Our brains formulate models of how the world works by integrating information about our visual, olfactory, tactile, and auditory environments into an existing body of knowledge. Algorithms based on the neural processes that perform these tasks should, in principle, be a useful strategy for making sense of the vast amounts of data generated in the fields of genomics, proteomics, metabonomics, and metabo-lomics. Neuronal networks that incorporate temporal coding have many useful characteristics. Because synchronous oscillations in the visual system scale with the size of the object, they can act as a context-dependent noise filter. The presence of a feature is indicated by identifying the neurons that respond to that feature, with the firing rate being proportional to the contrast or local signal strength. Contextual factors, such as the presence of cocircular elements, are encoded by the timing of the neuron's spikes. A network that includes the visual, olfactory, and hippocampal memory systems could encode a set of biomarkers and discern novel associations between different sets of markers. One of the key lessons learned from the brain that could be useful for bioinformatics is the possibility of binding data from these different sources to gain insight into the fundamental questions of biology.
Was this article helpful?