Background: In Intensive Care Unit (ICU) patients, dyspnea is frequent, severe and exerts unfavorable effects on the short, medium and long term. Detection and quantification rely on the patient's self-reporting abilities. However, more than half of the patients in the ICU are unable to report their sensations. Therefore, the risk is to miss the evaluation of dyspnea and the potential benefits associated with its control. Observational scales, based on physiological and behavioral changes related to dyspnea (such as the Mechanical Ventilation Respiratory Distress Observational Scale MV-RDOS), are promising alternative tools for the detection of dyspnea in non-communicating patients. However, their routine use is not standardized, is not supported by any recommendation, but above all, relies largely on the subjective observation of the facial expression of fear or the abdominal paradox. There is a need for alternatives to the visual analogue dyspnea scale (D-VAS) for the detection of dyspnea in non-communicating intubated patients. Analysis of brain cortical activity modifications during dyspnea could be an alternative to the dyspnea self-report (D-VAS) in the ICU and could improve the performance of observational dyspnea scales.

Hypothesis: 1) dyspnea during a spontaneous breathing trial (SBT) is associated with premotor cortex activation identifiable using functional Near-Infrared Spectroscopy (fNIRS); 2) replacing the items "abdominal paradox" or "facial expression of fear" by HbO2 level could improve the performance of the MV-RDOS to predict dyspnea in non-communicating intubated patients; 3) HbO2 level change identified using fNIRS performs well in predicting SBT outcome

Patient management will not be altered by the study. Patients will be included if the clinician in charge of the patient has decided to perform a SBT on the patient. SBT conditions will be those of current practice. The methods used to measure dyspnea will be those used in current practice in our department (D-VAS and MV-RDOS).

1. Types of measures and techniques used The presence of dyspnea will be defined by a positive response to at least 2 of the following questions: "Does participant feel short of breath?"; "Does participant feel short of air?"; "Is participant's breathing difficult?"; "Does participant have difficulty breathing?". The intensity of dyspnea will be measured by the D-VAS in communicating patients. The dyspnea-VAS will also define patients with clinically significant dyspnea ("D-VAS" > 3) or non-clinically significant dyspnea (D-VAS ≤ 3). Measurement of dyspnea by the MV-RDOS scale will be performed in all patients, and clinically significant dyspnea will be strongly suspected by the MV-RDOS value ≥ 2.6.

   Surface EMG of the extra-diaphragmatic inspiratory muscles (Alae Nasi and Parasternal) will be collected via self-adhesive surface electrodes (ECG Electrods, HG91TSG 48x34mm Kendall/Arbo, Covidien, Dublin, Ireland). Bilateral recording of the parasternal muscles will be performed by a pair of electrodes placed in the second intercostal space near the sternum. The recording of the Alae nasi muscles will be performed by placing an electrode on each nostril. Electrical signals of inspiratory muscle activity will be retrieved using the Labchart Peak Analysis MLS380/8 module to extract the root mean square (RMS) of the EMG (RMS-EMG). This envelope of the inspiratory RMS-EMG signal will be used to calculate the maximum EMG amplitude (EMGmax) and its area under the curve (EMGAUC). To minimize artifacts related to ECG activity, the parasternal EMG signal will be filtered before the RMS averaging process, using a low-pass filter (50-400 Hz).

   Electroencephalographic activity will be measured with an active electrode system comprising 30 electrodes positioned according to the international EEG 10-20 system, referenced to Fcz (EEG/NIRS device, Artinis Medical Systems®, Einsteinweg, The Netherlands). The impedance of the electrodes will be kept below 5 kΩ. The signals will be amplified and digitized at a frequency of 1000 Hz.

