Processing of prospective and retrospective duration estimates in medial prefrontal cortex

We, humans, created clocks to measure time and invented maps to navigate to locations. However, these tools only assist our natural capabilities for processing temporal and spatial information as we guide our lives through time and space – capabilities, that we share with other species. However, our bodies are not equipped with a sensory organ for the passage of time. Time is ultimately not a material object of the world for which we have a unique receptor system as we have ears for sound or eyes for light, with respective processing stages in the brain. The perception of duration – the interval between two successive time points of events – is essential for survival. Animals must integrate durations to reach sources of food or find mating partners or adapt fundamental cognitive processes such as decision-making and planning of action. The perception of time is not associated with specific sensory pathways but uses a highly distributed system in the brain. The medial prefrontal cortex appears to be one prominent member for time perception and duration processing. Cells in the frontal cortex exhibit climbing neural activation as a potential neural mechanism for the representation of duration. Climbing activity has been widely associated with mnemonic functions in temporal information processing. To get a better understanding of how duration estimation as a temporal integration process in the supra-second range is represented in the rodent brain, I performed in-vivo extracellular recordings in the medial prefrontal cortex of behaving Mongolian gerbils (Meriones unguiculatus). Specifically, I sought to find out if characteristic effects of magnitude estimation, like range effect, regression effect, and scalar variability, are captured by the encoding system. I designed and established a time estimation task, which required subjects to judge duration in a retrospective and prospective manner: First, subjects had to measure the experience of an unexpected time stimulus in hindsight and second reproduce this duration of experience in time passing. Experiments were performed in a virtual reality behavioral setup in which subjects traveled along a virtual linear corridor. Locomotion was tracked with an optical sensor-equipped treadmill. I showed that the behavioral paradigm I used is effective in reproducing known behavioral effects, such as range effect, regression effect, and scalar variability in experiments with humans. I succeeded in using the timing task paradigm with gerbils and gained comparable results to humans. My experiments demonstrate that gerbils can learn and perform complex time estimation experiments that are more intense in cognitive processing than time discrimination tasks or experiments based on peak procedure paradigms. With recordings in gerbil medial prefrontal cortex, I showed that prefrontal neurons respond with various patterns of neural activity to the passage of time. However, relevant patterns were mostly observed for prospective time estimation while reproducing a known time stimulus until a future point in time than for retrospective time estimation, where duration had to be judged for past events. The majority of prefrontal neurons responded by ramping neuronal activity. In doing so, the size of the time stimulus was encoded in gradually adapted total discharge rate, or by gradually adapted ramping speed in a generalized activity pattern until reaching a unified threshold. The adaptation of ramping speed was put into effect by temporally scaling the response pattern, notably towards the end of time reproduction. The applied scaling followed the prediction of stimulus sizes, yet with minor oversizing. Hence, a key discovery was that no matter the neurons’ response, the rate at which they adjusted their activity depended on the time interval required. I also investigated whether cells encoded the accuracy and precision of the duration estimate after the end of a trial, where animals were informed about their performance via visual feedback and appetitive reinforcement. On a populational level, the adaptation of firing rates to the level of accuracy demonstrated that regression effects, resulting from strategies to cope with uncertainty about sensory information, are represented by neuronal activity. Precision of duration estimates was represented by the magnitude of firing as well, although to an internal reference instead of an absolute external value. Using higher firing rates to provide higher information content for less regression and low variance corroborates the fact that the objective of timed behavior is maximal accuracy and minimal variance. Therefore, my results demonstrate that the medial prefrontal cortex in rodents profoundly provides timer functions at an internal clock stage. However, the prefrontal cortex does not exclusively code for time but also fulfills mnemonic functions by integrating the outcome and interaction of the decision with the environment. This integration might serve to compare the present outcome with previously stored values in memory. Activity during delay phase supports the idea that the prefrontal cortex concurrently acts as a memory stage integrating prior knowledge and updating posterior knowledge about the stimulus duration in accordance with Bayesian inference