Learned-non use of the paretic foot in gait after stroke: evidence from a study on joint power

Background Over a century ago Sherrington demonstrated in monkeys that a deafferented upper limb is doomed to loose motricity, which can be recovered if compensation through the spared limb is prevented. Taub named this phenomenon “learned-non use” (LNU). The author based his observation on the discrepancy between bedside examination and the showy limping of most stroke patients. In hemiparetic gait it is known that the power provided when the paretic limb is in rear position is 3 to 5 times lower, compared to the next double-stance. Yet, to date no proof is available that this asymmetry reflects LNU of otherwise recruitable muscles. In this work, a study on joint power during walking with flexed knees (“crouch” gait), gives support to this hypothesis. Methods Six healthy adults were enrolled as controls. A sample of five stroke patients were studied. The subjects had to walk on a split-belt force treadmill. EMG from lower limb were recorded wirelessly. A comfortable speed for most hemiplegic patients on the treadmill was 0.3 ms-1. Once equipped, the subjects were requested to walk on the treadmill at a speed of 0.3 ms-1. The walking trial was repeated with knees held at about 30° flexion. One series of 6 strides was analyzed. No subject reported unbalance or discomfort. Only the joint power in the sagittal plane is considered. For each muscle group, the sign of the power output (positive during muscle shortening, negative otherwise) was identified by the direction of joint rotation (e.g. positive power is assumed to be provided by the plantar flexor muscles, if the analysis of ground reaction forces led to evidence of power making the ankle to rotate towards flexion). Results Controls: stride length and stride periods ranged 0.46-0.53 m and 1.61-1.78 s. Compared to spontaneous gait, crouching entailed a 65% increase of ankle peak power. Patients: the stride length and periods on the “affected” step were 0.17-0.29 m and 1.20-1.32 s. Compared to controls, the ankle power was on average 37% lower on the affected side and 33% higher on the sound side. Power on the affected side was 54% lower compared to sound side. Compared to controls, the ankle power was 9.6% lower on the affected side and 23% higher on the sound side. Power on the affected side was 26% lower compared to sound side. Compared to spontaneous gait, crouching entailed a 145% increase of ankle peak power on the affected side, and a 54% increase on the sound side. Conclusions During hemiplegic gait virtually no power is provided at the ankle of the paretic lower limb. By contrast, during crouched gait the ankle resumes its propulsive role, through an active plantar flexion and an appreciable power output. The asymmetry is remarkably reduced. This indicates that lower limb weakness at least partly stems from an adaptive behavior fostering the use of the sound lower limb, despite an available residual power from the paretic limb.