Patients
Nineteen patients with stroke who had been diagnosed with poststroke hemiplegia at least 1 year ago were recruited from the welfare institute for the disabled in Gyeonggi Province. The patients who agreed to participate after receiving a detailed explanation about the experimental procedure were recruited. The inclusion criteria were individuals who could perform unassisted independent walking for at least 10 m and who were able to understand the therapist’s directions and scored 24 or above in the Mini Mental State Examination-Korean (MMSE-K). Participants with visual or hearing problems, vertigo or vestibular dysfunction, and orthopedic or cardiorespiratory problems were excluded from the study. All participants gave informed consent after detailed orientation of the study purpose and important matters prior to the experiment. The physical characteristics of the participants are shown in Table 1.
Experimental procedure and data collection
We conducted the experiment to measure and assess the variables of dynamic balance in the study participants after hearing white noise.
We used white noise for auditory feedback, which was presented for 20 minutes to the participant who was comfortably seated in a quiet place and wearing optic glasses that blocked vision to amplify the auditory effect within a short window. White noise used in this study was generated using the MC square X7 (GEOMC) equipped with 6 programs and 6 natural sounds that could be randomly mixed depending on the user’s taste and used simultaneously. For the purposes of this study, we used a mix of 1 program and 6 natural sounds (Figure 1). This program functions to regulate attention and biasing using pulsed sounds as external stimuli, which is thought to activate the reticula formation in the midbrain that induces blockage of alpha waves, while the thalamic tract acts to globally broadcast alpha. Beta waves induced during all states of conscious activity disrupt alpha waves, induced during relaxation and mediation (mental stability) with closed eyes, and theta waves, induced during states of creativity and learning.29
In this study, participants performed walking before and after hearing white noise to measure their dynamic balance abilities and CoP, CoM, and gait speed during walking. To measure the kinetical variable during walking, we used a force platform (AMTI, BP1200, USA) and measured the CoP, which is a classical variable used for balance testing.30 Participants were allowed to walk at preferred speed. We set the sampling rate to 1000 Hz and collected data using the Qualisys track manager (QTM, Qualisys, Sweden) software. The force platform was synced with and triggered by the QTM program by interconnecting the A/D boards using internal trigger cables. To record movements during walking, we used 8 infrared cameras (Oqus 300, Qualisys, Sweden) with a sampling rate of 100 Hz. The human body was estimated to be a rigid body of 14 segments (right/left feet, lower limbs, quadriceps, hands, upper limbs, biceps, head, and trunk) and a total of 46 reflection markers were attached. For each participant, we built the 3D coordinates of the walking space through Non-Linear Transformation (NLT) and recorded gait motion using a total of 8 infrared cameras. We simultaneously sampled the locations of the ground reaction force (GRF) using the force platform with the sampling rate of 1000 Hz. The CoMs for calculating the anterior-posterior (A-P) and medial-lateral (M-L) inclination angles in dynamic balance assessment were recorded using the 8 infrared cameras.
Data processing
We conducted kinematic and kinetic analyses on the collected data for derivation of variables using the Visual 3D software (C-motion, USA) and Matlab 2014a software (The Mathworks, USA).
To explore the kinematic variables, we set the space coordinates using the NLT method and defined the exact area by extracting the 3D coordinates of the reflection markers attached to the human body. Space coordinates were processed using the Visual 3D software (C-motion, USA) (Matlab R2014b (The Mathwork, USA).
The GRFs for kinetic variable derivation were stored in the QTM as 8-channel analog voltage values, which are then converted and outputted as a total of 3 digital values (Fx, Fy, Fz, Units: N). The positive values of Fx, Fy, and Fz were defined as the left, anterior, and upward perpendicular, respectively. To remove noise-driven errors, we low-pass filtered the data using a second-order Butterworth filter with a cutoff frequency corresponding to the 99% integral of the power spectrum density (PSD). After the GRF values had been derived, we normalized data across participants by dividing the number of frames per gait phase by the phase duration and showing the value in percentage.
Analysis variables
In this study, we measured the kinetic variable of CoP range and velocity, A-P/M-L inclination angles, and gait speed to test the effects of white noise on dynamic balance.
CoP range and velocity
Using the CoP derived from the GRF data obtained using the force platform, we calculated the M-L and A-P CoP range and CoP velocity as a function of time. CoP range was described as the difference between the maximum and minimum values and CoP velocity as the mean of instantaneous velocities.31
Using the anthropometric model as reference, we calculated the location of segmental CoM based on the proximal and distal markers of each segment and derived the CoM using the coordinates of the markers.32,31 We derived the locations of all segmental CoMs and then calculated the weighted sum of all body segments to obtain the whole-body CoM position data.30
We performed a 3D motion analysis to calculate the inclination angles. We analyzed the stance and swing phase data during walking from heel-contact to toe-off on the paretic and non-paretic sides. For each side, the A-P inclination angle in the sagittal plane and the M-L inclination angle in the frontal plane were defined as the angle between the vector connecting the CoM and CoP and the vertical axis. The position vectors of CoM and CoP were calculated using the inverse tan2 equation (Figure 2).
Gait speed
For gait speed, we analyzed the stance and swing phase data during walking from heel-contact to toe-off on the paretic and non-paretic sides.
Statistical analysis
To identify the pre-post white noise changes in the variables of dynamic balance – CoP range and velocity, CoM, A-P/M-L inclination angle, and gait speed – we conducted a paired samples t-test using a significance level of α=.05.