Respiratory muscle dysfunction is highly prevalent in patients with prolonged weaning from mechanical ventilation and is strongly associated with weaning failure. Efforts to strengthen the respiratory muscles, aimed at reversing or minimizing the impact of respiratory muscle weakness on clinical outcomes, have generally focused on the diaphragm with specific inspiratory muscle training (IMT) exercises. However, the effectiveness of these exercises and impact on clinical outcomes are not current practice in the majority of ICUs, as they are hardly feasible in ICU patients who often cannot be disconnected from the ventilator and cannot fully cooperate.

Promising results have been published concerning non-respiratory training techniques, which can also target the accessory muscles, particularly important in the presence of increased load to the respiratory system, as in the case of the weaning phase. These non-respiratory training techniques would have the advantage of not entailing disconnection of the patient from the ventilator. In particular, in healthy subjects, a quasi-isometric neck contraction, called neck flexion, appeared to generate greater or comparable recruitment of some principal and accessory muscles of respiration, when compared to conventional IMT. However, this has not been studied in patients requiring prolonged mechanical ventilation, for whom IMT with threshold loading devices remains the primary recommended rehabilitation strategy.

Therefore, the primary aim of the investigators is to assess the feasibility, tolerability, and safety of neck flexion and to compare them with IMT technique in patients with difficult and prolonged weaning from mechanical ventilation. Secondary aims are: i) to characterize which respiratory muscles are recruited and their level of activation at different levels of ventilatory assistance and ii) to assess which respiratory muscles are recruited and their level of activation during the two techniques and to compare these findings.

The hypothesis of the investigators is that neck flexion will be feasible (more than conventional IMT), well tolerated, and safe in patients with difficult and prolonged weaning. The investigators also hypothesize that, reducing the level of assistance and during unassisted breathing, a progressively increasing activation of the diaphragm, neck and trunk respiratory muscles, reflecting increased ventilatory load, will be fund. Finally, the hypothesis of the investigators is that the level of muscle activation/recruitment during neck flexion will be comparable or even greater to that occurring during IMT, as found in healthy subjects.

Finding a new and highly feasible rehabilitative technique, able to recruit and train the respiratory muscles (including accessory muscles), will have the potential to promote patients' weaning and improve all related clinical outcomes, and therefore to dramatically shift the paradigm about the role of rehabilitation in ICU.

The investigators will conduct a prospective longitudinal pilot study in patients receiving mechanical ventilation via endotracheal tube or tracheostomy.

The investigators will perform a set of baseline measurements (as early as possible after the patient reaches the ability to spontaneously trigger the ventilator).These measurements include: I) ultrasound measurements (thickness and thickening fraction) of the diaphragm, parasternal intercostals, sternocleidomastoid and of the abdominal muscles; II) surface electromyography (sEMG) of the diaphragm, sternocleidomastoid, scalene, and of the abdominal muscles. Ultrasound and sEMG measurements will be performed during the current level of ventilation (decided by the clinical team), during minimal level of ventilatory support and/or during unassisted breathing; III) maximal inspiratory pressure (MIP); IV) maximal voluntary contraction (MVC) of neck flexion.

As soon as patients can be disconnected from the ventilator I) IMT maneuvers and II) neck flexion maneuvers, will be started and performed once a week (each time in a randomized order), until patients are successfully weaned or for a maximum of 4 weeks. The investigators will also perform: III) ultrasound measurements; IV) surface electromyography of the muscles above specified during each IMT and neck flexion maneuver; V) MIP and MVC; these latter measurements will be repeated weekly and inspiratory muscle training and neck flexion intensity level targeted accordingly.

Inclusion Criteria:

• Patients receiving invasive MV via endotracheal tube who had failed a planned extubation or failed at least three spontaneous breathing trials (SBTs) or receiving MV via tracheostomy (who required tracheostomy because of difficult or prolonged weaning, failed extubation and/or prolonged ventilation) and able to spontaneously trigger the ventilator.
• Over or equal to 16 years of age
• Tolerating levels of pressure support of 10 cmH2O or lower (or equivalent support in NAVA or PAV+) for at least 15 minutes
• In stable clinical and hemodynamic conditions and adequate level of oxygenation (cardiac frequency ≤ 140 beats/minute, systolic blood pressure 90-160 mmHg, no or minimal vasopressors, PaO2/FiO2 over or equal to 150 mmHg)
• Able to understand and follow simple verbal instructions

Exclusion Criteria:

• Patients with a previously diagnosed severe neuromuscular disorder (such as amyotrophic lateral sclerosis, muscular dystrophy, multiple sclerosis, myasthenia gravis or spinal cord injury)
• Patients with chronic respiratory failure already ventilated before ICU admission
• Patients unable to collaborate or understand instructions

The two training techniques will be randomized using sealed opaque envelopes.