1 Introduction [1] [2,3] [4,5] [6, 7, see also 8] [9–11] [5] [12] [13,14] [15] A [16] [17] [7] [7,9,18] N d [19,20] d [21] [22] 2 Methods 2.1 Subjects ad libitum 2.2 Apparatus In Experiment 1, rewarded alternation took place on an elevated (∼1 m above the floor), low-walled wooden T-maze consisting of a start arm (80 cm long; 10 cm wide) joined to two identical goal arms (60 cm × 10 cm), with each arm surrounded by a 1 cm ridge. In Experiment 2, rewarded alternation took place on an elevated Y-maze (∼80 cm above the floor). Each of the three arms (50 cm × 9 cm) was surrounded by a 0.5 cm ridge and extended from a hexagonal central platform (14 cm diameter). The T- and Y-mazes contained stainless-steel food wells at the far end of each goal arm and each maze was located in a well-lit testing room containing prominent extra-maze cues such as wall posters. 2.3 Procedure In both experiments, rats were trained pre-operatively. Habituation to the maze took place over a five-day period. For the first three days rats were placed on the maze in pairs and left to explore and collect food rewards for 10 min. During days four and five, the rats were placed individually on the maze and allowed to explore freely for five minutes. By this stage all the rats ate from the food wells at the ends of the arms. Training on the rewarded alternation task then followed. Each trial in rewarded alternation had two runs: in the first, one of the goal arms was blocked, allowing the rat to enter only the other goal arm, whereupon it received 1 food pellet (45 mg Rodent Diet Formula A/I, Noyes, Lancaster, NH). During the second run, the block was removed and the rat was placed on the maze and given a free choice of either arm. Rats received two pellets for choosing the previously unvisited arm (i.e. for alternating). Choosing the arm previously visited in the sample run yielded no reward. The time between the sample and the choice runs was approximately 10 s (Experiments 1 and 2a) or 30 s (Experiment 2b). Left/right allocations for the sample and choice runs were pseudo-randomised over ten trials per day, with no more than three consecutive sample runs to the same side. The inter-trial interval was ∼3–4 min. Prior to surgery, each rat was trained on the task until they achieved a criterion of at least 80% correct alternation over two consecutive days. n n n Rats were allowed a minimum of five days post-surgical recovery after which they received training in the absence of any infusions. Rats failing to score 80% correct alternation on these sessions were given additional training. 2.4 Drugs 4 6 2 2 2 d On infusion days, the aliquots were defrosted and used to back-fill the microinjectors, which were connected via polyethylene tubing to a pair of Hamilton syringes (10 μl) driven by a microsyringe pump (SP250i, World Precision Instruments, England). Each rat was restrained firmly in a towel, the stylets were removed, and the microinjectors inserted through the guide cannulae into the hippocampus. Drug or vehicle (0.5 μl per side) was infused into the dorsal or ventral hippocampus at a rate of 1 μl/min (i.e. for 30 s), and the microinjectors were kept in place for an additional 60 s to allow the drug to diffuse. Infusions into each hemisphere were made simultaneously and were given 15 min prior to testing. In Experiment 1, rats received five infusions in total (saline and four doses of muscimol: 0.3 μg/side, 0.15 μg/side, 0.075 μg/side, 0.0375 μg/side). In Experiment 2a, rats received three infusions (PBS, AP5: 2.95 μg/side, muscimol: 0.15 μg/side) and in Experiment 2b, the same rats received a further 4 infusions (PBS × 2, AP5: 2.95 μg/side × 2), making seven infusions in total. Infusions were given in a counterbalanced, pseudo-random order with at least 48 h between each infusion. In addition, performance was assessed again 24 h after each infusion and rats failing to score 80% or more on these drug-free sessions received extra training before the next infusion was administered. 2.5 Histology On completion of the experiment, rats were perfused transcardially and their brains were removed, sectioned, and stained with cresyl violet for verification of the injection sites. 3 Results 3.1 Experiment 1 Fig. 1 t t p t p Fig. 2 2 5 S 23 F 1,21 p F 4,84 p F 4,84 p [23] F 1,21 p F 1,21 p F 1,21 p p p F 4,18 p F 4,18 Fig. 2 2 5 S 23 F 1,21 F 4,84 F 4,84 3.2 Experiment 2 Fig. 1 [21] 2 S 9 F 2,16 p p p Fig. 3 F 1,7 p F Finally, performance accuracy (90%) during drug-free testing carried out after the last infusion confirmed that there was no lasting damage due to cannulae implantation or the microinjection process. 4 Discussion [7,18] N d [24,25] [26] c-fos [27] [28] [29] [30] [29] [6] [9] Fig. 2 [7,9] [24,25] [5] [21] [19] [LTP; 31, 32] [20,33] [21] [19] [34] [35] In conclusion, the present research has confirmed and extended previous findings from lesion studies, namely that the dorsal hippocampus has a greater involvement in spatial working memory than the ventral hippocampus. Furthermore, NMDAR activation within the dorsal hippocampus makes an essential contribution to this aspect of hippocampal information processing.