Fig 1.
Setup diagram for experimental protocols.
A and B show the treadmill and overground walking entrainment experimental setups. The electrical stimulation (ES) device sits inside a drawstring backpack and connects to a handheld computer. An IMU sensor on the shoe also connects to the handheld computer to determine stride period from foot movement. The computer commands rhythmic ES to the right medial gastrocnemius muscle at the subject’s observed stride period. C shows the additional equipment used to test electromuscular response during electrical stimulation. In addition to the ES device, IMU sensor, and handheld computer, a tendon tensiometer is worn on the Achilles tendon and records data through a desktop data acquisition system. All equipment is synchronized wirelessly and ES is applied to the right gastrocnemius muscle at unpredictable gait phases while the subject walks on a treadmill.
Table 1.
Subject stimulation amplitudes, participation, and entrainment results.
Fig 2.
Real-time stride period calculation.
The magnitude of pitch angular velocity is used to detect strides. Heel-off (HO) is detected when pitch angular velocity magnitude exceeds the threshold after it was below the threshold for at least 0.16 s. Stride period (ts) is the time between consecutive HO’s and is averaged over the first 10 s of each trial.
Fig 3.
Flow chart for applying ES at unpredictable gait phase (Experiment C).
Average stride period over the first 10 s is used to create delay times to test approximately 8 gait phases during stance and 4 gait phases during swing, each repeated 7 times and in random order. Real-time pitch angular velocity is used to detect strides. After a random delay of 4-6 strides the electrical stimulation (ES) is applied to the right gastrocnemius muscle at the next gait phase in the sequence. The same procedure to calculate delays and stimulus phases was used for all three stimulus amplitudes.
Fig 4.
Examples of trials with entrainment of gait (Experiments A and B). The sequence of stimuli advances along the vertical axis, with the gait phase of each stimulus plotted on the horizontal axis. The final stimulus is placed within the 0-100% range of gait phase, and phase is unwrapped across gait cycles to show continuous drift. The left graph shows a two-minute treadmill trial that entrained to the ES (Experiment A). Successive stimuli drift later in the stride (a positive slope), indicating that the stride period of the subject before entrainment was less than the stimulation period. The right graph shows a two-lap overground trial that entrained to the ES (Experiment B). Successive stimuli drift earlier in the stride (a negative slope), indicating that the stride period before entrainment was greater than the stimulation period. The regions where entrainment occurred (gait phase within +/- 3% for 20 consecutive strides) are represented by a black vertical line.
Fig 5.
Calculating Δt for maximum effect from ES (Experiment C).
The blue curve shows the mean Achilles tendon wave speed for all non-ES strides for a representative subject. The purple curve shows a portion of wave speed for an ES stride (ends 0.3 seconds after ES). The square denotes the time and wave speed when ES was applied. The black lines show the difference between ES and non-ES wave speed, and the star shows the time and wave speed when the difference between the ES wave speed and the non-ES wave speed was greatest. Overall was calculated as the mean of individual Δt values from all ES across all subjects (
= 0.1465 s).
Fig 6.
Histograms of gait phase of electrical stimulus during entrained portions of treadmill (Experiment A) and overground walking (Experiment B). The left graph shows the distribution of gait phase of ES for entrained strides during treadmill walking and the right graph shows the distribution of gait phase of ES for entrained strides during overground walking. The curves show the kernel smoothing function fit for the histograms, and the locations of the peaks based on the fitted curves are denoted with vertical lines (in percent gait phase). For both treadmill and overground walking conditions, the gait phase of ES for entrained strides primarily occurred just before toe-off (TO) and heel strike (HS).
Fig 7.
Percent of entrainment for treadmill (Experiment A) and overground (Experiment B) trials.
The left graph shows that a greater percentage of subjects entrained to the electrical stimulus during treadmill walking than during overground walking. The middle graph shows that a similar percentage of trials exhibited entrainment in both conditions. The right graph shows that the percentage of total strides that occurred during periods of entrainment was significantly higher in treadmill than in overground walking (p < 0.001). Numbers in the bars indicate how many subjects, trials and strides were or were not entrained.
Fig 8.
Stride period variation from stimulation period (Experiments A and B). The graph shows the stride period mean square error (SPMSE) from the preset stimulation period for each subject’s treadmill and hallway trials. The *** represents p < 0.001. In general, the SPMSE was greater during overground walking, although not significant when combining all subjects.
Fig 9.
Achilles tendon wave speed from a representative subject with and without ES (Experiment C).
The black curve is the mean wave speed for all non-ES strides and the grey shading is the standard deviation band. The colored curves show local deviations in wave speed after 12 ES were applied – 8 during stance and 4 during swing. The squares show the start of ES and the stars show the wave speed after (0.1465 s) for the ES stride of the corresponding color. Traces from each stimulus are shown from 0 to 0.3 s after the time of each stimulus.
Fig 10.
Representative subject normalized Achilles tendon wave speed for grouped ES timing (Experiment C).
The black curve is the mean normalized wave speed for all non-ES strides and the grey shading is the standard deviation band. Wave speed values are normalized relative to the subject’s mean peak wave speed for strides without ES. The colored boxes show the range of gait phase for each ES timing band, and the asterisk (*) of the same color represents normalized wave speed at (0.1465 s) after each ES.
Fig 11.
Normalized Achilles tendon wave speed with and without ES at different gait phases (Experiment C).
Wave speed was higher with ES than without ES for most phases with Max stimulation (right graph), with effects diminishing at lower levels of stimulation. Wave speed values are normalized relative to each trial’s mean peak wave speed for strides without ES. The red stars show the mean normalized wave speed for all ES strides within each bin of gait phase, measured after stimulation began. The corresponding blue circles show the mean normalized wave speed for all non-ES strides at the same time. Error bars show standard deviation. The bottom axis shows the gait phase of the ES; the x-value of each data point represents the end of the associated bin, and the upper axis shows the gait phase when the wave speed measurement was recorded. Symbols: * p < 0.05, ** p < 0.01, *** p < 0.001.
Fig 12.
Normalized stride period for ES strides grouped by ES phase, compared to mean stride period for no-ES strides (Experiment C).
ES led to altered stride period with the direction and magnitude of the change strongly dependent on the gait phase when ES was applied. The strongest effects were during early swing and early stance, especially with Max stimulation. Stride period is normalized relative to each trial’s mean stride period for strides without ES. The grey shading is the standard deviation band for strides without ES. The boxplots show normalized stride periods (mean of stride period for the ES stride and subsequent stride) for all ES strides within each bin ES initiation phase. The red line denotes the median and the red dot denotes the mean. The bottom and top edges of the boxes indicate the 25th and 75th percentiles. The bottom and top edges of the whiskers indicate the 9th and 91st percentiles. Symbols: * p < 0.05, ** p < 0.01, *** p < 0.001.