Slide Slam E11
How Neural Activities during Encoding Novel Words and Resting Predict Memory Retrieval after Different Time Scales?
Yuejuan Wang1, Shuang Chen1; 1Department of Psychology, Zhejiang Normal University, Jinhua, China
To tell the fate of novel words by their learning process is crucial to reveal the mechanism for effective word learning. The widely accepted Subsequent Memory Effect indicated that the neural activity during encoding could predict the memory performance. Therefore, the present study aimed to explore the neural features of encoding for successful novel word learning. In addition, the neural pattern similarity was introduced to capture the neural correlates of the encoded information recently, whereas a debate about whether greater or smaller similarity is associated with better memory performance emerged, which need further exploration. Moreover, neural oscillations in the resting phase could predict subsequent memory. However, these subsequent memory effects on neural activity were mostly discovered within a short time lag, opening another question for the present study, that how neural activities during encoding novel words and follow-up resting predict successful retrieval after a longer time interval. Participants (N = 40, 33 females, Mage = 19.6 years) memorized 90 novel word-picture associations twice. After a 5-min quiet rest, the novel words were tested in a semantic category (living or not living) and a recognition (choosing picture from three alternatives) task. After one month, they were tested again. We coded the novel words categorized and recognized correctly as the remembered items and the rest as forgotten items for 5-min and 1-month intervals respectively. First, we captured the subsequent memory effect on LPC (800~1400ms) for both intervals as remembered novel words elicited more positive-going LPC than forgotten items (covering frontal and central regions for 5-min interval, while only frontal region for 1-month interval). These results revealed that typical SME on LPC could predict long-term memory retrieval at both short and long time lags. Second, within-item spatiotemporal pattern similarity of remembered items in 5-min interval was larger than that of forgotten items in an early time window (-70~120ms) at left frontal and central regions, while an opposite effect was found later (240~480ms) at frontal region. The pattern for 1-month interval was similar, showing a pre-stimuli effect (-440~-270ms) at right posterior region and inversed later (650~1100ms) at left frontal region. Early greater similarity associated with better memory might contribute to the specific item encoding, whereas the late smaller similarity for better memory might involve association encoding which aligned with the Encoding-variability hypothesis. Third, there was a negative correlation between the subsequent memory performance and the power of alpha oscillation in resting phase, which might reflect increased information processing. However, the activity during rest failed to predict the memory after one month. Our results manifested the classic subsequent memory effect after a long time interval for novel word learning and limited the prediction power of oscillation during rest for longer time memory. More importantly, we figured out a complex correlation between the neural activity of the encoding and the memory performance after short and long time intervals.