WIRELESS sensor networks (WSNs) are vulnerable to several types of attacks including passive eavesdropping, jamming, compromising of the sensor nodes, and insertion of malicious nodes into the network . Widespread adoption of WSNs, particularity for mission-critical tasks, hinges on the development of strong protection mechanisms against such attacks . Optimal Probabilistic Encryption for Secure Detection in Wireless Sensor Networks Due to the scarcity of resources, traditional wireless network security solutions are not viable for WSNs. The life span of a sensor node is usually determined by its energy supply which is mostly expended for data processing and communication . Optimal Probabilistic Encryption for Secure Detection in Wireless Sensor Networks Moreover, size and cost constraints of the nodes limit their memory size and processing power. Therefore, security solutions which demand excessive processing, storage or communication overhead are not practical. In particular, due to their high computational complexity, public key ciphers are not suitable for WSNs Optimal Probabilistic Encryption for Secure Detection in Wireless Sensor Networks. An important application of WSNs, which has been extensively studied in recent years, involves decentralized detection whereby the sensors send their (quantized) measurements to an ally fusion center (AFC) which attempts to detect the state of nature using the data received from all the sensors. Due to the broadcast nature of the wireless media, the sensors’ Optimal Probabilistic Encryption for Secure Detection in Wireless Sensor Networks data are prone to interception by unauthorized parties . In this paper we are concerned with data confidentiality in the presence of passive eavesdropping. In particular, weassume that the transmissions of the nodes are over insecure channels. An eavesdropping fusion center (EFC) is attempting to intercept the sensor’s messages and to detect the state of nature. Since the sensors’ data are used for hypothesis testing, security can be provided by degrading the detection performance of EFC. Optimal Probabilistic Encryption for Secure Detection in Wireless Sensor Networks The communication between the sensors and AFC (or EFC) is assumed to be over a parallel access channel where the sensors are connected to AFC (or EFC) by a dedicated channel. The dedicated channels are assumed to be independent and identical and are modeled by (noisy) discrete memoryless channels .