A new encoding technique based on construction of communication signals with respect to optimal perceptual pattern sensitivity of the cognitive receivers is proposed here. Communication signals are directly generated and adapted according to medium properties without a modulator stage. Receivers recognize signal patterns from a chosen frequency band. These patterns are matched to respecting data sequence within a glossary to recover the information. Our purpose is increasing the bandwidth efficiency via manageable Signal to Noise Ratio. Error free recoverability of the information encoded and carried by multiple pattern combinations in a noisy communication channel is the key factor for this purpose. A pattern based encoding and neural network based decoding technique is used in this study. It is shown that if the predefined communication patterns are appropriately chosen then the transmitted information by multiple sources can be decoded by the receivers in the same time interval, frequency band and location.