In an animal model, the researchers found that auditory cortex neurons might encode sounds differently than previously thought. Belonging to the group of sensory neurons, they respond to the stimulus quite rapidly, while they are unfocussed in the beginning, but get more selective during the phase of the duration, so the old theory. In the beginning of the sound transition, things seem to be diffusely encoded across the neuron population, but as the study points out, sound identity is more accurately encoded, so the individual can identify sounds quicker and act on that information. The more neurons are active, the faster the individual can react to the information - exactly why it comes in handy that the neural populations involved spike most and encode the beginning of the stimuli most accurately.
The study involved the recording of individual neurons; to test it in humans, the team needed to use noninvasive techniques that gave them the insight in many neurons together. Event-related potential (ERP) techniques were used to record the brain signals and neural activity synchronized at the onset of the stimulus, while fMRI was used to reflect the activity averaged over several seconds. Both techniques revealed neural encoding of stimulus identity.
