects, different stimulation frequencies, different stimulation current intensities, and in the presence of bursty impulse noise, etc. The M-wave extraction algorithm provided by the invention can remove the influence of the electric stimulation pulse artifact, reduce the time required for manually adjusting the threshold value, determining the parameters such as the extraction interval, the M-wave starting point and the ending point and the like in the fixed threshold value algorithm, and reduce the operation burden of the algorithm. In order to achieve the above purpose, the invention adopts the following technical scheme: An M-wave extraction method based on an adaptive threshold, the method comprising the steps of: Step S1, adaptively setting a stimulation period amplitude threshold value based on the mean value, standard deviation and quantile of an electromyographic signal of M waves to be extracted, obtaining a pulse coordinate sequence based on the threshold value and a signal gradient, calculating a pulse width sequence based on the pulse coordinate sequence, denoising the pulse width sequence, and dividing a stimulation period and a non-stimulation period according to the inter-pulse distance; step S2, adaptively setting a pulse amplitude threshold based on the maximum value and the minimum value of the stimulation period signal, positioning the stimulation pulse waveform of the stimulation period signal, filtering the stimulation period signal and removing the electrical stimulation artifact; And step S3, determining a starting point and an ending point of the M wave according to the positioned stimulation pulse waveform after removing the electrical stimulation artifact and the