Role of the Hippocampus in Instrumental Conditioning

Learning

The Role of the Hippocampus in Instrumental Conditioning

Laura H. Corbit and Bernard W. Balleine

The rational for the study is to shed light on a fundamental process occurring in instrumental learning that has not been well-researched. The authors hypothesize that instrumental learning consists of stimulus response (SR) processes and action-outcome (AO) processes. SR associations are encoded into procedural memory, whereas AO associations are encoded into declarative memory. The authors cite research that has stated that declarative memory is hippocampal dependent. Thus, the authors hypothesize that damage to the hippocampus should interfere with AO associations while not interfering with SR associations. In experiment 1A rats with dorsal hippocampal lesions and sham lesions (the IV) are tested for their sensitivity to outcome devaluation with a satiety treatment protocol (the DV) after being trained to press levers for two types of reinforces. The hypothesis was that lesioned rats would demonstrate a weakened devaluation effect. There was no effect in experiment 1A; in other words lesioned rats demonstrated a clear devaluation effect.

In experiment 1B the rats were exposed to different tasting reinforcers as some research has demonstrated that the incentive value of an outcome can be mediated by a "motivationally arbitrary stimulus" like taste. So the experiment was performed on a new group of rats, this time with outcomes that differ only in taste (grape vs. orange kool-aid). In the previous experiment the outcomes different in taste and texture as the two reinforcers were a liquid and a pellet, whereas in this experiment only different tasting liquids were used. Again no difference was observed. The authors speculated that perhaps AO processes in instrumental conditioning might be controlled by another brain area such as the amygdala.

However, not to be dissuaded, the authors then cite research that indicates that perhaps the hippocampus lesion rats can encode AO outcomes in training they may be insensitive to the causal consequences of their actions. Thus, the rats in experiment 1A were retrained and then one of the two outcomes was degraded. The hippocampal rats were expected not to demonstrate the same strong extinction effects on the degraded response compared to the sham lesion group. The only difference this time was that the lesioned rats demonstrated no overall difference in lever presses for degraded vs. non-degraded reinforcers during extinction trials whereas sham rats demonstrated markedly higher presses for non-degraded stimuli. Thus, this was viewed as supporting the hypothesis that the lesion rendered the rats insensitive to the casual consequences of their actions. So, according the authors, the hippocampus is important in instrumental learning in assessing the casual effects of their actions.

First, I find the overall reasoning that AO associations are encoded into declarative memory and SR into procedural memory interesting but not applicable to the population being studied in this paper. Rats have no declarative/procedural memory designation (they are incapable of declarative memory); therefore it is debatable that rates are capable of AO associations along the lines that these authors suggest. Rats perform based on SR associations and the authors' findings are more compatible with this than with their explanation. Since rats have no declarative memory capabilities the hypotheses and conclusions of this experiment cannot be based on theories of human memory. The hippocampus in rats does not encode aspects of declarative memory. Moreover, the first two experiments merely indicate that sated animals do not respond well to reinforced contingencies based on that domain; even extremely demented Alzheimer's patients or animals with extensive hippocampal damage will not eat or respond to food rewards when sated.