Reciprocal structure-function relationships underly both healthy and pathological behaviors in complex neural networks. Neuropathology can have widespread implications on the structural properties of neural networks, and drive changes in the functional interactions among network components at the micro-, and mesoscale level. Thus, understanding network dysfunction requires a thorough investigation of the complex interactions between structural and functional network reconfigurations in response to perturbation. However, such network reconfigurations at the micro- and mesoscale level are often difficult to study in vivo. For example, subtle, evolving changes in synaptic connectivity, transmission, and electrophysiological shift from healthy to pathological states are difficult to study in the brain. Engineered in vitro neural networks are powerful models that enable selective targeting, manipulation, and monitoring of dynamic neural network behavior at the micro- and mesoscale in physiological and pathological conditions. In this study, we first established feedforward cortical neural networks using in-house developed two-nodal microfluidic chips with controllable connectivity interfaced with microelectrode arrays (mMEAs). We subsequently induced perturbations to these networks by adeno-associated virus (AAV) mediated expression of human mutated tau in the presysnaptic node and monitored network structure and activity over three weeks. We found that induced perturbation in the presynaptic node resulted in altered structural organization and extensive axonal retraction starting in the perturbed node. Perturbed networks also exhibited functional changes in intranodal activity, which manifested as an overall decline in both firing rate and bursting activity, with a progressive increase in synchrony over time. We also observed impaired spontaneous and evoked internodal signal propagation between pre-and postsynaptic nodes in the perturbed networks. These results provide novel insights into dynamic structural and functional reconfigurations in engineered feedforward neural networks as a result of evolving pathology.