Emergence in Neuron Spikes

Jan Engelbrecht's group explores connections between organizing principles in physics and principles underlying computation in neurons.

 I.  With John Hopfield (Princeton) we study simple, biologically inspired neuron network models that compute through the development of synchrony in action potential timing.  For instance, in olfaction the development of synchronous firing patterns in different subsets of neurons can separate object and background odors.  Coincident spiking of particular neurons triggers appropriate "grandmother" neurons to fire.

  II.  With Rennie Mirollo (BC, Mathematics) we study how a spiking model neuron entrains to an oscillatory rhythm.  Frustration due to the competition between a neuron's firing rate and the rhythm period causes rich pattern formation.  Viewed as a phase transition, the dynamics of how a neuron locks to a drive exhibits model- independent, scaling in successive interspike intervals (ISI's) denoted by f_n-f_n-1.  Universality arises from "bottleneck" behaviour in the map that determines the spike times f_n = F(f_n-1).
  III.  With Chuck Stevens and Dimitri Klichko (Salk Institute) as well as Mike Hasselmo and Motoharu Yoshida (BU), and my student Kristie Loncich, we seek to demonstrate similar scaling in real neurons.  Initial recordings on CA1 pyramidal cells in hippocampal slices are very promising.