The effect of prior knowledge on post-encoding brain connectivity and its relation to subsequent memory
Introduction
Making associations between different kinds of information is an important way of building our knowledge system (Halford et al., 2010). Studies have shown that these associative processes can be facilitated by previous experiences or prior knowledge, namely the knowledge one has acquired and brings to bear in acquiring new information (Kan et al., 2009, Poppenk et al., 2010a, Reder et al., 2013, Sharon et al., 2011). At the neural level, it has also been shown that prior knowledge can modulate brain activity during encoding and retrieval to enhance new memory processing (Liu et al., 2017, Maguire et al., 1999, Poppenk et al., 2010b, Staresina et al., 2009, van Kesteren et al., 2010a, van Kesteren et al., 2010b). In addition, there is growing evidence that changes in resting functional connectivity after encoding reflect early memory consolidation (de Voogd et al., 2016, Gruber et al., 2016, Hermans et al., 2017, Staresina et al., 2013, Tambini et al., 2010, Tambini and Davachi, 2013), but there is little research on whether prior knowledge can also influence this process. In this paper we show that prior knowledge can promote post-encoding functional connectivity in regions implicated in memory and perception, and that this effect is related to subsequent associative memory.
A large body of animal literature has shown that post-encoding brain activity, e.g., in the hippocampus (HPC), may reflect memory replay or reactivation and contribute to memory stabilization (Foster and Wilson, 2006, Jadhav et al., 2012, Knauer et al., 2013, Ólafsdóttir et al., 2016, O'Neill et al., 2010, O'Neill et al., 2008, Oudiette and Paller, 2013, Silva et al., 2015, Singer and Frank, 2009, Sutherland and McNaughton, 2000, Wilson and McNaughton, 1994). Other brain structures such as the striatum, ventral medial prefrontal cortex (vmPFC), visual, and motor cortex are also involved in the replay of previous experiences during post-learning rest or sleep (Gomperts et al., 2015, Hoffman and McNaughton, 2002, Ji and Wilson, 2007, Pennartz et al., 2004, Ribeiro et al., 2004). In line with the animal literature, human functional neuroimaging studies have found that learning experiences can modulate resting-state brain connectivity (Albert et al., 2009, Deuker et al., 2013, Groen et al., 2011, Hasson et al., 2009, Urner et al., 2013, Wang et al., 2012, Zou et al., 2013), and that the HPC activity during post-encoding rest can be correlated with memory performance (Tambini et al., 2010, Staresina et al., 2013, Tambini and Davachi, 2013, de Voogd et al., 2016, Gruber et al., 2016, Hermans et al., 2017). These observations support the idea that post-encoding neural activity may reflect early memory consolidation processes by which newly-encoded memory becomes more stable. Intriguingly, evidence from recent animal and human neuroimaging studies has also shown that prior knowledge can facilitate this memory consolidation process (Hennies et al., 2016, Tse et al., 2007). This raises the important question of whether prior knowledge can directly affect brain activity during an immediate post-encoding time window to contribute to memory formation in humans.
Although a recent study found that higher sleep spindle density during post-encoding sleep predicted better schema-related memory (Hennies et al., 2016), to the best of our knowledge, only two human neuroimaging study (van Kesteren et al., 2010a; Schlichting and Preston, 2016) have directly examined prior knowledge effects on brain activity during post-encoding awake rest. However, both studies mainly focused on the connectivity between the HPC and vmPFC, based on the assumption that the vmPFC plays an important role in supporting existing structured mental representations or memory updating (van Kesteren et al., 2012, Zeithamova et al., 2012, Zeithamova and Preston, 2010). While this assumption is reasonable, other regions, such as the anterior temporal pole (aTPL), also have been shown to support prior knowledge (Kan et al., 2009, Liu et al., 2017, Sharon et al., 2011, Staresina et al., 2009) and should be considered in the study of prior knowledge effects on new learning.
To tackle these questions, we need to design experiments that can elicit different types or components of prior knowledge, such as semantics, social emotions, perceptions, or episodic memories, which can be supported by well-known distinct brain regions and the connections between them. Then, we can use fMRI to examine how functional connectivity among these different brain regions can be affected by prior knowledge manipulations in the preceding encoding tasks. Following this idea, in this study we designed an associative memory task that required forming associations between two items, a house and a face, for which there are well-delineated and differential localizations in the brain. The houses were always novel, but to manipulate prior experience, the faces were those of famous people or of people unfamiliar to the participants. Specifically, participants were asked to associate pictures of novel houses with pictures of either famous or nonfamous faces in different blocks. After each encoding block, participants rested in the scanner. Because famous faces can elicit rich stores of social/affective, semantic, and perceptual knowledge, as well as episodic memories, each supported by different brain systems (Fairhall and Ishai, 2007, Gobbini et al., 2004, Gobbini and Haxby, 2007, Ishai, 2008, Renoult et al., 2012, Ross and Olson, 2012, Simmons et al., 2010), this fame manipulation created two conditions, famous vs. nonfamous, in which prior knowledge effects at the brain level can be examined during post-encoding rest.
