CNBC Brain Bag
Title: Constraining neuronal networks using micro-technologies: Towards directionality in networks
Presenter: Ashwin Vishwanathan
Location: MI 3rd Floor Social Room
Additional Information: Please RSVP to Linda Moya (lhmoya@cmu.edu) by Friday, October 2, and indicate if you require a vegetarian meal.
Abstract: Persistent neuronal activity occurs in working memory and motor planning tasks. It has been proposed that reverberatory activity is maintained in such networks by virtue of recurrent connections. However, cellular mechanisms underlying such reverberatory activity remain unclear. Recently, it was shown that such reverberatory activity occurred in small neuronal networks. The synaptic mechanism underlying such reverberatory activity suggests an important role played by asynchronous calcium release. Indeed, upon suppressing asynchronous calcium release a significant decrease in reverberatory activity was reported. However in the above studies the networks were allowed to connect randomly. To understand the process further, these networks need to be made simpler. By constraining networks we hope to reduce the dimensionality of the network to a linearized culture. In this talk I will present a new method that we have developed to evoke directionality in networks.  Using this setup, we wish to study the properties of reverberatory activity and better understand the synaptic mechanisms that underlie this phenomenon.

Seminar
Title: Using the Past to Anticipate the Future: Detecting and Discriminating Unpredictable Communication Signals
Presenter: Leonard Maler, Ph.D.
Location: 1495 Biomedical Science Tower
Abstract: Sensory input can be classified in many ways. One key distinction is between signals involved in communication versus those related to environmental features. A second, less common classification, is into predictable versus unpredictable signals. Electric fish must both detect and discriminate signals related to navigation, prey capture and communication. The appearance of prey is unpredictable and the fish must first detect prey and then estimate its location. Electric communication signals can be either predictable or unpredictable. I will discuss how the predictable component of communication signals prepares, via feedback, sensory neural circuits to detect the unpredictable component.  As in many other species, specific electric signals are used in an aggressive context while others are used during courtship. The aggressive signals are stereotyped and can only be detected by population neural activity. In contrast, the courtship signals are variable and our analyses suggest that a small neural population can discriminate among them.  We conclude that males will use the rate of aggressive signals to decide on whether to fight or flee. Females, in contrast, can estimate the quality of the courtship signals in order to choose the fittest male suitor. Lastly I will provide evidence that spike bursts are used for detecting both prey and aggressive electric signals, while patterns of isolated spikes are used for estimating prey location and discriminating among courtship electric signals. I will address the cellular and network basis for emitting spike bursts or isolated spikes for specific signal types.