Lateral habenula integration of proactive and retroactive information mediates behavioral flexibility.

The lateral habenula (LHb) is known to play an important role in signaling aversive or adverse events that have happened or are predicted by cues under Pavlovian conditions. In rodents, it is also required for behavioral flexibility when changes in reward outcomes signal that strategies should be changed. It is not known whether the LHb also controls appetitive behaviors when an animal is able to utilize external cues proactively to guide upcoming decisions. In order to test this, male Long-Evans rats were trained to switch between two arms of a figure eight maze based on the tone presented prior to the choice. Importantly, the tones were switched every three to six trials so rats were able establish a response pattern before being required to switch. This caused rats to rely on both proactive (tones) and retroactive information (reward feedback) to guide behavior. Inactivation of the LHb with the GABA agonists baclofen and muscimol impaired overall performance by increasing both errors when the tones are switched (switch errors) as well as on subsequent trials (perseverative errors) indicating that both proactive and retroactive information are utilized by the LHb to guide behavioral flexibility. Once a correct choice was made in a given block, LHb inactivated rats did not make more errors than controls. A control study revealed that the LHb is not required for tone or reward magnitude discrimination per se. These results demonstrate for the first time that the LHb contributes to behavioral flexibility through utilizing both proactive and retroactive information when performing appetitive tasks.

Differential effects of estrogen on hippocampal- and striatal-dependent learning.

Estrogen's role in learning and memory may be to predispose animals to use specific cognitive strategies (Korol & Kolo, 2002). Specifically, estrogen may facilitate hippocampal-dependent learning, while at the same time attenuate striatal-dependent learning. As a stringent test of this hypothesis, place or response learning on an eight-arm radial maze was compared between ovariectomized (OVX) female Long-Evans rats and rats with chronic estrogen replacement (OVX+E; 5mg 17-beta estradiol 60-day release tablet). Reference and working memory errors were monitored separately for both place and response learning tasks. OVX+E rats learned the place task significantly faster than the response task, and faster than OVX rats. OVX rats required fewer days to reach criterion on the response task than OVX+E rats. Estrogen selectively enhanced reference memory performance, but only during place learning. The specific pattern of estrogen effects on learning suggests that future studies include verification of cognitive strategies used by animals.

Context-specific spatial representations by lateral septal cells.

To test whether the location coding of lateral septal cells is dependent on cue constellations, we examined single units in two different recording arenas on alternating days. Repeated recordings of lateral septal neurons in the same arena revealed that matching locations are encoded on separate days by about one third of the cells. The cells typically showed location-selective firing in only one of the two recording arenas and initially showed unrelated patterns when tested in a different recording arena. When tested for a second time in each recording arena, the initially dissimilar patterns were modified towards increased similarity between arenas. Simultaneously recorded hippocampal principal cells showed distinct place fields for each recording arena throughout the recording sequence. These results indicate that the initial reorganization of the lateral septal location coding may occur as a direct consequence of the hippocampal reorganization. Further septal reorganization is then partially independent of established place fields in the CA1 and CA3 area.Location-selective cells in cortical areas that receive projections from hippocampus proper (i.e. the subiculum and the entorhinal cortex) have not been shown to encode differences between recording arenas. Although some characteristics of this generalized coding scheme have also been found for location-selective lateral septal cells, the encoding of context information was generally preserved in the subcortical target cells of projections from the CA1 and CA3 area.