Brain-controlled cycling system for rehabilitation following paraplegia with delay-time prediction

Abstract

Abstract Objective. Robotic rehabilitation systems have been investigated to assist with motor dysfunction recovery in patients with lower-extremity paralysis caused by central nervous system lesions. These systems are intended to provide appropriate sensory feedback associated with locomotion. Appropriate feedback is thought to cause synchronous neuron firing, resulting in the recovery of function. Approach. In this study, we designed and evaluated an ergometric cycling wheelchair, with a brain–machine interface (BMI), that can force the legs to move by including normal stepping speeds and quick responses. Experiments were conducted in five healthy subjects and one patient with spinal cord injury (SCI), who experienced the complete paralysis of the lower limbs. Event-related desynchronization in the β band (18–28 Hz) was used to detect lower-limb motor images. Main results. An ergometer-based BMI system was able to safely and easily force patients to perform leg movements, at a rate of approximately 1.6 s/step (19 rpm), with an online accuracy rate of 73.1% for the SCI participant. Mean detection time from the cue to pedaling onset was 0.83 <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" overflow="scroll"> <mml:mo>±</mml:mo> </mml:math> 0.31 s. Significance. This system can easily and safely maintain a normal walking speed during the experiment and be designed to accommodate the expected delay between the intentional onset and physical movement, to achieve rehabilitation effects for each participant. Similar BMI systems, implemented with rehabilitation systems, may be applicable to a wide range of patients.