The AID2GAIT project aims to develop a biofeedback system with the aim of improving the outcomes of robot-assisted gait training (RAGT) in pediatric patients with cerebral palsy. The physiological signals of children during RAGT therapy sessions, acquired through non-invasive technologies, will be analyzed. These technologies specifically are: * wearable technology (smartwatch), from which the HRV (Heart Rate Variability) signal will be measured; * infrared thermography, from which the temperature in salient facial regions will be obtained; * fNIRS (functional near-infrared spectroscopy), from which information on brain activity and its changes over time will be obtained. Information on the kinematics of the exoskeleton used during RAGT will be extracted. The RAGT will be performed using the Lokomat orthosis (Hocoma), the most widely used exoskeleton in rehabilitation that facilitates a bilaterally symmetrical gait, as the individual actively tries to advance each limb during walking, combined with a patented dynamic body weight support system.
The entire RAGT treatment will consist of 12 sessions of robotic gait rehabilitation administered over 4 weeks. The protocol consists of a block paradigm: children will be asked to actively move during the robotic training for the first 30 seconds, followed by a rest period of 30 seconds. The paradigm will be provided in 10 blocks for a total duration of 30 minutes. During each single session, infrared (IR) thermal video, smartwatch signals (HRV- heart rate variability) and the kinematic output of the robot will be recorded. During the first (T1) and the last session of the RAGT (T12), fNIRS will also be recorded together with the assessment of motor skills, revealed by clinical scales. The assessment of fNIRS and clinical scales will be useful to understand the global efficacy of the treatment. Before the start of the experimental trials, parents will be extensively informed about the purpose and protocol of the study and will sign an informed consent form. For each session, the assessment of the psychophysiological state of the patients will be based on the estimation of the state of physiological parameters, such as heart rate variability (HRV), recorded by a smartwatch and on the assessment of the emotional state of the child by means of an infrared thermal imaging system. The information on the state of the child and the robot will be assessed in real time and will constitute the input data for a machine learning-based model capable of classifying the level of patient engagement. Based on this information, the physiotherapist, who assists the child during the training sessions, will be able to intervene and modify the parameters of the exoskeleton (e.g. push force, body weight support, treadmill speed, range of motion and hip and knee offset). Furthermore, the effectiveness of the entire treatment will be assessed through the administration of clinical tools commonly used in clinical practice and by the assessment of brain activation by means of a non-invasive and portable neuroimaging technique, the fNIRS. These assessments will be conducted by comparing the first and the last training session.
Study Type
OBSERVATIONAL
Enrollment
11
Biofeedback based system development to enhance Robotic Assisted Gait Training in cerebral palsy pediatric patients
Fondazione Centri di Riabilitazione Padre Pio Onlus - Gli Angeli di Padre Pio
San Giovanni Rotondo, FG, Italy
RECRUITINGAnalysis of variations in the motor skills of patients from T0 (1st RAGT) to T2 (12th RAGT), assessed by GMFM88 clinical scale
Assessment by Gross Motor Function Measure (GMFM-88): it measure changes in GMF with intervention through 88 items. The scoring range is from 0 to 3 with higher scores meaning a better outcome. The clinical scale is administered before RAGT#1 (i.e. T0) and after RAGT#12 (i.e. T2).
Time frame: One month (4 weeks) from the beginning of the 1st RAGT (T0)
Analysis of variations in the motor skills of patients from T0 (1st RAGT) to T2 (12th RAGT), assessed by MAS clinical scale
Assessment by Modified Ashworth Scale (MAS): The test is performed by extending the patients limb's first from a position of maximal possible flexion to maximal possible extension. The scoring range is from 0 to 4 with lower scores meaning a better outcome. The clinical scale is administered before RAGT#1 (i.e. T0) and after RAGT#12 (i.e. T2).
Time frame: One month (4 weeks) from the beginning of the 1st RAGT (T0)
Analysis of variations in the motor skills of patients from T0 (1st RAGT) to T2 (12th RAGT), assessed by WeeFIM clinical scale
Assessment by Functional Independence Measure for Children (WeeFIM): it measures the need for assistance and the severity of disability. Scores range from 0 to 4, with lower scores indicating higher levels of disability. The clinical scale is administered before RAGT#1 (i.e. T0) and after RAGT#12 (i.e. T2).
Time frame: One month (4 weeks) from the beginning of the 1st RAGT (T0)
Analysis of variations in the motor skills of patients from T0 (1st RAGT) to T2 (12th RAGT), assessed by PEDSQL clinical scale
Assessment by Pediatric Quality of Life Inventory (PedsQL): modular approach to measuring health-related quality of life. The items of the four Scales (Physical Functioning, Emotional Functioning, Social Functioning, and School Functioning) are grouped together. The scoring range is from 0 to 4 with lower scores meaning a better outcome. The clinical scale is administered before RAGT#1 (i.e. T0) and after RAGT#12 (i.e. T2).
Time frame: One month (4 weeks) from the beginning of the 1st RAGT (T0)
Measurement of heart rate variability of patients
During RAGT #1(i.e T0), RAGT#6 (i.e. T1), RAGT#12 (i.e. T2) heart rate variability (HRV) measurement will be performed by a smartwatch. The HRV signal will be monitored and metrics such as mean, standard deviation, root mean square, power spectral density will be extracted in time windows of 30 seconds. The above mentioned parameters are correlated with autonomous nervous system activity and will be used as input data for a AI based model able to classify the patient's engagement level during RAGT.
Time frame: Two years from the beginning
Measurement of fNIRS of patients
During RAGT #1(i.e T0), RAGT#6 (i.e. T1), RAGT#12 (i.e. T2) functional near infrared spectroscopy measurement will be performed by a fNIRS cap. fNIRS allows to measure oxy- and deoxyhemoglobin oscillations in the frontal, prefrontal and motor cortex areas. Given the ecological nature of the experiment, a GLM-based algorithm will be applied to automatically identify the onset and duration of cortical activations. The canonical GLM metrics (beta-values and t-statistics) indicative of brain activity will be evaluated in time windows of 30 seconds and averaged across all good channels. The above mentioned parameters are correlated with central nervous system activity and will be used as input data for a AI based model able to classify the patient's engagement level during RAGT and to assess neural plasticity in patients.
Time frame: Two years from the beginning
Measurement of infrared imaging of patients
During RAGT #1(i.e T0), RAGT#6 (i.e. T1), RAGT#12 (i.e. T2) infrared imaging (IRI) measurement will be acquired by an infrared camera. The thermal signals in salient areas of the face (nose tip, nostrils, corrugator, chin, and perioral area) will be monitored (tempertures in °C), and metrics such as mean, standard deviation, kurtosis, skewness and LF (\[0.04-0.15\] Hz), HF (\[0.15-0.4\] Hz), and LF/HF components will be extracted in time windows of 30 seconds. The above mentioned parameters are correlated with autonomous nervous system activity and will be used as input data for a AI based model able to classify the patient's engagement level during RAGT.
Time frame: Two years from the beginning
Measurement of exoskeleton kinematics
During RAGT #1(i.e T0), RAGT#6 (i.e. T1), RAGT#12 (i.e. T2) the torques relative to both knees and both hips will be acquired (in Nm). Metrics such as mean, standard deviation will be extracted in time inetrvals of 30 seconds. The above mentioned parameters are correlated with the participation level and with the activity of the patient during RAGT and will be used as input data for a AI based model able to classify the patient's engagement level during RAGT.
Time frame: Two years from the beginning
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.