Deal all,
I have sEMG data for triceps push-down exercise. the data was recorded throughout the duration of cycle. I want to separate the concentric and eccentric phases, and the condition is I don't have any kinematics information (no video recording or any other biomechanic information).
During our pilot testing, i recorded few videos and checked for phase (eccentric or concentric) detection. I found that the there is clear indication of transition (after zooming in one cycle on 22 inch computer screen) from concentric to eccentric phase. Further, from RMS plot, transition from concentric to eccentric phase is clearly visible from steep drop in the RMS value. attached figures show the scenario.
Both the video data and the visual segmentation nearly always matched (with difference of 15 to 30ms) for our pilot testing. So, I decided not to record the video as it increase the complexity. Is it a coincidence that in those pilot cases, the change in phase and video record happen to be at the same time? Can the concentric and eccentric phases be separated like this i.e. without any information on kinematics?
The eccentric phase of a lift occurs when a muscle contracts while lengthening, and the concentric phase of a lift occurs when a muscle contracts and shortens .
To detect a change in frequency between concentric and eccentric phase without information on kinematics, I suggest you to try a median and mean frequency analysis in sEMG as explain in Chapter 6, section 6.3 of my book Applied Biomechatronics Using Mathematical model (more info in link: www.garzaulloa.org) , free download dataset with Matlab example in link: https://data.mendeley.com/datasets/9j6sd5d8x9/1 ). Nevertheless, the real meaning of the phases between eccentric/concentric is with kinematics.
I say that you are wasting your time if you do not have kinematic data. There will be too much overlap between the observable range of median/mean sEMG frequency values for eccentric and concentric contractions.
Gordon L Warren thanks Sir for the insight.
I am not relating frequency or any other parameter. Well, there is a clear overlap between the two phases in both temporal and spectral features, and I am aware of it. The point is, while transition between concentric and eccentric i.e. there is a small (very little) pause. The point I am referring to is full elbow extension. Apparently, that point is detectable from the visual analysis of raw EMG data.
While moving from full extension to the neutral position, there is (nearly) always a gap of around 100 to 200ms. Further, while at full extension, the sEMG amplitude is highest and there is a sudden drop in that.
Are these arguments valid?
Jorge Garza-Ulloa Thanks Sir.
the mean and median frequency analysis is part of my analysis.
What I observed is, though the two phases show different frequencies in one cycle, the range of overall frequencies is largely overlapping. So can't rely on these parameters as they are questionable.
Depending on your signal processing and statistical knowledge, this paper may help. There are pattern classification techniques that can accurately differentiate between concentric and eccentric phases.
Article Paired pattern classification of electromyographic intensity...
- Badges
- Science topic