19 Signals were then passed through a BNC adapter chassis that was interfaced with an analog-to-digital board within a personal computer. These signals were then converted to ground reaction force vectors and moments. Data were filtered using a second order recursive low-pass Butterworth digital filter with an estimated optimum cutoff frequency

of 12.53 Hz.19 A customized LabVIEW (National Instruments Corp., Austin, TX, USA) software program computed A/P and M/L TTS. A/P and M/L components of the ground reaction force data were analyzed separately for each subject, but the same procedure KU-55933 clinical trial was used for both components. First, the last 10 s of the ground reaction forces were analyzed to find the smallest absolute ground reaction force range for each component.19 These ranges were accepted as the optimal range of variation values.19 A/P and M/L components of the ground reaction force data were then rectified.19 An unbounded

third order polynomial was fit from the peak force to the last data point for each component.19 TTS for each component was the point where the unbounded third order polynomial was equal to or less than the respective optimal range of variation value.19 Average A/P and M/L TTS values for each treatment condition were computed in PASW version 18.0 (SPSS, Inc., Chicago, IL, USA). Alpha level was set a priori at p ≤ 0.05 to indicate statistical STAT inhibitor significance. One-tailed paired samples t tests compared SRS to control conditions for A/P and M/L TTS. Effect size d values were calculated for each t test. 22 Average percent improvements for each TTS measure were also computed for all subjects and average improvement of eight subjects who

improved with SRS (subjects who did not improve were removed). No improvements were defined as increased TTS with SRS over a control condition. Lastly, to provide insight on why some subjects did not improve with SRS, we computed effect size d values for comparing responders and non-responders on frequency of sprains, frequency of “giving-way”, and score on the AJFAT. SRS significantly improved A/P TTS over the control condition (SRS = 1.32 ± 0.31 s, Control = 1.74 ± 0.80 s; 3-mercaptopyruvate sulfurtransferase t(11) = −2.04, p = 0.03; d = 0.76). The average percent improvement for A/P TTS with SRS was 24% (n = 12) and increased to 34% (n = 8; SRS = 1.32 ± 0.35 s, Control = 2.01 ± 0.86 s) when four subjects who did not improve were removed. SRS did not affect M/L TTS (SRS = 1.95 ± 0.40 s, Control = 1.92 ± 0.48 s; t(11) = −0.20, p = 0.42; d = −0.07). The average percent improvement for M/L TTS with SRS was 2% (n = 12) and increased to 15% (n = 8; SRS = 1.75 ± 0.30 s, Control = 2.06 ± 0.50 s) when four subjects who did not improve were removed. Using effect size d values to detect mean differences, non-responders had greater mean values than responders on frequency of sprains, frequency of “giving-way”, and score on the AJFAT.