Martin Paul Nawrot (AG Neuroinformatics, Institute for Biology - Neurobiology, Free University of Berlin, Germany)

The Bernstein Center Freiburg

Bernstein Seminar

Martin Paul Nawrot

AG Neuroinformatics
Institute for Biology - Neurobiology
Free University of Berlin, Germany

 

Dynamics of odor processing and
reward prediction in the honeybee brain

Abstract:
Behavioral studies have demonstrated that bees can recognize learned odors that are presented for only 200ms, and they express the learned behavior within less than 500ms after stimulus onset. Thus, odor processing, memory recall, and behavioral decision making must be completed within a short period of time. In a first study (Krofczik, Menzel & Nawrot, 2009) we investigated neural correlates of odor processing in intracellular recordings from antennal lobe projection neurons (PNs). We found that the neural population signifies odor identity within approx. 50-150ms (rate code). Information about odor identity is also represented in the spatio-temporal pattern of odor responses (latency code).
In a second study we model the neuronal response dynamics by spike-frequency adapting (SFA) neurons which can explain the dominant phasic-tonic responses in PNs and temporal sparseness in the 3rd order olfactory neurons, the mushroom body Kenyon Cells (KCs). The latter result is in contrast to existing models that explain KC sparseness.
In a third study (Strube-Bloss, Nawrot & Menzel) we analyzed extracellular single unit recordings from mushroom body extrinsic neurons (ENs) during an associative conditioning paradigm. ENs read out converging inputs from the KCs and provide the mushroom body output. The population of ENs reliably encodes odor-reward associations 150 ms after stimulus onset during retention tests after a 3h consolidation phase (LTM), but not during training (STM).
We conclude that odor encoding and reward prediction represent rapid neural processes. We hypothesize that the neural output of the mushroom body is used by a downstream network to form decisions and to initiate the appropriate behavioral response. In ongoing research projects we device models of spiking neural networks designed to accomplish odor processing, the formation of associative memories, and the expression of learned behaviors.

• Krofczik S, Menzel R and Nawrot MP (2009) Rapid odor processing in the honeybee antennal lobe network.
  Frontiers in Computational Neuroscience 2:9
• Strube-Bloss M, Nawrot MP, Menzel R (2010) Mushroom Body Output Neurons Encode Odor-Reward Associations.
  Journal of Neuroscience (in press)

Host: Stefan Rotter