This study takes an analytic approach to delve into the dynamics of pyramidal neurons, focusing on interplay of the basal and perisomatic regions. A significant portion of recurrent excitatory connections converge onto the basal dendrites, generating N-Methyl-D-Aspartate (NMDA) driven depolarizations, also known as NMDA-spikes. The time-scale of NMDA-spikes significantly surpasses the neuronal membrane time-scale by a factor of five, underscoring their relevance in information integration. The voltage dynamics within the dendric subunits remain unaffected by somatic spikes, enabling their independent modeling. Despite their importance, the effects of NMDA-spikes on network dynamics remain elusive, necessitating comprehensive exploration. To decipher the influence of NMDA-spikes on network dynamics, we employ two distinct meanfield approaches. In the first approach, we employ binary neurons, leveraging a stereotypical connectivity pattern between dendritic and somatic compartments. This approach sheds light on the underlying mechanisms of network dynamics, with a specific focus on the phase transition from negligible to finite NMDA activity states. The second approach trades the analytical tractability for biological plausibility, employing the master equation formalism, coupled with a semi-analytical method tailored to multiple compartmental neurons. Through these two complementary approaches, we are able to predict the characteristic relationship between recurrence ratio and population output and aim to unravel the intricacies of pyramidal neuronal behavior, shining a light on the impact of NMDA-spikes on network-level dynamics.