Publications updated 12/12/13.
Hearing relies on the brain’s ability to represent auditory information. We use multiple approaches to investigate this function in owls and chicken. Birds are not just fascinating creatures but offer advantages for understanding the neural coding and computations underlying hearing.
Computations underlying auditory
spatial tuning: Models of human sound localization materialized in the owl’s
The plausibility of
sound-localization models remains in question. Our studies in barn owls
directly address this issue. In the owl’s brain, the signals collected by the
ears are compared to infer sound direction. This computation underlies the
highly efficient ability of owls to catch prey in darkness. In fact, the owl’s
brain performs many of the operations proposed in sound localization models.
Cross-correlation describes the response of neurons to the extent that theorems
apply; averaging-like processes reduce noise; multiplication underlies the
selectivity of space-specific neurons. Thus, a bird’s brain possesses
tremendous power to understand neural computations. The question how the
evolutionary pressure for optimality has driven the owl’s brain to achieve
elegance in a mathematical sense guides our long-term plans.
Interaural time and level differences determine the receptive field of ICx cells in auditory space.
(modified from Pena and Konishi 2002)
Coding sound identity together with
space and time
The auditory system encodes both the
identity and location of sounds. The spectral features over time are essential
for identifying sounds, such as in vocalizations. However, this information
must be encoded in parallel with cues about sound direction, such as differences
in time and level between the ears. Our studies examine how information flows
and how enhanced selectivity for ‘what’ and ‘where’ emerges in the ascending
STRFs in the cochlear nuclei predict spike-time reliability
(modified from Steinberg and Pena 2011)
Brain representation of auditory
space: The biased owl
owls can very accurately localize sounds near the center of gaze, they
underestimate the direction of sources in the periphery. This behavioral
bias is also observed in other animals and in humans. This behavior and the
underlying neural implementation can be predicted by statistical inference;
Brian Fischer showed that the mapping of auditory space in the owl’s midbrain
could explain how statistical inference takes place. These conditions are
likely in other cases, as shown by studies of the oblique effect in visual
perception in humans. To perform statistical inference, it is critical that the
brain represents the relationship between sensory information and the
environment, as well as the statistics of the environment. We plan to elucidate
how this happens in collaboration with Brian Fischer and Terry Takahashi.
Non-uniform spatial tuning and surround bias in the owl's midbrain Representation of space predicted by Bayesian theory
(modified from Knudsen 1982 and Wang et al 2012) (modified from Fischer and Pena 2011)
Techniques used in the
We are interested in recording techniques that may help us address scientific
questions. We have used in vivo intracellular
recordings to investigate space selectivity in the owl’s brain and in vivo cell-attached recording in auditory coincidence detectors to overcome the
difficulty in isolating coincidence detector neurons. We developed in vitro recordings in chicken to bridge
our systems approach and cellular and synaptic mechanisms.
In vitro recording in the chicken midbrain.
(modified from Penzo and Pena 2009)
Intracellular labeling and in vivo recording of a space-specific neuron.
(modified from Pena and Konishi 2001)
In vivo physiology in free-field
We built a hemispheric speaker array to more efficiently
manipulate auditory-scene context and map receptive fields in space and time. The speaker array consists of 144 speakers, with 10 degree resolution at the densest portion of the array (towards the center). The stereotax is now mounted on top of a turn table so we can orient the owl's heading in any direction in the azimuth.
Speaker array for sound stimulation in free field Pupillary dilation reflex recovers with deviant stimulus.
In vivo imagining using fiber optic endoscope
collaboration with Kazuo Funabiki we set up in
vivo deep-brain imaging in the lab. Using a bundle of nano fiber optics that can be inserted deep into the
brain we plan to monitor the activity of clusters of neurons in the bird’s
brain. Calcium imaging using Oregon green labeling will allow us to visualize neural activity in real time.
Endoscope consists of a bundle of nano optic fibers and an electrode. Labeled neurons visualized with the endoscope.