X-Sender: rnewton@pophost.scu.edu.au
Mime-Version: 1.0
Approved-By: Robert Newton
Newsgroups: bit.listserv.biomch-l
Date: Wed, 6 Nov 1996 16:26:20 +1100
Reply-To: Robert Newton
Sender: Biomechanics and Movement Science listserver

From: Robert Newton
Subject: Smoothing and Ensemble of EMG
To: Multiple recipients of list BIOMCH-L


We are in the process of a gait analysis of 6 below knee amputees and 6
normal subjects walking on level ground as well as up and down two grades.
We recorded EMG and force plate data as well as sagital plane video at 50 Hz.

For the analysis we intend to compare each amputee on a case by case basis
with the normal data. I am looking for suggestions on how to condense the
force and EMG data for the normal subjects to provide mean and SD data for

At the moment we intend to use a similar technique to Winter, J. Biomech,
1988. For the EMG the data is rectified and then a 3 Hz low-pass filter is
applied. The data is then normalised with respect to one stride (Stride
Time as 0-100%) and the maximum EMG value (0-100% of peak activity). For the
force data, the force would be expressed in body weights and normalised in
time to one stride. I was then going to create an ensemble by averaging each
data set across trials and all the normal subjects. This would be the
normal ensemble data.

For each amputee, I was going to create an ensemble across the repeat trials
and compare this qualitatively to the ensemble for the normal subjects under
the same conditions.

Does anyone have other suggestions or references which would be suitable for
this type of comparison? As usual I will post a summary of the replies.

Robert Newton
Lecturer, Centre for Exercise Science and Sport Management
Southern Cross University
PO Box 157, Lismore, NSW 2480, AUSTRALIA
Telephone: Int + 61 66 203 234 Facsimile: Int + 61 66 203 880
Email: rnewton@scu.edu.au Web Page: http://www.scu.edu.au/hlthsci/exsci/
Personal Web Page:

Return to PE 506 page