Weaning from mechanical ventilation (MV) is a complex process in which patients are liberated from the ventilator. Prolongued weaning and weaning failure, defined as the need for reintubation, have different adverse effects, including prolongation of MV, intensive care unit (ICU) and hospital stay, and are also associated with increased incidence of ventilator-associated pneumonia and high mortality. The rate of weaning failure is high, even when the classic extubation criteria are met, so it is necessary to improve scores that allow predicting and determining the ideal time for MV withdrawal. The aim of this study is to design a new multimodal index to predict and optimize weaning results in a personalized way, based on the use and interpretation of data derived from continuous monitoring of critically ill patients. The new multimodal index, in addition to classical respiratory parameters, will include parameters related to patient-ventilator interaction (asynchronies), diaphragmatic function, cardiovascular status and autonomic nervous system function (ANS). The investigators have designed a prospective observational study that will include 126 critical patients from a medical-surgical ICU that meet the classical criteria for weaning. The management of the patients, as well as the weaning process, will be carried out following the usual protocol. In addition to the classical weaning predictor data, data on the patient-ventilator interaction and the function of the autonomic nervous system will be collected by means of specific software (BetterCare). Cardiovascular and diaphragmatic function will be evaluated using ultrasound. Based on the advanced analysis of data from different devices collected throughout the mechanical ventilation period, it will be designed a personalized weaning score that should improve the accuracy of the decision-making process and therefore reduce morbidity and mortality. Additional benefits would include lowering health care costs without increasing adverse events.