Biomedical Engineering Reference
In-Depth Information
9.4 Respiration
The duets between horneros are a spectacular display of complex rhythms.
However, many birds show a wide variety of intersyllabic times. How could
these be generated? This section contains a speculative model of their gener-
ation, inspired by anatomical descriptions of neural nuclei.
9.4.1 Periodic Stimulation for Respiratory Patterns
In the case of the oscine songbirds, much is known about the structure of the
nuclei involved in the generation of song. However, much of the description
of the avian nervous system focuses on the forebrain nuclei and pathways
[Wild 2004]. The extratelencephalic projections are described as if they were
simple followers of the activity in the forebrain nuclei. In fact, this is unlikely
to be true, since brain-stem nuclei are receptors of the sensory feedback which
informs the song system of the requirements for air. Brain-stem nuclei receive
both information from the telencephalic nuclei and sensorimotor feedback,
and are therefore a place where very interesting dynamics could be expected.
Horneros are suboscines. These birds are believed (after work with the east-
ern phoebe [Kroodsma and Konishi 1991]) to develop normal song without
auditory feedback. They also seem to lack the telencephalic nuclei widely
studied in songbirds. They are expected to have a nucleus DM, projecting to
the nucleus XIIts, as well as to respiratory nuclei. The respiratory pathways
are expected to show similar structures in oscines and suboscines.
Wild and others [Wild 2004, Sturdy et al. 2003] have described the path-
ways involved in the control of respiration during quiet respiration and
singing. The respiratory rhythm is generated or conveyed by the rostral
nucleus of the ventrolateral medulla (RVL), which innervates the retroam-
bigualis (RAm) nucleus. This nucleus projects to expiratory motor neurons.
On the other hand, the nucleus PAm projects to inspiratory motor neurons,
and receives feedback from sensors which update the nervous system about
the dynamics of the air sacs (see Fig. 9.7a). The details of the connections
are not yet known, but on the basis of this partial list of observations we can
build a computational model in order to explore the possible solutions that
this respiratory system can display. Interestingly enough, the effects emerging
out of this architecture can be present in both oscines and suboscines.
9.4.2 A Model
Recently, Trevisan and coworkers [Trevisan et al. 2005] proposed a simple
model to translate these anatomical observations into a computational model.
In it, they described the air sac dynamics in terms of a geometric variable
measuring the variation
x
of the air sac volume at atmospheric pressure.
The sacs are modeled as damped masses, driven by expiratory and inspi-
ratory muscles (a departure from what happens with mammals, where the