   Cerebral perfusion will be assessed using a 27-channel fNIRS tool (EEG/NIRS device, Artinis Medical Systems®, Einsteinweg, The Netherlands). This fNIRS device uses two wavelengths of near-infrared light (695 and 830 nm) to measure relative changes in oxyhemoglobin and deoxyhemoglobin at a sampling rate of 10 Hz. The transmitter and detector optodes are placed 3 cm apart. The cortical areas between each pair of transmitters and detectors are called channels. Anatomically, the channels correspond to the cortical regions located 2-3 cm below the surface of the skin and scalp. The optodes are placed on the forehead and scalp, with the lowest optodes placed along the T4-Fpz-T3 line, defined by the 10/20 system. The fNIRS signals will be processed as described by Schecklmann et al. Oxyhemoglobin, deoxyhemoglobin, and total hemoglobin are derived from the optical densities using the modified Beer-Lambert law. Corrective factors will be applied to remove motion artifacts. An average oxyhemoglobin and deoxyhemoglobin waveform will be generated for each channel. The regions of interest being the premotor cortical areas (supplementary motor area).

   SBT failure is defined by the occurrence and persistence for at least 5 minutes of one of the following criteria: SpO2 (pulsed oxygen saturation) ≤ 90% or PaO2 (partial oxygen pressure) ≤ 50 mmHg with FiO2 (Inspired oxygen fraction) ≥ 50%, PaCO2 (partial pressure of carbon dioxide in arterial blood.) > 50 mmHg, pH < 7.32, respiratory rate > 35/min, heart rate > 140/min, systolic blood pressure > 180 mmHg or < 90 mmHg.

2. Sequence of experimental steps Patients will be placed in a ventilatory weaning test with 0 cmH2O of pressure support and 0 cmH2O of end-expiratory pressure. A quantification of dyspnea will be performed for all patients using the MV-RDOS score and by Dyspnea VAS (D-VAS) for communicative patients before the start of SBT, every 10 minutes during SBT and at the end of SBT. Recordings of respiratory movements, airway flow, EMG and ECG will be made at the same time, per 10-minute period, before the start of the weaning test, during the 30-minute weaning test and 10 minutes after the end of the weaning test.

3. Statistical analysis The two groups clinically significant vs. non-clinically significant dyspnea will be compared on their brain activation indices (HbO2 and HbR) using the non-parametric Mann-Whitney test for continuous variables and the Chi2 test for categorical variables. The two SBT "success" vs. "failure" groups will be compared on their brain activation indices, using the same modalities. The correlations between their brain activation indices and dyspnea intensity, between their brain activation indices and EEG, and between their brain activation indices and surface EMG will be tested using Spearman's correlation coefficient. The performance of the brain activation indices and the modified MV-RDOS in predicting dyspnea or SBT failure will be estimated by calculating the area under the curve of ROC curves. An observed difference will be considered significant if the probability "p" of a type I error is ≤ 0.05.

Inclusion Criteria:

• Invasive mechanical ventilation > 24h
• Ability to realize a spontaneous breathing trial, defined by the following criteria: Effective cough, no excessive tracheo-bronchial secretions, resolution of the disease that prompted the intubation, heart rate ≤ 140/min, pressure systolic blood pressure between 90 and 160 mmHg, no or very low dose of amines, SpO2 > 90% in 40% FiO2 (or PaO2/FiO2 > 150), positive expiratory pressure < 8 cmH2O, respiratory rate ≤ 35/min, tidal volume > 5ml/kg of theoretical body weight, respiratory rate / tidal volume ratio < 105 cycles/min/L
• Decision by the clinician in charge of the patient to perform a SBT,
• Patient or relative consent obtained.

Exclusion Criteria:

• Minor patient,
• Pregnant or breastfeeding woman
• Agitation/sedation: RASS ( Richmond Agitation-Sedation Scale) > 2 or < 2,
• Person under legal protection (guardianship, curators), safeguarded by justice.

Any intensive care patient undergoing invasive mechanical ventilation deemed suitable for a ventilatory weaning test

The procedures carried out with the French data privacy authority (CNIL, Commission nationale de l'informatique et des libertés) do not provide for the transmission of the database, nor do the information and consent documents signed by the patients.

Consultation by the editorial board or interested researchers of individual participant data that underlie the results reported in the article after deidentification may nevertheless be considered, subject to prior determination of the terms and conditions of such consultation and in respect for compliance with the applicable regulations.

Beginning 3 months and ending 3 years following article publication. Requests out of these time frame can also be submitted to the sponsor

Researchers who provide a methodologically sound proposal.