TY - JOUR
T1 - Evaluation of force-sensing resistors for gait event detection to trigger electrical stimulation to improve walking in the child with cerebral palsy
AU - Smith, Brian T.
AU - Coiro, Daniel J.
AU - Finson, Richard
AU - Betz, Randal R.
AU - McCarthy, James
N1 - Funding Information:
Manuscript received August 14, 2000; revised April 17, 2001, September 19, 2001, and March 12, 2002. This work was supported by Shriners Hospital for Children under Research Grant 8530. The authors are with Shriners Hospitals for Children, Philadelphia, PA 19140 USA (e-mail: [email protected]). Publisher Item Identifier S 1534-4320(02)05847-3.
PY - 2002
Y1 - 2002
N2 - Force-sensing resistors (FSRs) were used to detect the transitions between five main phases of gait for the control of electrical stimulation (ES) while walking with seven children with spastic diplegia, cerebral palsy. The FSR positions within each child's insoles were customized based on plantar pressure profiles determined using a pressure-sensitive membrane array (Tekscan Inc., Boston, MA). The FSRs were placed in the insoles so that pressure transitions coincided with an ipsilateral or contralateral gait event. The transitions between the following gait phases were determined: loading response, mid- and terminal stance, and pre- and initial swing. Following several months of walking on a regular basis with FSR-triggered intramuscular ES to the hip and knee extensors, hip abductors, and ankle dorsi and plantar flexors, the accuracy and reliability of the FSRs to detect gait phase transitions were evaluated. Accuracy was evaluated with four of the subjects by synchronizing the output of the FSR detection scheme with a VICON (Oxford Metrics, U.K.) motion analysis system, which was used as the gait event reference. While mean differences between each FSR-detected gait event and that of the standard (VICON) ranged from +35 ms (indicating that the FSR detection scheme recognized the event before it actually happened) to -55 ms (indicating that the FSR scheme recognized the event after it occurred), the difference data was widely distributed, which appeared to be due in part to both intrasubject (step-to-step) and intersubject variability. Terminal stance exhibited the largest mean difference and standard deviation, while initial swing exhibited the smallest deviation and preswing the smallest mean difference. To determine step-to-step reliability, all seven children walked on a level walkway for at least 50 steps. Of 642 steps, there were no detection errors in 94.5% of the steps. Of the steps that contained a detection error, 80% were due to the failure of the FSR signal to reach the programmed threshold level during the transition to loading response. Recovery from an error always occurred one to three steps later.
AB - Force-sensing resistors (FSRs) were used to detect the transitions between five main phases of gait for the control of electrical stimulation (ES) while walking with seven children with spastic diplegia, cerebral palsy. The FSR positions within each child's insoles were customized based on plantar pressure profiles determined using a pressure-sensitive membrane array (Tekscan Inc., Boston, MA). The FSRs were placed in the insoles so that pressure transitions coincided with an ipsilateral or contralateral gait event. The transitions between the following gait phases were determined: loading response, mid- and terminal stance, and pre- and initial swing. Following several months of walking on a regular basis with FSR-triggered intramuscular ES to the hip and knee extensors, hip abductors, and ankle dorsi and plantar flexors, the accuracy and reliability of the FSRs to detect gait phase transitions were evaluated. Accuracy was evaluated with four of the subjects by synchronizing the output of the FSR detection scheme with a VICON (Oxford Metrics, U.K.) motion analysis system, which was used as the gait event reference. While mean differences between each FSR-detected gait event and that of the standard (VICON) ranged from +35 ms (indicating that the FSR detection scheme recognized the event before it actually happened) to -55 ms (indicating that the FSR scheme recognized the event after it occurred), the difference data was widely distributed, which appeared to be due in part to both intrasubject (step-to-step) and intersubject variability. Terminal stance exhibited the largest mean difference and standard deviation, while initial swing exhibited the smallest deviation and preswing the smallest mean difference. To determine step-to-step reliability, all seven children walked on a level walkway for at least 50 steps. Of 642 steps, there were no detection errors in 94.5% of the steps. Of the steps that contained a detection error, 80% were due to the failure of the FSR signal to reach the programmed threshold level during the transition to loading response. Recovery from an error always occurred one to three steps later.
KW - Cerebral palsy
KW - Children
KW - Foot switches
KW - Functional electrical stimulation
KW - Gait
UR - https://www.scopus.com/pages/publications/0035992805
U2 - 10.1109/TNSRE.2002.1021583
DO - 10.1109/TNSRE.2002.1021583
M3 - Article
C2 - 12173736
AN - SCOPUS:0035992805
SN - 1534-4320
VL - 10
SP - 22
EP - 29
JO - IEEE Transactions on Neural Systems and Rehabilitation Engineering
JF - IEEE Transactions on Neural Systems and Rehabilitation Engineering
IS - 1
ER -