Studies have found that flipped classroom teaching (FT) improves learning compared to lecture-based teaching (LT). However, whether the structured teacher-student interaction--the key feature of FT--plays an essential role in enhancing learning remains unclear, as do its neural underpinnings. Here, we compared three teaching conditions: FT with a video lecture and structured interaction, LT with a face-to-face lecture and spontaneous interaction, and control teaching (CT) with a video lecture and spontaneous interaction. The fNIRS-based hyperscanning technique was used to assess the interbrain synchrony (IBS) from teacher-student dyads. Results showed that the learning was significantly improved in FT than in LT and CT, and FT significantly increased teacher--student IBS in left DLPFC. Moreover, the IBS and learning improvements were positively correlated. Therefore, these findings indicate that the structured teacher-student interaction is crucial for enhancing learning in FT, and IBS serves as its neural foundation.