Independent component analysis algorithm, a recently developed multivariate statistical data analysis technique, has been successfully used for signal extraction in the field of biomedical and statistical signal processing. This paper reviews the concept of ICA and demonstrates its usefulness and limitations in the context of surface electromyogram related to hand movements and facial muscles. In the first experiment, ICA has been used to separate the electrical activity from different hand gestures. The second part of our study considers separating electrical activity from facial muscles. In both instances, surface electromyogram has been used as an indicator of muscle activity. The theoretical analysis and experimental results demonstrate that ICA is suitable for the identification of different hand gestures using sEMG signals. The results identify the unsuitability of ICA when the similar techniques are used for the facial muscles in order to perform different vowel classification. This technique could be used as a prerequisite tool to measure the reliability of sEMG based systems in rehabilitations and human computer interaction applications.