Principal components analysis as an evaluation and classification tool for lower torso sEMG data

Neuro-rehabilitation robot-assisted assessments of synergy patterns of forearm, elbow and shoulder joints in chronic stroke patients

BACKGROUND

Abnormal synergy is one of the major motor deficits in stroke patients. Abnormal muscle synergies, in conjunction with weakness and spasticity, interfere with voluntary movements and restrict the range of motion. This study aimed to quantify abnormal synergies in the affected upper limbs of chronic stroke patients by using a neuro-rehabilitation robot.

METHODS

Twelve chronic stroke patients and eight age-matched control subjects were recruited to perform rectilinear tracking movements in four horizontal directions (back-forth, two oblique directions at 45 degrees , and right-left). Kinematic, kinetic and electromyogram data were recorded and used to develop two biomechanical indices and one electromyogram assessment index based on principal component analysis.

FINDINGS

Significant differences between upper limbs of control subjects and the affected side of stroke patients were observed in all three assessment indices. Higher correlation between the elbow joint angle and the forearm pronation/supination torque, higher variation of the forearm torque, and abnormal co-contraction of the elbow and shoulder muscles were observed in the affected limbs of stroke patients. The difference was more prominent in the right-left direction and the oblique direction contra-proximal to ipsi-distal.

INTERPRETATION

The proposed assessment indices could be employed to quantify the abnormal synergies in stroke patients. Rectilinear tracking along the right-left direction and the oblique direction of contra-proximal to ipsi-distal is more suitable for assessing abnormal synergies.

Principal components of frequency domain electromyograms for muscular fatigue analysis

The parameters commonly employed for muscular fatigue (MF) analysis, the mean frequency (F(mean)) and median frequency (F(median)) of surface EMG (SEMG) spectra, did not present consistent results for exercises performed at low to moderate intensities. For overcoming this limitation, the present study proposes the use of principal component (PC) analysis of SEMG spectra for MF monitoring. The SEMG from vastus lateralis muscle of 24 young male subjects were recorded during a maximal effort test in cycle ergometer. Epochs of 200 ms SEMG, extracted from each period of muscle activation of each subject were used for estimating the amplitude spectra, which were employed to obtain the first two PCs. The PC coefficients from the first 40 spectra were used as a reference for calculating the standard distance. This index presented a consistent increase along the exercise, while the values of F(mean) and F(median) did not present any pattern. Comparisons between initial and final values from all subjects show no significant changes in F(mean) and F(median) (Student t test, P > 0.05) and a significant increase (P < 10(-3)) in standard distance, indicating this index as an alternative to MF analysis in cyclical exercises.

Classification of EMG signals using PCA and FFT

In this study, the fast Fourier transform (FFT) analysis was applied to EMG signals recorded from ulnar nerves of 59 patients to interpret data. The data of the patients were diagnosed by the neurologists as 19 patients were normal, 20 patients had neuropathy and 20 patients had myopathy. The amount of FFT coefficients had been reduced by using principal components analysis (PCA). This would facilitate calculation and storage of EMG data. PCA coefficients were applied to multilayer perceptron (MLP) and support vector machine (SVM) and both classified systems of performance values were computed. Consequently, the results show that SVM has high anticipation level in the diagnosis of neuromuscular disorders. It is proved that its test performance is high compared with MLP.

Automated Design of Robust Discriminant Analysis Classifier for Foot Pressure Lesions Using Kinematic Data

In the recent years, the use of motion tracking systems for acquisition of functional biomechanical gait data, has received increasing interest due to the richness and accuracy of the measured kinematic information. However, costs frequently restrict the number of subjects employed, and this makes the dimensionality of the collected data far higher than the available samples. This paper applies discriminant analysis algorithms to the classification of patients with different types of foot lesions, in order to establish an association between foot motion and lesion formation. With primary attention to small sample size situations, we compare different types of Bayesian classifiers and evaluate their performance with various dimensionality reduction techniques for feature extraction, as well as search methods for selection of raw kinematic variables. Finally, we propose a novel integrated method which fine-tunes the classifier parameters and selects the most relevant kinematic variables simultaneously. Performance comparisons are using robust resampling techniques such as Bootstrap 632+ and k-fold cross-validation. Results from experimentations with lesion subjects suffering from pathological plantar hyperkeratosis, show that the proposed method can lead to approximately 96% correct classification rates with less than 10% of the original features.