The production effect indicates that the items could be better remembered in loud reading compared to silent reading. MacLeod et al (2010) put forward the distinctiveness account to explain this effect, suggesting that items read aloud are particularly distinctive compared to items read silently, pronunciation improves memory. The distinctiveness account has been supported by many studies. In loud reading, articulation movements and sound whilst such do not exist in silent reading. In other words, loud reading involves explicit articulation movement and sound compared with silent reading. However, the effect of articulation movement and sound on the production effect is still not clear. In the present study, we recruited 32 right-handed healthy university students (14 boys, 18 girls, Mage±SD=21.56±1.32years). They are all Chinese native speakers. The participants were asked to complete reading tasks in three reading modes: loud reading; silent mouthing (read silently while “mouthing” the words); silent reading (read silently without mouth movements). Functional near-infrared spectroscopy (fNIRS) was used to examine the brain cortical oxygenation changes during the tasks. Specifically, fNIRS data was acquired using the LABNIRS/16 (Shimadzu Corporation, Kyoto, Japan), with a 5×6 multichannel probe frame including 15 sources and 15 detectors, 49 channel montage, and the main areas of interest including the Primary Motor Cortex, Broca area, and Wernicke area. Before the official experiment, the participants practiced three reading modes. fNIRS caps were placed on the participants’ heads after the training was completed successfully. During the experiment, each participant completed three reading blocks, with each block containing three, one-minute reading passages (loud reading; silent mouthing; silent reading). Within the blocks, each reading task was intermixed with a 12s inter-stimulus interval (ISI) with a fixed cross-presenting on the screen. Sixty-second rest periods were placed before each block and after the final reading block. During the rest periods, the participants were instructed to look at a fixed cross in the middle of the monitor and to relax. The order of three different passages was presented randomly to each participant. There were no unfamiliar words in all passages, and they were homogeneous in difficulty and familiarity. After reading all passages, a recognition memory test was conducted immediately, and the whole experiment lasted around 20 minutes. The results indicate: (a)The memory performance of loud reading (p<0.001) and silent mouthing (p<0.001) was significantly better than that of silent reading, and there was no significant difference between that of loud reading and silent mouthing (p=0.21); (b)The Primary Motor Cortex (p=0.013), Broca's area (p=0.007) and Wernick's area (p<0.001) showed significantly higher activations in cortical oxygen levels during loud reading than silent reading; the Primary Motor Cortex (p=0.016), Broca's area (p=0.008) and Wernick's area (p<0.001) were significantly higher activated during silent mouthing than silent reading; (c)Wernick's area showed greater activation during loud reading than silent mouthing (p=0.005). Our results suggested that articulation movement plays a more important role in the production effect than sound.