Inferring Elapsed Time from Stochastic Neural Processes

Many perceptual processes and neural computations, such as speech recognition, motor control and learning, depend on the ability to measure and mark the passage of time. However, the neural mechanisms that make such temporal judgements possible are unknown. A number of different hypotheses have been advanced, all of which depend on the known evolution of a neural or psychological state, possibly through oscillations or the gradual decay of a memory trace. We suggest a new model, which instead exploits the fact that neural and sensory processes, even when their precise evolution is unpredictable, exhibit statistically structured changes. We show that this structure can be exploited for timing, and that reliable timing estimators can be derived from the statistics of the processes. This framework of decoding time from stochastic processes allows for a much wider array of neural implementations of time estimation than has been considered so far, and can simultaneously emulate several different behavioral findings, which so far have only been understood in psychological terms.