The Participants with stable vital signs were scheduled for specialized rehabilitation training upon enrollment, and targeted rehabilitation programs were selected based on the patient's specific limb function impairment, conducted by experienced rehabilitation therapists.

The patients were instructed to watch the recorded video and imitate the movements of their affected limbs. After the video ended, only the instructions and synchronized music were played. The patient was required to sit in front of a mirror table, with the mirror placed between both arms and facing the healthy side. The patient was required to recall the video based on the verbal instructions while using the healthy side to perform movement training. They were also asked to observe the reflected motion of the healthy limb in the mirror, simultaneously imagining the affected limb moving in sync (the patient focused on the front of the mirror, avoiding direct eye contact with the healthy hand, to create the sensation that it was the affected hand performing the movements).

The surface electromyography (sEMG) was used to assess the coordinated contraction rates of the triceps brachii on the affected upper limb. The surface electromyography device was used for detection, with the most prominent part of muscle belly of triceps brachii serving as the monitoring points. During the maximum isometric voluntary contraction (MIVC) state of elbow flexion or extension, two channels were used to separately record the integrated electromyographic values (iEMG) of the triceps brachii, and the antagonists. Measurements were taken three times to obtain the maximum value, and the coordinated contraction rate (CR) during elbow flexion or extension under MIVC conditions was calculated using the formula: CR (%) = Antagonist muscle iEMG / (Agonist muscle iEMG + Antagonist muscle iEMG).

the pain level of the affected shoulder was measured using the Visual Analog Scale (VAS), with 0 indicating no pain and 10 indicating unbearable severe pain. The lower score indicated the lighter pain.

The anxiety was assessed using the Generalized Anxiety Disorder-7 (GAD-7), including items related to fear, worry, attention, etc. The total score ranged from 0 to 21, which was positively correlated with potential anxiety. In the previous study, the GAD-7 scale demonstrated a Cronbach's α-coefficient of 0.879. The scores range from 0 to 21. The higher scores mean a worse outcome.

The depression was assessed using the Patient Health Questionnaire-9 (PHQ-9), with aspects including mood swings, optimism, sleep quality, appetite, etc. The total score of PHQ-9 ranged from 0 to 27, which was positively correlated with potential depression. In the previous study, the PHQ-9 scale demonstrated a Cronbach's α-coefficient of 0.913.The scores range from 0 to 27.The higher scores mean a worse outcome.