Optimization of threshold values for automated analysis of EMG recordings.

1
Optimization of threshold values for automated
analysis of EMG recordings.
University of Zurich, Switzerland
L.M. Gallo, P. Rompré, G.J. Lavigne and S. Palla
2
Long-time EMG
Orofacial motor activityduring sleep is still a
controversial issue.
3
Ambulant long-time EMG recording

• Solberg et al., 1975
• Burgar Rugh, 1983
• Lotzmann et al., 1992
• Bowley JF et al., 1993
• Rivera-Morales McCall, 1995
• Gallo Palla, 1995

4
Normative study
Gallo LM, Salis Gross SS Palla S Nocturnal
masseter EMG activity of healthy subjects in a
natural environment.J Dent Res 78(8)1436-1444,
1999
5
Recording system
6
Subjects of the normative study
21 asymptomatic volunteers
(6 females and 15 males)mean age 31 (22 to 37)
7
Signal analysis

• A0 amplitude threshold
• t0 stand-by time
• TE1, TE2 duration
• Amax1, Amax2 maximum amplitude

8
Signal levels
9
Normative study
10
Aim of this study
Optimization of a simple threshold criterion
for best separation in the automated analysesof
EMG recordings in bruxers and controls.
11
Polysomnography EMG
12
Present study
Data from a single masseter channel of
polysomnographic recordings.
13
Reformatting of POLYSOM signal
14
Data analysis

• N number of episodes
• Amean mean amplitude
• Amax maximum amplitude
• T duration
• I integral amplitude over time

15
Subjects
10 bruxers (5 f 5 m, aged 23-39 y) 10
controls (5 f 5 m, aged 21-45 y)
16
Variation of threshold A0
17
Sleep duration
bruxers 824 039controls 825 028
18
Example Amax
du
bruxers
controls
threshold
19
Best separation of analysis parameters
A0 10 MVC
20
Bruxers (threshold A0 10 MVC)
21
Controls (threshold A0 10 MVC)
22
Comparison bruxers-controls
23
Difference controls-bruxers at 10 MVC
24
Separation bruxers/controls
A lower threshold of 10 MVC appears to be an
effective criterion yielding the best separation
between bruxers and controls in an automated
analysis of masseter EMG recordings.