Flexible spatial learning requires both the dorsal and ventral hippocampus and their functional interactions with the prefrontal cortex

Abstract

When faced with changing contingencies, animals can use memory to flexibly guide actions, engaging both frontal and temporal lobe brain structures. Damage to the hippocampus (HPC) impairs episodic memory, and damage to the prefrontal cortex (PFC) impairs cognitive flexibility, but the circuit mechanisms by which these areas support flexible memory processing remain unclear. The present study investigated these mechanisms by temporarily inactivating the medial PFC (mPFC), the dorsal HPC (dHPC), and the ventral HPC (vHPC), individually and in combination, as rats learned spatial discriminations and reversals in a plus maze. Bilateral inactivation of either the dHPC or vHPC profoundly impaired spatial learning and memory, whereas bilateral mPFC inactivation primarily impaired reversal versus discrimination learning. Inactivation of unilateral mPFC together with the contralateral dHPC or vHPC impaired spatial discrimination and reversal learning, whereas ipsilateral inactivation did not. Flexible spatial learning thus depends on both the dHPC and vHPC and their functional interactions with the mPFC.

Keywords: executive function; hippocampus; learning; memory; prefrontal cortex; spatial memory; temporal lobe.

FIGURE 1

Experimental design. (a) Spatial discrimination and reversal task. In each session, rats learned an initial spatial discrimination (ID; e.g., “go East”) followed by a series of spatial reversals (R1, R2, etc.; e.g., “go West”, “go East”, etc.) in a plus maze. (b) Rats were trained on the spatial reversal task and then implanted with cannulae targeting bilateral mPFC and either bilateral dorsal HPC or bilateral ventral HPC. Ovals show dorsal view of rat head, with cannulae indicated by circles. The effects of inactivating targeted brain areas on initial spatial discrimination and reversal learning were then tested. Targeted brain areas were injected with saline or muscimol 20 min before spatial reversal task performance. C, Targeted brain areas. Filled circles indicate injection sites. Bilateral (1) mPFC, (2) dHPC or (3) vHPC; (4) ipsilateral or (5) crossed mPFC-dHPC; (6) ipsilateral or (7) crossed mPFC-vHPC

FIGURE 2

Cannula locations. Black circles indicate histologically verified cannula tip locations for injections targeting mPFC (left), dHPC (middle), and vHPC (right), shown on coronal plates from Paxinos and Watson (2007). AP coordinates relative to bregma are shown

FIGURE 3

Bilateral dHPC or vHPC inactivation impaired spatial learning. Inactivating either HPC subregion (a) decreased choice accuracy and (b) increased the number of trials performed and (c) errors during initial discrimination learning. (d) Inactivating either HPC subregion increased both incorrect goal arm and incorrect start arm entries. Goal and start arm error data are pooled across ID and R1. Plots show means ± SEMs. A set of asterisks indicates a significant difference between saline and muscimol (significant ANOVA main effect of drug) over the conditions spanned by the bar. *p < .05, **p < .01, ***p < .001

FIGURE 4

Bilateral mPFC inactivation impaired spatial reversal more than initial discrimination learning and selectively increased goal arm errors. More so during the first reversal than the initial discrimination, muscimol (a) decreased choice accuracy and (b) increased number of trials performed and (c) errors. (d) Muscimol increased incorrect goal arm entries, but had no effect on incorrect start arm entries. Plots show means ± SEMs. A set of asterisks indicates a significant difference between saline and muscimol (significant post hoc t-tests (a and c) or ANOVA main effect of drug (b and d)) over the conditions spanned by the bar. *p < .05, **p < .01, ***p < .001

FIGURE 5

Crossed, but not ipsilateral, inactivation of the mPFC and either the dHPC or vHPC impaired spatial learning. Crossed inactivation (a) decreased choice accuracy and (b) increased numbers of trials performed and (c) errors. (d) Crossed, but not ipsilateral, inactivation increased incorrect goal arm entries. A small increase in start arm errors was detected across crossed and ipsilateral inactivation. Plots show means ± SEMs. A set of asterisks indicates a significant difference between saline and muscimol (significant post hoc F-test (a–d left) or ANOVA main effect of drug (d right)) over the conditions spanned by the bar. *p < .05, **p < .01, ***p < .001