In this paper a novel event triggered neural network (NN) based adaptive controller is presented for linear systems with multiple uncertainties. Controller design is primarily based on an observer, called the modified state observer (MSO). MSO is used to approximate uncertainties online, with two tunable gains, which allow fast approximation without inducing high frequency oscillations. On the other hand state information is transmitted to the feedback loop only when required in an aperiodic fashion. This aperiodic update is triggered by a dynamic condition based on errors. Consequently, this event triggered control (ETC) not only reduces the control computations, but also bring down the communication cost. Lyapunov analysis is used to show the stability of the system as well as to develop the event sample triggering condition. Efficacy of the proposed controllers is demonstrated using a numerical example.