Sutherland & Rudy (1989) proposed configural association theory that the hippocampus plays a role in configuration of compound stimuli. This theory assumed that the hippocampus is necessarily for solving non-linear task but is not necessarily for solving linear task. This study examined the relationship hippocampal theta and non-linear and linear task. The chapter 1 presents a review of previous research. The chapter 2 showed that the hippocampal theta power during negative patterning task (A+, B+, AB-) as non-linear task was higher than simple discrimination task (A+, B-) as linear task. The chapter 3 showed that the hippocampal theta power during positive patterning task (A-, B-, AB+) as non-linear task was not higher than simple discrimination task (A+, B-) as linear task. The chapter 4 reanalyzed the data of chapter 2 and 3 by time-frequency analysis. The chapter 5 compared the hippocampal theta power during negative patterning (A+, B+, AB-), simultaneous feature negative (A+, AB -) and simple discrimination task (A+, B-). The result revealed that the hippocampal theta was the highest in negative patterning task, followed by simultaneous feature negative and simple discrimination task. In the chapter 6, I discussed these result of this study and proposed the new hippocampal function theory that the hippocampus is important for the modulation of expectation value of reinforcement for compound stimulus.