Sensor number in simplified insole layouts and the validity of ground reaction forces during locomotion

Research attempted to validate simplified insoles with a reduced number of sensors to facilitate clinical application. However, the ideal sensor number is yet to be determined. The purpose was to investigate the validity of vertical ground reaction forces in various simplified pressure sensor insoles and to identify an optimal compromise between sensor number and measurement performance. A Kistler force plate (1000 Hz) and 99-sensor Pedar-X insole (100 Hz) obtained force data of 15 participants during walking and jogging. Eight simplified insole layouts (3-17 sensors) were simulated. Layout performances were expressed as Pearson's correlation coefficients (r) with force plate as reference and coefficient of variation. Differences were assessed via repeated-measures ANOVA as partial eta square ($$) at p < .05. All layouts correlated with the force plate (r = .70-.99, p < .01). All layout performances were higher in jogging than in walking by r = +.07 ± .04 ($$=.28-.66, p < .05). The three- and five-sensor layouts yielded the lowest correlation (r = .70-.88) and the highest coefficient of variation (11-22%). Layout performances improved constantly from 7 to 11 sensors. The optimal compromise between simplification and measurement performance, quantified via change in correlation per sensor number, was found in the 11-sensor layout, recommendable for practical settings to improve monitoring and adjusting protocols.