Gait Kinematic Dependent Plantar Stimulation

Abstract

Vibro-tactile stimulation is often used to stimulate the plantar surfaces during walking to augment or replace the pattern of natural sensory feedback. However, many of these methods supply patterns of vibro-tactile feedback that are independent of the individual's real-time gait kinematics. If the purpose of the stimulation is to augment the natural feedback provided by the tactile receptors, the additional vibrations should change depending on the real-time movements similar to those natural receptors. For this reason, we created a method of applying plantar vibro-tactile stimulation through tactor embedded insoles that activate and deactivate based on the kinematic phases of walking in real-time. We demonstrate and validate a gait-like pattern of tactor activations. This pattern sequentially stimulates the plantar surfaces to follow the natural progression of the stance phases of walking. Gait events such as heel-strike, midstance, heel-lift, and toe-off were used to reliably drive a natural pattern of tactile stimulation across the plantar surface during walking. This real-time detection of gait events produced only small differences when compared to post-analysis detection methods. Overall, this controller can be used for a multitude of live feedback systems to attempt to better understand how real-time feedback is used during different tasks such as walking.

Fig. 1.

Equipment and subject set-up. (A) C-2 tactors with washer on top to ensure comfort and vibration perception. (B) Tactor activation boxes were attached to fanny pack with velcro and tape that was worn by subject. The tactors within the insoles were connected to these tactor boxes via connecting wires coming out of the shoes. (C) These wires were attached to the shanks of the subject to limit wire movement while walking on the treadmill. These boxes were connected to a computer with the real-time controller via USB connection.

Fig. 2.

Flow chart describing the real-time controller for gait-like stimulation. Stance detection uses the heel and toe marker velocity changes for determining heel-strikes and toe-off times. The sequence of tactor actiations is controlled by the AP ankle and hip marker positions to determine midstance, and the vertical position difference between the heel and toe markers to determine heel-lift. If the limb is in swing, all tactors are off.

Fig. 3.

Tactor activations and data driving those activations from a subject that received the gait-like stimulation while walking. The far right shows the location of the active tactor set. (A) The heel tactors were activated from the heel marker velocity. (B) The tactor set at MT5 was controlled by finding the difference between the ankle and hip marker locations in the AP direction. (C) The final tactor set at MT1 turned on from the difference between the heel and toe marker vertical positions became greater than 3cm. (D) The MT1 tactors turn off by the toe marker velocity change.

Fig. 4.

The real-time controller provides feedback according to gait phases. Average durations of ON time for each tactor set calculated from both the real-time controller and post-hoc analysis across all subjects. The offsets between bars depicts the time delays of the real-time controller.