Differences of interference myogram patterns from forearm muscles at different variants of the hand grip

Background. Most modern bioprostheses use two sensors located on the flexor and extensor muscles of the hand and fingers to register muscle activity, which severely limits the functionality of the device and complicates switching between grips. One of the solutions to this problem is to use an array of more sensors to recognize different grips using a neural network.

Aim. To evaluate the possibility in principle of using an array of 8 sensors for recognizing different grips.

Materials and methods. Research was conducted on 23 healthy volunteers, whose surface myograms were recorded at 13 different grips. An array of surface-type electroneuromyographic sensors (n = 8) with bipolar electrodes and electroneuromyographic signal amplification factor of 2000 times, myograph and software are of own development were used as a technical base for the research.

Results. For most subjects, the differences in the investigated parameters when comparing grips with each other reached thousands of percent, especially when comparing the product of frequency and mean amplitude. In some pairs the differences were less significant and amounted to less than 400 %. The proportion of pairs with reliable differences varies from subject to subject when comparing the product of frequency and mean amplitude and ranges from 71 to 98 %. The average value for the whole group is 87 %. When comparing frequency only, the variation ranges from 67 to 93 %, with an average of 78 %.

Conclusion. For most subjects, the majority of grips are confidently differentiated from each other. Due to pronounced individual peculiarities, the decision to use this or that parameter to control the bioprosthesis for a definite person is individual and should be made after a thorough neurophysiological research. Some people will require training in order to develop a new motor stereotype.

1. Plotnikov A.A. Control algorithm of a prosthetic human hand. Matritsa nauchnogo poznaniya = Matrix of Scientific Knowledge 2019;(6):87–4. (In Russ.).

2. Bezyazychnyy V.F., Yeliseichev E.A., Vorobyev P.S. et al. Review of ways of reading EMG signals in the forearm area for controlling of bionic upper limb prostheses. Biomeditsinskaya radioelektronika = Biomedicine Radioengineering 2023;26(1):35–10. (In Russ.). DOI: 10.18127/j15604136-202301-04

3. Gorokhova N.M., Golovin M.S., Chezhin M.A. Methods for controlling upper limb prostheses. Nauchno-tekhnicheskiy vestnik informatsionnykh tekhnologiy, mekhaniki i optiki = Scientific and Technical Journal of Information Technologies, Mechanics and Optics 2019;19(2):314–2. (In Russ.). DOI: 10.17586/2226-1494-2019-19-2-314-325

4. Budko R.Yu., Chernov N.N., Budko A.Yu. et al. Forecession electromyography recognition and gestures selection for protesis control. Modelirovanie, optimizatsiya i informatsionnye tekhnologii = Modelling, Optimization and Information Technology 2019;24(1):17, 18. (In Russ.). DOI: 10.26102/2310-6018/2019.24.1.017

5. Phinyomark A., Phukpattaranont P., Limsakul C. Investigating long-term effects of feature extraction methods for continuous EMG pattern classification. Fluctuation Noise Lett 2012;11(4):1250028. DOI: 10.1142/S0219477512500289

6. Branchukova D.A., Peregudova O.O., Killer A.I. et al. Neural network management of bionic prosthesis. Vestnik molodyozhnoy nauki Rossii = Bulletin of Youth Science of Russia 2020;(1):15. (In Russ.).

7. Bhagwat S., Mukherji P. Electromyogram (EMG) based fingers movement recognition using sparse filtering of wavelet packet coefficients. Sādhanā 2020;45(3):11. DOI: 10.1007/s12046-019-1231-9