Search

Animals smelling in the real world use a small number of receptors to sense a vast number of natural molecular mixtures, and proceed to learn arbitrary associations between odors and valences. Here, we propose a new interpretation of how the architecture of olfactory circuits is adapted to meet these immense complementary challenges. First, the diffuse binding of receptors to many molecules compresses a vast odor space into a tiny receptor space, while preserving similarity. Next, lateral interactions “densify” and decorrelate the response, enhancing robustness to noise. Finally, disordered projections from the periphery to the central brain reconfigure the densely packed information into a format suitable for flexible learning of associations and valences. We test our theory empirically using data from Drosophila. Our theory suggests that the neural processing of olfactory information differs from the other senses in its fundamental use of disorder.

The N -body problem in a double well requires new features for quantum information processing, macroscopic quantum superposition, and other fundamental studies of quantum many body physics in ultracold atoms. One needs (a) tilt, and (b) to go beyond the single-particle ground state in each well, i.e., to two or more energy levels. For (a), we show that a small potential difference between the wells, or tilt, causes the decoherence of Schrodinger cat states. However, these states reappear when the tilt can be compensated by atom-atom interactions; these tilted cat states constitute partial cats that are protected from decoherence by the many body wavefunction. For (b), we provide explicit criteria for when two energy levels are needed to describe the state space. For typical experimental parameters, two levels are indeed required for creation of cat states., Comment: 4 pages, 3 figures