How the brain stores information for short periods of time

The authors of the study, now published in the journal Nature Neuroscience, investigated the persistent activity in a hindbrain circuit responsible for eye movements in zebrafish larvae. This circuit, the so-called oculomotor system, gives the command for rapid eye movement by way of special nerve cells that produce a short-lived succession of action potentials. On the one hand, this "burst of fire" reaches the neurons responsible for movement in the eyes and triggers a "saccade," a rapid movement of the eye. On the other hand, it is also transmitted to a second cell population, the so-called neural integrator for eye movements, where the speed signal is integrated mathematically and a position signal is created. This signal is then transmitted to the motor neurons, thus producing -- in fish as well as in humans -- a stable eye position following the rapid eye movement. The neural integrator keeps up this signal for several seconds, until a new saccade is initiated.

The persistent activity in the neural integrator for eye positions is never perfect, as the eyes gradually drift back to their point of rest after a saccade. The authors thus had the possibility of measuring the dynamics of the system during spontaneous eye movements in the dark and testing the model without the measurements being distorted by saccade commands or visual feedback.

The authors discovered that, contrary to previous belief, the cells of the neural integrator for eye movements do not constitute a homogeneous population and that existing models for explaining persistent activity in the oculomotor system will have to be reconsidered. The scientists demonstrated that the integrator neurons do not posses a uniform dynamics and that the neurons are distributed in the hindbrain with the help of their integrator time constants.

These findings provide new evidence on the organization and functioning of circuits with persistent activity and suggest a potential explanation for their low susceptibility to failure. The study is an important milestone in the quest of network neuroscience to explain the functioning of local circuits and thus close the gap between the functioning of a single neuron and the production of behavior.