Patient machine interface for the control of mechanical ventilation devices.

Area of Science:

• Neuroscience
• Biomedical Engineering
• Respiratory Medicine

Background:

• Mechanical ventilation (MV) is crucial for patients with respiratory failure, but current interfaces offer limited control.
• Brain Computer Interfaces (BCIs) are typically used for motor control, leaving potential applications in other clinical areas unexplored.
• Non-invasive methods are preferred for MV due to its transient nature and patient conditions.

Purpose of the Study:

• To explore the potential of non-invasive Brain Computer Interfaces (BCIs) for controlling mechanical ventilators (MV).
• To propose a novel Patient Ventilator Interface (PVI) enabling ventilator control via brain activity during varying conscious states.
• To investigate the use of electroencephalography (EEG) signals for direct brain-ventilator communication.

Main Methods:

• Developed a Patient Ventilator Interface (PVI) schema for controlling ventilators.
• Utilized scalp-recorded electroencephalography (EEG) signals as a non-invasive communication pathway.
• Analyzed EEG data to discriminate between inspiration and expiration periods during voluntary breathing.

Main Results:

• Demonstrated the feasibility of using non-invasive EEG signals for controlling a ventilator.
• Achieved 92% accuracy in discriminating inspiration and expiration periods from EEG data in a healthy subject (10-fold cross-validation).
• Presented a functional schema for the proposed Patient Ventilator Interface (PVI).

Conclusions:

• Scalp-recorded EEG signals offer a viable non-invasive communication channel for controlling mechanical ventilators.
• The proposed Patient Ventilator Interface (PVI) represents a novel approach to assist patients requiring mechanical ventilation.
• Further research is needed to address the advantages and obstacles for this BCI application in clinical settings.