Price spikes have long been a feature in energy markets; but five-sigma spikes have also become more common in equity markets in the past three years than at any time since the 1940’s. Spiking has significant implications for pricing and risk management; particularly since spikes tend to cluster. Such price behaviour has made life hard for users of trend following and momentum models. However; on-going research using spiking neural networks may be able to bridge the gap between the cognitive models of neuro-science and the more mainstream deep neural network models currently popular in machine-learning; by using biologically-realistic models of neurons for financial market prediction. This talk will examine such issues; report some preliminary results and suggest directions for future research.