Sampling Rates Influence Interpretation of Data from Force Plate Center of Pressure Measures of Postural Control

Abstract

Mean velocity of center of pressure (MVCOP), calculated from force plate data, has been used frequently to assess postural control in a variety of populations. However, there is little consistency in the literature with respect to the sampling rate selected for data collection and analysis. PURPOSE: To analyze force plate data obtained at three different sampling rates to determine the most physiologically relevant rate for the assessment of postural control. METHODS: MVCOP was calculated for twenty-six female figure skaters from force plate measurements taken from two different 15 second trials (quiet single leg stance, and a single leg jump landing from an 8 inch box). Data were collected at a sampling rate of 1,000 Hz, and subsequently downsampled to 100 Hz and 10 Hz in MS Excel by selecting every 10th or 100th point, respectively, for analysis. Pearson's correlations were utilized to identify relationships between MVCOP at each sampling rate and subject characteristics (age, height, weight, and years of competition experience). Significance was set a priori at α = 0.05. RESULTS: When the data was analyzed at 1,000 Hz the MVCOP calculation was almost exclusively dependent on subject weight for both quiet stance (R2 = 0.95; p < 0.05) and for the jump landing test (R2 = 0.92; p < 0.05). At 100 Hz, the dependence of MVCOP on subject weight was eliminated with respect to the jump landing test (R2 = 0.09; p > 0.05), but was still a significant influence for quiet stance (R2 = 0.47; p < 0.05). There was an inverse correlation between years of competition experience and MVCOP for quiet stance only (R2 = 0.44; p < 0.05). By downsampling to 10 Hz we were able to eliminate subject weight as a factor for both quiet stance (R2 = 0.10; p > 0.05) and for the jump landing test (R2 = 0.15; p > 0.05). However, there was no longer a correlation between years of competition experience and MVCOP. CONCLUSION: The data suggest it is inappropriate to analyze MVCOP at 1,000 HZ as the signal is highly influenced by subject weight, thereby making the measurement insensitive to changes in postural control. However, it was unclear as to whether 100 Hz or 10 Hz produced the most physiologically relevant information.