Surface Electromyography Study of Fatigue in Diabetic Neuropathy

Overview

Fatigue in diabetic neuropathy compromises patients' physical activity and poses questions on how to plan correct rehabilitation training. Conclusive interpretation of muscular mechanisms of fatigue in diabetic neuropathy has not yet been achieved. Among the various instrumental evaluations for fatigue, multichannel surface electromyography (sEMG) is a recognized tool that permits the study of myoelectric manifestations of fatigue. Aim of the study is to assess if differences in myoelectric manifestations of fatigue between patients affected by diabetic neuropathy exist after an aerobic or endurance training.

The instrumental session consisted of the registration of sEMG signals during electrically stimulated and voluntary contractions according to a consolidated standardized protocol. The investigated muscle is the Anterior Tibial. Each subjects sat comfortably on a chair with his/her ankle flexed at 90° degree and knee extended; the leg was fixed at 90 degrees, in the isometric brace fixed on a plane (MISO1, LISiN Bioengineering Centre, Turin Polytechnic, Italy Torque was measured with a modular brace incorporating two independent torque transducers (model TR11, CCT Transducers, Torino, Italy). The signal from the two torque meters were amplified, summed and displayed by means of a visual feedback system, which provided the subject with information regarding the torque level produced. Torque signal were stored to be analysed later. The sEMG signal of the right Anterior Tibial was investigated with a flexible adhesive linear array of 16 electrodes (silver bars 10 mm long, 1 mm diameter, 10 mm apart) in single differential (SD) configuration. The optimal position and orientation of the array was determined at moderate contraction levels by visual inspection of the signal. It provided clear motor unit action potentials with similar propagation in the two directions from the neuromuscular junction to the tendons. The reference electrode was positioned on the patient's leg. The skin was cleaned by slightly abrading it with abrasive gel before positioning the array. Since sEMG variables are affected by muscle temperature, the skin temperature was monitored with an electronic thermometer throughout the whole examination and was kept between 31.5 C° and 32.5 C° 30. The protocol consisted in three evaluations: one stimulated contraction and two voluntary contractions, according to a standardized protocol. The stimulated contraction was executed through a button stimulation electrode (size: 10 mm) positioned on the motor point using a monopolar configuration, a frequency of 25 Hz for a duration of 30 seconds and a supramaximal stimulation. The motor point was selected as the position of the stimulated electrode on the skin where the M-wave showed the maximum amplitude for a specified stimulation intensity; the supramaximal stimulation level was defined as the current intensity above which there was no significant increase of the amplitude of the M-wave or the maximum level tolerated by the subjects. A rest period of 10 minutes after stimulation was observed in order to avoid cumulative fatigue phenomenon. The subject then performed two test-contractions by dorsiflexion of the foot against the resistance given by the braces, in order to get acquainted with the procedure and to verify the correct posture and position of the array. The subject was subsequently asked to produce three maximal voluntary contractions (MVC) lasting 3 seconds each with a rest period of 2 minutes in-between. The reference MVC, expressed in Nm, was established as the maximum of the three measurements. The last MVC measurement was followed by a 10-minute rest period. The subject then produced two voluntary contractions each lasting 30 seconds: one contraction at 30% MVC and one at 60% MVC with a 5-minute rest in between. A visual biofeedback was used to help the subject maintaining the requested contraction level; furthermore, the subjects were verbally encouraged to obtain the best outcome during their performance. The EMG signals were filtered with a 10-500 Hz bandwidth filter, amplified (EMG 16-16 channel amplifiers LISiN Bioengineering Centre Turin Polytechnic). They were sampled at 2048 Hz during voluntary contractions and 1024 Hz during electrically elicited contractions. Signals were digitised by a 16 bit A/D converter (DAQCARD-6024E National Instruments, Austin, Texas, USA) and stored on the disk of a personal computer. Signal processing was performed using MATLAB. EMG variables of interest were: mean normalized frequency (MNF), average rectified value (ARV) and muscle fibre conduction velocity (CV). Spectral (i.e. MNF), amplitude (i.e. ARV) and CV variables were computed with numerical algorithms described in previous papers. CV was estimated from the consecutive double differential signals showing the best signal propagation; MNF and ARV were estimated from the single differential channel in the middle of the channels used for CV estimation. Epoch length for EMG variable estimation was 0.5 seconds without overlapping. A linear regression was used to fit all the scatter graphs of the EMG variables with time. The rate of change was defined as the slope of the regression line. The normalized rate of change for all variables was defined as the ratio between the slope and the intercept (initial value of sEMG variables) expressed as percentage. Physiological myoelectric manifestations of muscle fatigue consist in reduction of MNF and CV and increase of ARV.