Andrew Tsourkas' pictureAndrew Tsourkas

Cornell University

Dr. vanPutten

Supported by NSF-REU

Determining the precise onset of myoelectric activity has been a problem for some time. Several methods have been developed to detect the onset, but experiments show that these methods all involve a systematic error. The Maximum Likelihood Method (MLM), however, eliminates this systematic error and has shown to be extremely consistent in determining the precise onset of myoelectric activity.

A simulation of myoelectric signals, with known onset times, were generated by Matlab. The Maximum Likelihood Method and several other methods were then used to find the onset time of the signal. The MLM showed no systematic error while the other methods reported substantial systematic errors.

One hundred samples of actual EMG signals were produced by the contraction of the tibialis anterior. The Maximum Likelihood Method and several other methods were also used to find the onset time of the actual EMG signals. The differences between the onset time found by the MLM and the other methods on the actual EMG signals reflected the differences found in the simulation tests. This suggests that the Maximum Likelihood Method is accurate in determining the onset of myoelectric activity

--------------

Studies show that older females are injured due to falling much more often that younger females. Whether this is because of differences in strength, balance, response time, etc. in unknown. Females from two different age brackets (18-25 and 0-70) were tested in an attempt to answer this question.

The subjects were pulled either right or left, twenty-four times, in random order. Their reactions were recorded using force plates and infra-red sensors. These recordings were then formatted through Matlab. Visual observations in this experiment did not point to any obvious differences between the younger and older females response.

at23@cornell.edu