transitlab

Harry Erwin talking about the types of behaviours he is modelling in silico

Last week I attended the second day of the Conference on Natural Computing and Intelligent Robotics at Sunderland University. This meeting was organised by NCAF (the Natural Computing Applications Forum)  meeting hosted by the Hybrid Intelligent Systems group at Sunderland.

An Approach to Developing a Library of Voltage-gated Channels by Harry Erwin (University of Sunderland).

This paper was a demonstration of the ability of a neuroscience-engineering collaboration to unravelling a complex process, in this case how sound is processed in the inferior collicular centre. The computational modeling arose out of the need to discover how the nerves in the IC worked. His group is using computational models to mimic the responses of four populations of nerve cells believed to be important in hearing in the IC.  Although not conversent with the algebra of the voltage-gated channels, I was interested in the explanatory potential of the models. One of the nice advantages over previous modelling of these channels was to anchor the models to DNA sequence terminology, thus providing a way to verify the in silico models in vivo. This would be by using genetic labeling technologies such as fluorescent proteins(see related articles), however it would only tell of the presence of the channel, not the activity.  This still needs to be addressed. His models provide predictions that can be checked in vitro, and give the engineers ideas of how to replicate hearing in robots. The insight that there is more than one behaviour in neurons in the IC, means that just using one type in robotic hearing is limiting.

From this and the other talks of the day, it is apparent that processing sound is computational intensive, having four populations of neurons that respond to different components of the sound means that immediate the processing is decreased. Each neuron ‘tags’ an input with sound, or vowel or consonant.

So what are scientists doing to reduce the amount of comutation needed? They look at simpler organisms, with smaller brains, to see if there are physical or neuronal structures that reduce the amount of computation needed. Lizards and Bats are being used for their use of sound in insect capture and object avoidance.

Related articles by Zemanta

• New calcium reporter for two-photon imaging in vivo
• Biomimetic designs strike again: Rescuebots