Extensive training and hippocampus or striatum lesions: Effect on place and response strategies

Abstract

The hippocampus has been linked to spatial navigation and the striatum to response learning. The current study focuses on how these brain regions continue to interact when an animal is very familiar with the task and the environment and must continuously switch between navigation strategies.

Rats were trained to solve a plus maze using a place or a response strategy on different trials within a testing session. A room cue (illumination) was used to indicate which strategy should be used on a given trial. After extensive training, animals underwent dorsal hippocampus, dorsal lateral striatum or sham lesions. As expected hippocampal lesions predominantly caused impairment on place but not response trials. Striatal lesions increased errors on both place and response trials. Competition between systems was assessed by determining error type. Pre-lesion and sham animals primarily made errors to arms associated with the wrong (alternative) strategy, this was not found after lesions.

The data suggest a qualitative change in the relationship between hippocampal and striatal systems as a task is well learned. During acquisition the two systems work in parallel, competing with each other. After task acquisition, the two systems become more integrated and interdependent. The fact that with extensive training (as something becomes a “habit”), behaviors become dependent upon the dorsal lateral striatum has been previously shown. The current findings indicate that dorsal lateral striatum involvement occurs even when the behavior is spatial and continues to require hippocampal processing.

Introduction

When moving through the environment (e.g. foraging for food, avoiding danger, looking for a mate, returning home) animals must be able to flexibly use multiple navigations strategies. Different brain systems process the same environmental information in multiple parallel ways allowing for a variety of behavioral strategies (see [1], [2]). Two exemplar systems are the hippocampus system and the dorsal lateral striatum system. The hippocampus is necessary for processing spatial information. Rats with damage to either the hippocampus or inputs to the hippocampus are impaired in the spatial water maze [3], [4], win-shift radial arm maze [5], [6], and contextual fear conditioning [7], [8]. The dorsal lateral striatum is thought to mediate associations between a stimulus and a reinforced response [9], or between actions and outcomes [10]. Damage to the dorsal striatum causes impairments on egocentric maze learning tasks [11], visual cued discrimination [12], and win-stay radial arm maze [6], [13]. When initially learning to navigate through an environment, both rats [14] and humans [15], [16] start out using spatial strategy (using landmarks as cues) associated with activity in the hippocampus; though with increased training the strategy shifts to a motor response (using body turn or movement information) associated with activity in the striatum [17], [18], [19].

Two paradigms used to differentiate the strategy used by the hippocampus and striatum are the T-maze probe and the single strategy training task. The T-maze paradigm consists of training animals to find food at one end of a maze shaped like a “T”. Probe trials reveal whether the choice was based on spatial location (e.g. east arm) or motor response (e.g. right turn). Initially, rats use a spatial strategy, but after further training will switch to a motor response; this is true across days [14] or within a single training session [20]. T-mazes that probe animals for strategy use are based on a single trial where the animal has only two arms to choose from. However by giving the animal only two choices, any reduction in one strategy (e.g. place) would necessitate an increase in the choice of the other arm, indicating increased use of the alternative strategy.

The single strategy training paradigm examines how rapidly rats acquire either a place or response task in a plus maze. Inactivating a brain system makes it harder for animals to learn a task based on strategies linked to that system [3], [11]. There is also evidence of competition between systems, such that inactivation of one system may facilitate learning of the other strategy [6], [13], [21], [22], [23].

In contrast to the above findings regarding acquisition, less is known regarding the continuous interplay between the hippocampus and the striatum. It is plausible that in a familiar environment under natural conditions animals should be able to use multiple strategies depending on the situation. Packard and McGaugh's [14] data suggest a continuous interaction between systems, since inactivation even after task acquisition can change behavioral strategy. The present study examined the effect of lesions on a task where over-trained rats continuously switched between a spatial (place) and a motor response strategy [24]. Using a 4 arm plus maze allowed the animal to choose a goal arm that could be associated with a place strategy, a motor response, or an arm associated with neither strategy (Fig. 1a). In addition, animals' ability was examined on a spatial reference, spatial working memory, fixed motor-response and learning a new motor response.

Of current interest was the degree to which 1) the hippocampus (HIP) and striatum (STR) are involved in place and response navigation after a task is well learned, 2) lesions to one system enhance the use of the other intact system, 3) how these lesions affect new learning/reversals.