Similar to the previous study that focused on the encoding phase (Liu et al., 2017), we included the HPC, aTPL, vmPFC, parahippocampal place area (PPA), and fusiform face area (FFA) as our regions of interest (ROIs). We chose these ROIs because the literature has shown that the medial temporal lobe, especially the HPC, plays an important role in associative encoding (Davachi, 2006, Davachi and Wagner, 2002, Eichenbaum et al., 2007) and post-encoding memory reactivation (O'Neill et al., 2010). Previous studies have also shown that the vmPFC and aTPL support social evaluative and semantic processing (Etkin et al., 2011, Grabenhorst and Rolls, 2011, Luo et al., 2010, O'Reilly, 2010, Patterson et al., 2007, Roy et al., 2012) and contribute to prior knowledge facilitation effects (Kan et al., 2009, Liu et al., 2017, Ross et al., 2011, Sharon et al., 2011, Staresina et al., 2009, van Kesteren et al., 2012). The two posterior perceptual regions, i.e., the PPA and FFA, support the processing of house and face stimuli, respectively (Kanwisher, 2010), and their activation can be affected by prior knowledge related to familiar faces (Bar et al., 2008, Gobbini and Haxby, 2006, Liu et al., 2014). Importantly, all of these ROIs, which support different components of memory processing, showed increased activity during face-house associative encoding when prior knowledge was involved (Liu et al., 2017), raising the possibility that they also play a role in post-encoding memory processes. Therefore, by comparing the functional connectivity among these ROIs between the famous post-encoding and nonfamous post-encoding rest periods, the current study enables us to investigate how prior knowledge affects post-encoding brain activity and impacts subsequent memory.
The literature on memory consolidation has shown that the medial temporal lobe, especially the HPC, is crucial for supporting new memory processing, whereas neocortical regions may play a greater role in supporting already consolidated memories (Frankland and Bontempi, 2005, Moscovitch et al., 2005, Nadel and Moscovitch, 1997, Squire and Alvarez, 1995). This switch in neural substrates may also underlie changes in the nature of these memories following consolidation (Nadel and Moscovitch, 1997, Wiltgen and Silva, 2007, Winocur and Moscovitch, 2011). Because it has been found that prior knowledge can facilitate memory consolidation (Tse et al., 2011, Tse et al., 2007), we hypothesized that the anterior cortical regions, namely, the vmPFC and aTPL, should form stronger resting connectivity with the HPC, PPA, and FFA after the famous compared to the nonfamous encoding condition, reflecting stronger cortical involvement from anterior brain regions during early memory consolidation when prior social/evaluative or semantic knowledge was involved. These connectivity measures should also better predict associative memory performance in the famous than the nonfamous condition. Moreover, our earlier finding that the HPC and FFA/PPA activations were stronger in the famous than nonfamous encoding condition (Liu et al., 2017), would lead to the prediction that the connectivity between the HPC and PPA/FFA should also be stronger during the famous post-encoding than the nonfamous post-encoding rest, reflecting stronger episodic binding processes when prior knowledge is present.
Section snippets
Participants
Twenty young adults (12 females), between 18 and 24 years of age (Mean = 21.3, SD = 1.49), were recruited from the University of Toronto's St. George campus. They were all right-handed, native English speakers, and free of any psychiatric or neurological conditions. The participants were paid $76 and gave their informed consent. The study was approved by the Research Ethics Board at Baycrest Center for Geriatric Care (University of Toronto).
Overview
There were 3 resting scans in this experiment (Fig. 1
Behavioral results
First, we found that associative memory performance (hit rate - false alarm rate) was better for the famous (M = 0.36, SD = 0.21) than the nonfamous (M = 0.19, SD = 0.14) condition (t(19) = 4.60, p < 0.0002, Cohen's d = 1.03). The correlation between associative memory performance in the two conditions was also significant (r = 0.47, p = 0.035). Interestingly, participants' age predicted associative memory performance in the famous (r = 0.60, p = 0.005), but not the nonfamous condition (r
Discussion
Using an explicit encoding task in which participants associated novel houses with either famous or nonfamous faces, we investigated how associative encoding with or without prior knowledge involvement differentially affected post-encoding brain connectivity. We hypothesized stronger post-encoding connectivity in the famous than non-famous condition among regions whose activation at encoding was found to play a role in associative memory performance. For the most part, our results were
Conclusion
The current findings, consistent with our hypotheses, suggest that when prior knowledge is involved, the HPC, vmPFC, and aTPL, which support prior episodic, social-evaluative/schematic, and semantic memories, respectively, continue to interact with each other and the posterior perceptual brain regions (e.g., the PPA and FFA) during the post-encoding rest to facilitate off-line processing of the newly formed memory and lead to better memory for it. Our findings may also provide preliminary
Author contributions
Z.-X.L. and M.M. designed research, Z.-X.L. performed research, Z.-X.L. analyzed data with help from M.M., and C.G., Z.-X. L., M.M., and C.G. wrote the paper.
Conflicts of interest
No conflict of interest.
Acknowledgement
This study was supported by Natural Sciences and Engineering Research Council of Canada RGPIN 8347 to M. M. Z.-X. L. was also supported by NSERC PDS scholarship. We would like to thank Dr. Kang Lee at the Institute of Child Study, University of Toronto, for his helpful comments and suggestions. We also gratefully acknowledge Marilyne Ziegler for her help in programming the experimental task, and Yasha Khatamian and Annette Weekes-Holder for their help in designing the fMRI protocol.
References (109)
- et al.
The resting human brain and motor learning
Curr. Biol.
(2009) - et al.
A component based noise correction method (CompCor) for BOLD and perfusion based fMRI
NeuroImage
(2007) - et al.
Dissociating measures of associative memory: evidence and theoretical implications
J. Mem. