Action potentials carry nerve signals in individual neurons and somehow encode information. My question is if it is possible to read that information carried by the nerve signal and feed it into a computer to understand what information it contains?
Neural info is contained in spike trains, mainly in terms of spike frequency and related parameters. These spike trains in individual neurons can be measured by techniques like patch clamp and Ca2+ imaging. These techniques are all pretty sophisticated and computer assisted for sure.
However, by recording of action potentials in e.g. the optic nerve, we cannot possibly know what the person is seeing. Nonetheless, by cleverly mapping responses of conglomerates of neurons we can make estimates of what the information carried in the nerve fibers mean. For example:
- Mapping the responses of an ensemble of place neurons in the hippocampus has been successfully used to predict where a mouse is exactly located in a maze (Guger et al., 2011);
- Gross potentials measures with EEG can be used to move artificial limbs, as shown in physically handicapped people. In essence, by mapping out the responses of motor areas in the brain, it can be predicted where a person is thinking their limb is moving.
- Guger et al., Front Neurosci (2011); 5 85