Train control systems provide for the safe operation of railways by synchronizing the movement of trains and maintaining their separation along shared railway tracks. However, the use of wireless data transmissions in Communications-Based Train Control (CBTC) systems has rendered railways vulnerable to cyber attacks. The current body of research concerning security, specifically intrusion detection, for CBTC does not consider that most railway operators deploy this system alongside a secondary train control system, known as an external interlocking. The integration of these two train control systems allows for the operation of both CBTC-capable and CBTC- incapable trains along the same railway. The IEEE 1474.1 standard for CBTC terms this mixed-mode operation. Mixed-mode operation allows both CBTC-capable and CBTC-incapable trains to operate alongside each other. This thesis proposes a security solution for mixed-mode operations in an integrated CBTC-external interlocking railway based on the functional requirements and performance standards of IEEE 1474.1. This addresses the aforementioned gap in the literature by securing the wire- less communications of such railways with machine learning-based intrusion detection. The intrusion detection methods proposed in this work were evaluated in a simulated railway environment that integrated the CBTC system with an external interlocking to provide the functionality of mixed-mode operations. Multiple machine learning models have been trained on the resultant data transmissions of both systems under normal and attack conditions. The subsequent results of their performance provide valuable insight into which models can best meet the real-time requirements of an integrated CBTC-external interlocking railway to ensure the integrity and availability of the system.