1. Clinical data 1.1 General data There were 58 cases in this group, 37 males and 21 females. The minimum age was 4 years old, the maximum age was 12 years old, and the median age was 7 years old. Using Asthworth’s muscle tone assessment criteria, all patients had muscle tone above grade 3, and were patients with spastic cerebral palsy with bilateral lower extremity involvement as the main cause. There were 28 cases with a pointed foot gait due to triceps spasm alone and 30 cases with both scissor and pointed foot gait due to spasm of the whole lower limbs. There were 44 cases (76%) who could walk independently. There were 14 cases (24%) who could walk with the support of one or both hands. 1.2 Surgical method: General anesthesia with tracheal intubation, prone position with head low and hip high, posterior median lumbar incision, resection of the spinous process and pushing plate. The posterior median epidural incision was made longitudinally to reveal the cauda equina nerve, and the segment of the spinal nerve root was determined according to the position of the intervertebral foramen, generally the L5 nerve root was the thickest. The posterior roots are generally wider and flatter than the anterior roots, lighter in color, and located on the posterior side of the anterior roots, so that gentle stimulation with a detachment hook will not cause muscle contraction. The posterior roots were divided into two to four bundles according to the natural division line, and each bundle was stimulated with a nerve threshold meter. The threshold of each bundle was measured, and the bundle with the lower threshold was cut off, and the general cut-off ratio was L2-L330-40%, L5-S140-50%, and L4 was generally not cut off. The incision is flushed, the dura is sutured, the drainage tube is placed, and the incision is closed. 1.2 Combination of surgical approaches For patients with acromioclavicular gait without adductor spasm, L5 and S1 selective posterior spinal nerve rhizotomy was performed. For patients with both scissor and acromioclavicular gait, L2, L3, L5 and S1 selective posterior spinal nerve rhizotomy was performed in 17 patients in the early stage, and in 13 patients in the late stage, a combination of L5 and S1 selective posterior spinal nerve rhizotomy plus adductor rhizotomy was performed. 1.4 Rehabilitation training was carried out preoperatively and postoperatively, including muscle strength training, training to increase joint mobility, low back muscle training, and movement coordination training. For patients who could not walk independently, insistence on crawling training was emphasized. Three weeks after the operation, the patient was taken out of bed with a lumbar brace. 2. Results After follow-up from 3 months to 63 months, all patients showed obvious decrease in muscle tone after surgery, and gradually recovered to near normal level within 2 weeks, and the rate of spasticity release reached 100%, and tendon reflex hyperactivity, ankle tremor and scissor gait all disappeared completely. The gait of 44 patients who could walk independently before surgery improved significantly after surgery, and 21 patients could walk independently on one foot. 14 patients who had to hold hands before surgery could walk independently in 9 cases. For patients with both scissor gait and acromion gait, the combination of small-scale selective posterior spinal nerve root amputation plus medial adductor amputation had the same efficacy as large-scale selective posterior spinal nerve root amputation, and both patients were completely relieved of scissor gait, with easy separation of the legs and no recurrence. Some patients also showed improvement in upper limb spasticity, drooling, strabismus, epilepsy, and speech disorders.