Even though Magnetic Resonance Imaging (MRI) exams are performed up to 16 times per every 100 inhabitants each year, patient comfort and acceptance rates are strongly compromised by exposure to loud acoustic noise. Here we present a system for acoustic noise cancellation using anti-noise derived from predicted scanner sounds. In this approach, termed predictive noise canceling (PNC), the acoustic fingerprint of an MRI system is obtained during a 60 s calibration, and used to predict anti-noise for arbitrary scan procedures. PNC achieves acoustic noise attenuation of up to 13 dB across a wide range of clinical MRI sequences, with spectral noise peak reduction of up to 96.76 % occurring between 0.6 and 1.2 kHz. These results suggest that predicted scanner noise can achieve substantial in-bore noise cancellation with the prospect of providing a cheap and scanner-independent solution for improved patient comfort.

中文翻译：

---

通过预测噪声消除降低 MRI 中的声学噪声负担

尽管每年每 100 名居民要进行多达 16 次磁共振成像 (MRI) 检查，但患者的舒适度和接受率仍会因暴露在巨大的噪音中而受到严重影响。在这里，我们提出了一种使用从预测的扫描仪声音中得出的抗噪声来消除声学噪声的系统。在这种称为预测噪声消除 (PNC) 的方法中，MRI 系统的声学指纹是在 60 秒的校准过程中获得的，并用于预测任意扫描过程的抗噪声。 PNC 在各种临床 MRI 序列中实现了高达 13 dB 的声学噪声衰减，0.6 至 1.2 kHz 之间的频谱噪声峰值降低高达 96.76%。这些结果表明，预测的扫描仪噪声可以实现显着的孔内噪声消除，并有望提供廉价且独立于扫描仪的解决方案，以提高患者的舒适度。