NOTE: The clinical study was approved by the Medical Ethics Association of the Fifth Affiliated Hospital of Guangzhou Medical University (NO. KY01-2019-02-27) and has been registered at the China Clinical Trial Registration Center (No. ChiCTR1800017487 and entitled, "The multiple modal tasks on gait control and motor cognition after stroke").

1. Recruitment

1. Recruit stroke patients with the following inclusion criteria: patients meeting the diagnostic criteria for cerebrovascular disease of the Neurological Branch of the Chinese Medical Association (2005); cerebral infarction confirmed by computed tomography or magnetic resonance imaging; damage to the unilateral cortex or with a subcortical lesion; ability to walk independently, Brunnstrom stage ≥ 4 stages; Modified Ashworth Scale¹¹ ≤ 2 points; meeting the requirements of three-dimensional (3D) gait analysis and the ability to tolerate the whole process; and the ability to give informed consent.
2. Ensure the following exclusion criteria are met: congestive heart failure, deep vein thrombosis of the lower extremities, malignant progressive hypertension, respiratory failure or other diseases, and serious risk of falling.
3. Obtain written informed consent from all patients before beginning the study.

2. Clinical evaluation

1. Record the demographic characteristics of the patient including the name, gender, date of birth, level of education, chief complaint, current medical history, past history, medical treatment, and current medications.
2. Cognitive function assessment
   1. Ask the patient to complete the Mini-Mental State Examination (MMSE)¹² record the patient's responses to a 30-question scale with a total score of 30 points for cognition evaluation, which involves the following seven aspects: time orientation, position orientation, instant memory, attention and computing power, delayed memory, language, and visual space.
      ​NOTE: The scores of MMSE are closely related to the level of education. The normal cognitive standard is illiteracy > 17 points, primary school > 20 points, and junior high school > 24 points¹³.
   2. Ask the patient to complete the Montreal Cognitive Assessment (MoCA)¹⁴ record the patient's responses to an 11-question scale with a total score of 30 points for cognition evaluation, which involves the following eight aspects: attention and concentration, executive function, memory, language, visual structure skills, abstract thinking, calculation, and orientation.
      ​NOTE: The normal cognitive standard is ≥ 26 points. If the subject has been educated for less than 12 years, they should add 1 point to the score¹⁵.
3. Walking ability assessment
   1. Conduct the 10-m walk test (10 MWT)¹⁶. Ask the patient to perform three consecutive trials at a self-selected pace for safety, comfort, and higher speed, respectively. Record the time taken to walk to the middle 6 m in each trial (to exclude acceleration and deceleration effects).
   2. Conduct the timed up and go test (TUGT)¹⁷. Ask the patient to perform three consecutive TUG trials (stand up, walk 3 m, turn, walk back, and sit down) at a self-selected pace for safety and comfort¹⁸.

3. 3D gait analysis

1. Patient preparation
   1. Inform the patient about the precautions and the purpose of the experiment.
   2. Ask the patient to wear tight underwear to fully expose the neck, shoulders, waist, and lower limbs.
   3. Record the values of various anthropometric indicators including height, weight, bilateral width of the ankle joints, bilateral knee diameter, pelvic width, bilateral pelvic depth, and bilateral leg length.
   4. Place 22 markers on key points of the patient based on the Davis protocol¹⁹: three markers on the trunk (7^th cervical vertebrae, shoulders on both sides); three markers on the pelvis (both sides of the anterior superior iliac spine and ankle joint); six markers on the thigh (bilateral femoral greater trochanter, femoral condyle, and middle point of femoral greater trochanter and femoral condyle on the same side); six markers on the calf (bilateral humeral head, lateral ankle joint, and middle point of humeral head and lateral ankle joint on the same side); four markers on the foot (the fifth metatarsal head and the heel on both sides) (Figure 1).
   5. Click on the Start button of the 3D gait analysis system, and make a new profile for the patient.
   6. Enter basic patient information and previously measured parameters.
2. Standing data acquisition
   1. Instruct the patient to maintain an upright position on the force plate for at least 3-5 s to gather the baseline data.
   2. Click on the Proc_Davis_Standing button to quickly check the position of the marker.
3. Walking task data acquisition
   1. Determine the random order of three walking tasks by drawing lots.
   2. Ask the patient to walk on the walking pass for five trials at a self-selected comfortable speed, which is marked as Task 0 (consider the single walking task as the Baseline task).
   3. Ask the patient to walk while holding a bottle of water on the walking pass for five trials at a self-selected comfortable speed, which is marked as Task 1 (simple dual-motor task).
      NOTE: Ask the patient to hold a 550 mL bottle of water in the unaffected hand while holding the arm position of the shoulder joint at 0° and elbow flexion at 90°.
   4. Ask the patient to walk across the line in the middle of the walking pass for five trials at a self-selected comfortable speed, which is marked as Task 2 (complex dual-motor task).
      ​NOTE: Place a soft ruler in the middle of the walking pass before Task 2 data acquisition.

4. Data processing and analysis

1. Select the middle three trials of each walking task to be processed to ensure the patient is stable.
2. Identify each gait cycle with two consecutive heel stride points on the same side.
3. Mark the toe-off point in each gait cycle²⁰.
4. Click on the Proc_DavisHeel+GI_AE button to compute the kinematic parameters of gait, as well as the computation of the Gait Performance Score (GPS) index.

5. Data extraction and statistical analysis of interest

1. Select region of interest parameters from the processed data, which include special-temporary parameters (stance phase, swing phase, single stance, double stance, cadence), joint angle parameters (trunk obliquity (frontal plane), trunk tilt (sagittal plane), trunk rotation (transversal plane), pelvic obliquity (frontal plane), pelvic tilt (sagittal plane), pelvic rotation (transversal plane), hip flex-extension, hip ab-adduction, hip rotation, knee flex-extension, ankle dorsi-plantarflexion, and GPS index.
2. Calculate DTC values based on the following formula^[10]:
   ([single-task gait velocity – dual-task gait velocity]/ single-task gait velocity) × 100 (1)
3. Perform the statistical analysis (see the Table of Materials) using the methodology described previously^20,21.
   1. Present parametric data as means and standard deviation if normally distributed or as medians if not.
   2. Use the paired t-test to compare the differences in kinematic parameters between patients in Task 1 and Task 2 conditions.
   3. Use one-way analysis of variance to compare three different tasks (Task 0, Task 1, and Task 2) of the kinematic parameters. Set statistical significance at P < 0.05.