Temporal information encoding in isolated cortical networks.

Abstract

Time-dependent features are present in many sensory stimuli. In the sensory cortices, timing features of stimuli are represented by spatial and temporal code. A potential mechanism by which cortical networks may perform temporal-to-spatial conversion is "reservoir computing." The state of a recurrently-connected network, or a reservoir, represents current and prior inputs. In this experimental study, we determined whether the state of an isolated cortical network represents temporal information in the inputs. We used patterned optogenetic stimulation of dissociated primary rat cortical cultures. We delivered input sequences of patterns where one of the patterns occurred at different times in each sequence. The state of these experimental networks contained information about input sequences for at least a second, with at least 100-ms precision. Accurate classification required many neurons, suggesting that timing information was encoded via population code. Trajectory of the network state was largely determined by spatial features of the stimulus, with temporal features having a more subtle effect. The duration of spatial information retention was > 2 s, similar to duration of short-term memory in the primary visual cortex. We concluded that local reservoir computation may be a plausible mechanism for temporal-to-spatial code conversion in sensory cortices.