<p>A novel aperiodic adaptive event-triggered mechanism (AAETM) is adopted to address the synchronization issue of chaotic neural networks with actuator saturation. This control strategy is designed based on more general aperiodic sampling intervals. In particular, to save limited communication resources, an adjustable function related to the softsign function is introduced into the AAETM. Then, the improved sampled-data synchronization criteria are derived by combining the AAETM with the two-side looped-functional method and sawtooth-characteristic-based free matrix integral inequality. Finally, a numerical example is given to verify the superiority of the proposed method.</p>