Cells of the rodent medial entorhinal cortex (EC) possess cellular properties hypothesized to underlie the spatially periodic firing behaviors of 'grid cells' (GC) observed in vivo. Computational models have simulated experimental GC data, but a consensus as to what mechanism(s) generate GC properties has not been reached. Using whole cell patch-clamp and computational modeling techniques this thesis investigates resonance, rebound spiking and persistent spiking properties of medial EC cells to test potential mechanisms generating GC firing. The first experiment tested the voltage dependence of resonance frequency in layer II medial EC stellate cells. Some GC models use interference between velocity-controlled oscillators to generate GCs. These interference mechanisms work best with a linear relationship between voltage and resonance frequency. Experimental results showed resonance frequency decreased linearly with membrane potential depolarization, suggesting resonance properties could support the generation of GCs. Resonance appeared in medial EC but not lateral EC consistent with location of GCs. The second experiment tested predictions of a recent network model that generates GCs using medial EC stellate cell resonance and rebound spiking properties. Sinusoidal oscillations superimposed with hyperpolarizing currents were delivered to layer II stellate cells. Results showed that relative to the sinusoid, a limited phase range of hyperpolarizing inputs elicited rebound spikes, and the phase range of rebound spikes was even narrower. Tuning model parameters of the stellate cell population to match experimental rebound spiking properties resulted in GC spatial periodicity, suggesting resonance and rebound spiking are viable mechanisms for GC generation. The third experiment tested whether short duration current inputs can induce persistent firing and afterdepolarization in layer V pyramidal cells. During muscarinic acetylcholine receptor activation 1-2 second long current injections have been shown to induce persistent firing in EC principal cells. Persistent firing may underlie working memory performance and has been used to model GCs. However, input stimuli during working memory and navigation may be much shorter than 1-2 seconds. Data showed that input durations of 10, 50 and 100 ms could elicit persistent firing, and revealed time courses and amplitude of afterdepolarization that could contribute to GC firing or maintenance of working memory.

Le courant cationique activé par l'hyperpolarisation (courant H) est impliqué dans l'organisation temporelle de l'activité neuronale. Nous montrons que le courant H améliore la synchronisation et la régularité des oscillations thêta dans l'hippocampe et dans le néocortex in vitro. Il détermine les oscillations thêta par la définition de résonnance intrinsèque de membrane, la fidélité de décharge et le couplage entre potentiels postsynaptiques et décharge. Les cinétiques du courant H sont modulés par l'AMPc. Nous montrons que l'augmentation de l'activité synaptique provoque une augmentation de la concentration en AMPc intracellulaire qui pourrait réguler les oscillations de l'activité de réseau. Ces résultats illustrent le rôle clef d'un courant intrinsèque dans l'organisation temporelle de l'activité des neurones en réseau. La modulation des propriétés cinétiques du courant H peut agir comme un régulateur de fréquence.