Remembering the past during new learning: the temporal dynamics of integrative encoding
Liu, Zhenghao LU ; Bramao, Ines LU
and Johansson, Mikael
LU
(2022)
International Conference of Cognitive Neuroscience
- Abstract
- Memories may integrate elements experienced in different events. For instance, meeting a woman leaving her house, and later meeting another woman entering the same house, may allow us to infer that the two women live together. Such memory representations are thought to rely on integrative encoding mechanisms, allowing us to make inferences about the world and generalize knowledge to entirely new situations. This study uses multivariate pattern analysis (MVPA) of electrophysiological data to elucidate the temporal dynamics of integrative encoding. We adapted the associative inference task to investigate cross- episode memory integration. Participants first learned overlapping paired associates, AB and CB pairs, that comprised a word and a... (More)
- Memories may integrate elements experienced in different events. For instance, meeting a woman leaving her house, and later meeting another woman entering the same house, may allow us to infer that the two women live together. Such memory representations are thought to rely on integrative encoding mechanisms, allowing us to make inferences about the world and generalize knowledge to entirely new situations. This study uses multivariate pattern analysis (MVPA) of electrophysiological data to elucidate the temporal dynamics of integrative encoding. We adapted the associative inference task to investigate cross- episode memory integration. Participants first learned overlapping paired associates, AB and CB pairs, that comprised a word and a picture (a face or a bird). Memory was later tested for the AC inferred associations, and for the direct AB and CB associations. MVPA was used to measure the online reactivation of previously learnt AB memories during encoding of the CB events. Reliable reactivation of AB was observed between 1.5 and 3 s. Participants (n=29) were divided into subgroups according to their AC performance: good and poor generalizers. Interestingly, while both groups showed comparable levels of AB reactivation during CB learning, the reactivation effects were associated with different consequences. For good generalizers, reactivation was predictive of later AC inference performance, suggesting that an integrated ABC representation was formed during BC learning. This interpretation is corroborated by behavioral data showing that direct and indirect inferred associations were retrieved equally fast. Altogether, this suggests that AC inference in this group of participants was based on integrative encoding mechanisms. Conversely, for poor generalizers, the reactivation of AB was instead negatively correlated with BC retrieval performance, indicating that AB reactivation impaired BC learning. Additionally, this group of participants took longer to retrieve inferred associations than direct associations, suggesting that successful AC inference was accomplished by flexibly recombining the direct AB and CB associations during retrieval. The present study extends previous literature by revealing the temporal dynamics of memory integration and providing an account of inter-individual differences in the capacity to make inferences across distinct episodes. (Less)
Please use this url to cite or link to this publication:
https://lup.lub.lu.se/record/074b9c77-f99d-4855-94b2-4b965bbd862d
- author
- Liu, Zhenghao
LU
; Bramao, Ines
LU
and Johansson, Mikael
LU
- organization
- publishing date
- 2022
- type
- Contribution to conference
- publication status
- published
- subject
- keywords
- Memory integration, EEG, MVPA, integrative encoding, flexible retrieval
- conference name
- International Conference of Cognitive Neuroscience
- conference location
- Helsinki, Finland
- conference dates
- 2022-05-18 - 2022-05-22
- project
- When remembering affects new learning: Temporal dynamics of memory integration revealed by EEG and machine learning techniques
- Learning and remembering: The cognitive neuroscience of memory for real-world events
- language
- Swedish
- LU publication?
- yes
- id
- 074b9c77-f99d-4855-94b2-4b965bbd862d
- alternative location
- https://www2.helsinki.fi/sites/default/files/atoms/files/icon2022_poster_abstracts.pdf
- date added to LUP
- 2022-05-29 18:02:11
- date last changed
- 2022-06-27 18:27:25
@misc{074b9c77-f99d-4855-94b2-4b965bbd862d,
abstract = {{Memories may integrate elements experienced in different events. For instance, meeting a woman leaving her house, and later meeting another woman entering the same house, may allow us to infer that the two women live together. Such memory representations are thought to rely on integrative encoding mechanisms, allowing us to make inferences about the world and generalize knowledge to entirely new situations. This study uses multivariate pattern analysis (MVPA) of electrophysiological data to elucidate the temporal dynamics of integrative encoding. We adapted the associative inference task to investigate cross- episode memory integration. Participants first learned overlapping paired associates, AB and CB pairs, that comprised a word and a picture (a face or a bird). Memory was later tested for the AC inferred associations, and for the direct AB and CB associations. MVPA was used to measure the online reactivation of previously learnt AB memories during encoding of the CB events. Reliable reactivation of AB was observed between 1.5 and 3 s. Participants (n=29) were divided into subgroups according to their AC performance: good and poor generalizers. Interestingly, while both groups showed comparable levels of AB reactivation during CB learning, the reactivation effects were associated with different consequences. For good generalizers, reactivation was predictive of later AC inference performance, suggesting that an integrated ABC representation was formed during BC learning. This interpretation is corroborated by behavioral data showing that direct and indirect inferred associations were retrieved equally fast. Altogether, this suggests that AC inference in this group of participants was based on integrative encoding mechanisms. Conversely, for poor generalizers, the reactivation of AB was instead negatively correlated with BC retrieval performance, indicating that AB reactivation impaired BC learning. Additionally, this group of participants took longer to retrieve inferred associations than direct associations, suggesting that successful AC inference was accomplished by flexibly recombining the direct AB and CB associations during retrieval. The present study extends previous literature by revealing the temporal dynamics of memory integration and providing an account of inter-individual differences in the capacity to make inferences across distinct episodes.}},
author = {{Liu, Zhenghao and Bramao, Ines and Johansson, Mikael}},
keywords = {{Memory integration; EEG; MVPA; integrative encoding; flexible retrieval}},
language = {{swe}},
title = {{Remembering the past during new learning: the temporal dynamics of integrative encoding}},
url = {{https://www2.helsinki.fi/sites/default/files/atoms/files/icon2022_poster_abstracts.pdf}},
year = {{2022}},
}