AIM: To analyze the reliability and describe the parameterization with inertial
sensors, of Romberg test in people who have had a stroke.
METHODS: Romberg's Test was performed during 20 seconds in four different
setting, depending from supporting leg and position of the eyes (opened eyes /
dominant leg; closed eyes / dominant leg; opened eyes / non-dominant leg; closed
eyes / non-dominant leg) in people who have suffered a stroke over a year ago.
Two inertial sensors (sampling frequency 180Hz) were placed in lumbar (L5-S1) and
in the trunk (T7). The test was performed three times for the four different
settings. The outcome variables were extracted in each of the axes (X, Y, Z) (figure
1). We considered as the maximum, minimum and mean velocity and the magnitude of the angular
displacement for each of the sensors (Figure 2 shows an example
of direct extraction of variables). Statistical Analysis: descriptive
analysis of all outcome variables for each axis and sensor. Further
analysis of the internal consistency of the measure was
performed by analysis interclass correlation (ICC) with a
confidence interval of 95%.
RESULTS: Values obtained after statistical analysis show levels
of reliability ranging from 0.61 (Z axis speed - eyes closed / nondominant
leg) and 0.92 (Y axis offset - opened eyes / dominant leg).
The descriptive results of all outcome variables are shown in Table 1.
CONCLUSION: Making inertial sensors in trunk and / or lumbar, inertial sensors are reliable tools for
parameterizing Romberg test in different settings in people who have suffered stroke over a year ago.