Static hand posture classification based on the biceps brachii muscle synergy features.

With modern upper limb myoelectric prostheses, many movements can be produced when many control signals are available. Some of them could originate from the multifunctional biceps brachii which appears to be composed of up to six individually innervated compartments. Given its compartmental nature, this muscle behaves as if composed of many individual muscles acting in synergy to achieve a given motor task. Through an appropriate synergy model, its EMG signals could thus be used to produce some of the many control signals required with modern upper limb myoelectric prostheses. Exploring that possibility, muscular synergy which is usually applied to a group of different muscles, was tested on the biceps brachii only. A non-negative matrix factorization method was applied on pre-recorded data consisting of 8 surface electromyographic signals collected across the biceps of 10 normal subjects who, in Seat or Stand posture, held their hand in 3 different postures. We found that muscular synergies can be extracted from the biceps brachii. With a learning process and a classifier, it was also possible, between a pair of static hand postures, to identify which one was used when a given record was made. The mean score of correctly detected hand posture was >80% for our subjects.