Abstract

The time taken to react voluntarily to a stimulus is far longer than can be accounted for by ordinary processes of nerve conduction and synaptic delay, and varies unpredictably from trial to trial. Though random, the distribution of reaction times usually follows a relatively simple law, which in turn can be explained by the LATER model, in which a decision signal, representing belief in the existence of the target, rises in response to incoming sensory evidence from an initial value to a criterion level at which action is initiated. But the rate of rise fluctuates randomly from trial, to an extent that cannot be explained by sensory noise at the input. These conclusions, confirmed by recording from neurons in the frontal eye fields, suggest that the randomness of reaction time is due to a deliberate, gratuitous, neural process which confers certain biological advantages as well as having implications for general ideas about the nature of ‘free will'.