The investigators hypothesized that increasing non intentional leaks could increase work of breathing and could lead to patient/non-invasive ventilation (NIV) asynchrony.

The main objective is to compare ventilatory pattern (work of breathing, flow, pressure) under NIV with and without non-intentional leaks in 10 awakened healthy subjects and 10 awakened and asleep obese hypoventilation syndrome (OHS) patients.

Methods: While the subjects will be under NIV, several levels of leaks will be simulated in a random order with an automatically opening valve. Breathing pattern (work of breathing,flow, pressure, thorax and abdominal movements) will be recorded by Polygraphy. Healthy subjects will be recorded only during awakened state. OHS patients will be recorded both during awake and sleep sates.

Analysis: A repeated measures analysis of variance (ANOVA) will compare work of breathing according to the different levels of leaks.

Inclusion Criteria:

• OHS patients, in stable stade, teated with nocturnal non invasive ventilation for at least three months.
• Healthy subjects without known pathology

Exclusion Criteria:

• OHS patients: instable cardio-respiratory status and/or acute cardio-respiratory failure three month before the inclusion. Latex allergy, oesophageal varices.

10 healthy subjects recruited in professionnal area 10 patients with Obesity hypoventilation syndrome addressed to the sleep laboratory for clinical assessment