The time steps 1 5, 6 12, 13 15 correspond to the scene, delay, and response periods of the task, respectively, matching the relative timing of these periods in the actual location- scene association experiment (Wirth et al., 2003).
Surprisingly a hierarchical cognitive map representing object-scene associations by asymmetric connections is formed in a several second encoding period.
Such memory structure is not a simple trace of input sequence but organizes individual object-scene associations into the whole object arrangement similar to a human cognitive map [22].
On the other hand, regardless of whether they correspond to correct or incorrect location-scene associations, cells that lose the competition at a given trial receive or do not receive reward, depending on whether the behavioral response at that trial is incorrect or correct, respectively.
Here, it is shown that the learning-related spiking activity of some hippocampal neurons (changing cells; Wirth et al., 2003) in monkeys performing an associative learning task (location-scene association task; Wirth et al., 2003) is correlated with the reward-predictive value signal V t) generated by a TD learning algorithm that is trained on the same task.
Constant gigging and his surf-scene association is how he built his enormous following, but the Radio 1 A-listing of every single he has released has helped, too.
In the location-scene association task, this probability is 0.25.
The location-scene association task consists of learning arbitrary associations between complex visual scenes and superimposed target locations (Fig. 1).
This result suggests that the changing cells may have changed their activity before the estimated neural change points during the location-scene association task.
These observations suggest that the change in hippocampal neural activity leads behavior during the learning of the location-scene association task.
