Home

About

Research

Events

Contact

print

Banbury Center Workshop Series

Neurocomputational Strategies: From Synapses to Behavior (1998)
Functional Organization of the Thalamus and Cortex and their Interactions (1999)
Toward Animal Models of Attention and Consciousness (2000)
Persistent Neural Activity (2000)
Can a Machine Be Conscious? (2001)
Communication in Brain Systems (2004)
Neurobiology of Decision-Making (2005)
Computational Approaches to Cortical Functions (2006)
New Frontiers in Studies of Nonconscious Processing (2007)
Theoretical and Experimental Approaches to Auditory and Visual Attention (2008)
Searching for Principles Underlying Memory in Biological Systems (2009)
Neuronal Response Variability and Cortical Computation (2011)
Connections and Communication in the Brain (2014)

Alex Bäcker

Learning Pattern Recognition In The Olfactory System

California Institute of Technology

The olfactory system is the oldest and most primitive of sensory modalities, whose function is to serve as a general-purpose pattern recognition device to relate sensory experiences to those stored in memory. The exact pattern of activation of olfactory receptors, however, is seldom repeated. This implies that recognition has to be robust to 'slight' changes in odor composition. I will show that such robustness is not present in the naïve brain, but rather is developed as a function of exposure to an odor.

I will show a novel type of plasticity that, contrary to the adaptation that the olfactory system is better known for, exhibits a sensitization selective to low odor concentrations in response to exposure to the same odor at high concentrations. Finally, I will demonstrate that, by lowering neurons' odor detection thresholds, this effect enhances the robustness of responses to changes in odor concentration, providing specific predictions for behavioral experiments. Finally, I will show that the duration of this priming effect suggests that it may act to provide a bias to the system that lowers the signal to noise ratio needed to classify a percept as an odor, using information across multiple odor puffs to lower false negatives and improve recognition by tuning specifically to an odor likely to be encountered.

Then, if time allows, I will show an algorithm, related to the Sellers algorithm for evolutionary distances between genetic sequences, that allows us to 'recognize' what odor was presented to an animal in any given trial by looking at the neural responses that occur in response to the odor and show very recent work extending the algorithm described above to the case of multiple neurons and prove a theorem enabling the problem to be solved in a fraction of the time it used to take to solve it.