Reducing the effects of background noise during auditory functional magnetic resonance imaging of speech processing: qualitative and quantitative comparisons between two image acquisition schemes and noise cancellation.

Purpose: The intense sound generated during functional magnetic resonance imaging (fMRI) complicates studies of speech and hearing. This experiment evaluated the benefits of using active noise cancellation (ANC), which attenuates the level of the scanner sound at the participant's ear by up to 35 dB around the peak at 600 Hz.

Method: Speech and narrowband noise were presented at a low sound level to 8 listeners during fMRI using 2 common scanning protocols: short ("continuous") and long ("sparse") temporal schemes. Three outcome measures were acquired simultaneously during fMRI: ratings of listening quality, discrimination performance, and brain activity.

Results: Subjective ratings and discrimination performance were significantly improved by ANC and sparse acquisition. Sparse acquisition was the more robust method for detecting auditory cortical activity. ANC reduced some of the "extra-auditory" activity that might be associated with the effort required for perceptual discrimination in a noisy environment and also offered small improvements for detecting activity within Heschl's gyrus and planum polare.

Conclusions: For the scanning protocols evaluated here, the sparse temporal scheme was the more preferable for detecting sound-evoked activity. In addition, ANC ensures that listening difficulty is determined more by the chosen stimulus parameters and less by the adverse testing environment.