Editor, Editors, USER, admin, Bureaucrats, Check users, dev, editor, Interface administrators, lookupuser, oversight, Push subscription managers, Suppressors, Administrators, translator, Widget editors
17,889
edits
Difference between revisions of "Bilateral Trigeminal neuromotor organic symmetry"
Bilateral Trigeminal neuromotor organic symmetry (view source)
Revision as of 16:09, 8 March 2022
, 2 years ago→Results
Gianfranco (talk | contribs) |
Gianfranco (talk | contribs) |
||
| Line 284: | Line 284: | ||
==Results== | ==Results== | ||
The table 1 | The table 1 shows the list of row EMG value which were subsequently normalized and weighed in order to train – as input – the ANN. In Table 2 we report instead the descriptive and comparative statistical results. | ||
Standardized asymmetry (*) and Standardized Kurtosis (**) outside -2 to +2 range indicate significant deviations from normality and could compromise the validity of many | With regard to the descriptive statistical aspect we can consider the mean and SD values for onset latency (1.96 msec ± 0.18 msec vs. 2.01 msec ± 0.21 msec), amplitude (5.76 mV ± 2.01 mV vs''.'' 5.89 mV ± 2.51 mV) and integral area (11.09 mV/msec ± 4.45 mV/msec vs''.'' 11.27 mV/msec ± 4.34 mV/msec) for right and left masseter, respectively. | ||
[[File:Symmetry 2.jpg|thumb|'''Table 2:''' <br /> | |||
The table shows statistically descriptive data and ''p''-values between sides for onset latency, amplitude and integral area of the bR-MEPs.<br /> | |||
Standardized asymmetry (*) and Standardized Kurtosis (**) outside -2 to +2 range indicate significant deviations from normality and could compromise the validity of many statistical procedures.|alt=|500px]] | |||
The Kruskal-Wallis test shows a statistically not significant difference between the medians (confidence level 95%); in fact, we obtained <math>p-value</math> as 0.33, 0.96 and 0.86 between sides for latency, amplitude and the EMG integral area, respectively for the <sub>b</sub>R-MEPs (Table 2). | |||
In this study, ANN is used to predict the values of the right side by inputting values from the left side. Then the left- and right-side values are combined in a ratio called | In this study, ANN is used to predict the values of the right side by inputting values from the left side. Then the left- and right-side values are combined in a ratio called 'correlation coefficient'. Then the correlation coefficient is computed for the actual observed values (right/left) and then, additionally, for the ratio of the ANN-derived right/left. | ||
The MSE and the <math>R^2</math> were computed to test the ANN. If the ANN estimated the correlation coefficients with zero error, MSE must be 0 and the <math>R^2</math> must be 1. In comparison of the correlation coefficients which computed from the EMG signals of the right and the left muscles and the outcomes of the ANN (Table 3) which were trained with the normalized features computed from the EMG of the left muscles, it can be seen that the outcomes of the ANN are closest to the correlation coefficients (Table 4), and that ANN is able to compute correlation coefficients, based on features of only the left muscles, with almost zero errors. | The MSE and the <math>R^2</math> were computed to test the ANN. If the ANN estimated the correlation coefficients with zero error, MSE must be 0 and the <math>R^2</math> must be 1. In comparison of the correlation coefficients which computed from the EMG signals of the right and the left muscles and the outcomes of the ANN (Table 3) which were trained with the normalized features computed from the EMG of the left muscles, it can be seen that the outcomes of the ANN are closest to the correlation coefficients (Table 4), and that ANN is able to compute correlation coefficients, based on features of only the left muscles, with almost zero errors. | ||