Data moving across networks has increased thanks to technological developments, which has also created network security issues. The sophistication of these threats makes conventional intrusion detection systems unable to adequately manage these fresh security issues. Consequently, since neural-based intrusion detection systems can manage dynamic and complicated data, focus has turned to leveraging them to address these challenging cyber security issues. This work investigates how neural-based intrusion detection systems address challenging cyber security issues. Thus, for this work, important neural-based models of relevance are CNN and MLP models that have been shown to be useful in capturing abnormalities in big datasets and so pertinent for intrusion detection systems. The NSL-KDD dataset was applied for the experiment and served to support the idea on the application of neural-based models in intrusion detection. Deep learning models show, based on their accuracy, F1-score, and recall values, far higher performance in spotting anomalies than conventional models. These findings allow one to deduce that neural-based intrusion detection systems present a strong method for network the detection and prevention of cyber security events.