Click here to


Are you sure ?

Yes, do it No, cancel

Signal Processing Analysis of Breathing Rates Show Improved Spectral Coherence When Human Subjects Receive Musical Melody and Voice Instruction Guidance

J Bayouth1*, T Patton1 , S Gerard2 , W Shao2 , G Christensen2 , J Reinhardt2 , A Baschnagel1 , (1) University of Wisconsin, Madison, WI (2) The University of Iowa, Iowa City, IA


(Tuesday, 7/31/2018) 10:00 AM - 10:30 AM

Room: Exhibit Hall | Forum 9

Purpose: Clinically useful 4DCT requires patients to breathe consistently throughout the scanning study. The purpose of this work was to quantify patient respiratory pattern consistency when provided musical melody and voice instruction audio guidance.

Methods: Forty-five human subjects were accrued for an IRB-approved prospective clinical trial, receiving multiple helical 4DCT scans (Edge, Siemens). The external respiratory surrogate was an infrared reflective block (RPM, Varian). Subjects were instructed to follow a musical melody with breathing instructions directing them to breathe in and breathe out; the melody and voice instructions were set to 12.5-14 breaths per minute (bpm). Each subject was used as their own control. The trace data was acquired for > 60 seconds pre-scan, and during two sequential 4DCT scans. Their compliance to the directed respiratory pattern was determined using spectral coherence, the square of the cross-spectral density between the two traces, divided by the auto-spectral density of each individual trace. Identical signals produce a coherence of unity, unrelated signals have a coherence of zero. Variance of breathing period, inhale and exhale time, peak and valley position, and amplitude were also determined.

Results: The respiratory patterns of subjects during audio guidance were significantly more consistent than without instruction. The median coherence with musical melody and voice instruction (80%+/-19%) was superior to coherence without instruction (33%+/-22%, p<0.001). Median coherence was above 50% for 42 of 45 subjects during audio instruction, and for 5 of 45 without instruction. The standard deviation of the following parameters were all significant (p<0.02): breathing period (0.9sec vs. 0.4sec), inhale (0.4sec vs. 0.3sec) and exhale time (0.8sec vs. 0.4sec), and breathing amplitude (3mm vs. 1mm)

Conclusion: Signal processing analysis of human subject breathing rates during 4DCT show that musical melody and voice instruction guidance improved the consistency of patient respiratory amplitude, drift, breathing rate, and spectral coherence.


Gating, Image Artifacts, Motion Artifacts


TH- External beam- photons: Motion management (intrafraction)

Contact Email