Computational protocol: Vocal Tract Articulation in Zebra Finches

Similar protocols

Protocol publication

[…] One observation made on the X-ray videos is that OEC expansion is caused by a posterior-ventral movement of the hyoid skeleton. Therefore we posterior-ventrally displaced the hyoid skeleton in 0.5 mm steps to gradually increase OEC expansion and evaluate its filtering characteristics while playing frequency sweeps through the vocal tract of freshly sacrificed zebra finches. We did so for three different beak gapes.The birds used for this experiment were euthanized with an overdose of Nembutal (300 mg/kg body weight) in the pectoral muscle. Afterwards a small incision was made posterior from the lower jaw to expose the urohyal bone of the tongue apparatus. A cord was knotted around this bone which was later attached to a micromanipulator that could be moved in 0.5 millimeter steps. Subsequently, the syrinx and a part of the trachea were made accessible by dissecting the birds ventrally between the clavicles following the sternum. The trachea was intersected just above the splitting into the two primary bronchi and a short silastic tube which was fitted over the port of a small speaker (Knowles WBHC NB-68438C, Itasca, Illinois, USA) was inserted into the trachea so that the speaker was placed in the same position where otherwise the syrinx would have been . The dissected tissue was then agglutinated with tissue adhesive (Superglue 90–120 CPS, World Precision Instruments, Inc., Sarasota, Florida, USA) and the head of the bird was fixed in a stereotaxic device in such a way that the bill was positioned vertically. A thin metal wire (0.7 mm diameter) was stuck between the tips of mandible and maxilla and fixed with tissue adhesive to keep the beak gape constant. During the experiment acoustic measurements with three different beak gapes were taken whereas in the first series the beak was kept open at ca. 4.0 millimeters which represented a wide opening as observed on the X-ray videos only during some notes. In the second series the beak was kept open at 1.0 millimeter, a range frequently observed during natural zebra finch song. In the third series the beak was closed completely. Within each series the position of the hyoid skeleton was changed stepwise by displacing the urohyal bone ventrally in 0.5 millimeter steps in order to model the expansion of the OEC as observed on the X-ray videos. The maximal ventral movement of the urohyal bone varied between birds and series with a minimal displacement of 4.0 millimeters and a maximal displacement of 6.5 millimeters. The acoustic measurements took place in the sound-attenuating chamber described above. For every position of the tongue apparatus within all three series a linear frequency sweep (0.3 to 10 kHz in 1 second) constructed with PRAAT was played through the vocal tract of the birds using a sound card (CDX-01 CardDeluxe, Digital Audio Labs, 1266 Park Road Chanhassen, MN 55317). The sound emitted from the beak was then recorded with a Sennheiser MKH50 microphone vertically directed at the beak from 3 cm distance and immediately recorded in PRAAT with the same sound card (44.1 kilosamples/s, 16 bit resolution). After the experiment we checked for every bird whether the speaker was still attached to the trachea, which was the case for all five birds. To ensure that differences between spectra of recorded sweeps were caused by differences in articulation and not by position-dependent filtering due to remaining room resonances, we took care that the exact position of both the microphone and the bird preparation did not change between recordings. We also measured speaker output at approximately the same position where the beak was during recordings in order to correct for frequency response deviations of the speaker system by subtracting the dB values of the speaker output from the measured spectrum. Although remaining resonances might still affect the data slightly this impact can be considered rather small and does not change the general results. The data were analyzed by calculating the long-time average spectrums (Ltas function in PRAAT; 100 Hz bin width) of the recorded sound sweeps, and comparing them between different articulatory states. The latter was done using custom-written scripts in the scientific computing environment SciPy version 0.7 . […]

Pipeline specifications

Software tools MUSCLE, SciPy
Databases THE-DB
Applications WGS analysis, Nucleotide sequence alignment
Organisms Homo sapiens, Dipturus trachyderma, Taeniopygia guttata