Adaptive speech enhancement using directional microphone in a 4-T MRI scanner

Area of Science:

• Biomedical Engineering
• Acoustics
• Signal Processing

Background:

• Magnetic Resonance Imaging (MRI) generates significant acoustic noise, hindering patient-operator communication.
• Effective communication is crucial for patient comfort and cooperation during MRI procedures.
• Existing noise reduction methods may interfere with speech intelligibility.

Purpose of the Study:

• To evaluate an adaptive speech enhancement (ASE) system for MRI environments.
• To reduce MRI scanning noise while preserving speech signal integrity.
• To ensure clear communication between patients and MRI operators.

Main Methods:

• Developed an ASE system utilizing differential directional microphones.
• Implemented two-stage adaptive filters with normalized least mean square algorithms.
• Tested the system with echo planar imaging (EPI) and gradient echo multi-slice (GEMS) sequences on a 4T MRI scanner.

Main Results:

• Achieved 1.4 dB and 3.3 dB speech enhancement for EPI and GEMS noise, respectively, via cepstral distance assessment.
• Demonstrated noticeable speech recovery and clear waveform post-ASE treatment.
• Observed additional noise reduction when combining ASE with spectral subtraction (SS).

Conclusions:

• The proposed ASE system shows significant potential for mitigating MRI acoustic noise.
• Combining ASE with spectral subtraction enhances communication effectiveness.
• This approach can improve the patient experience and operational efficiency in MRI settings.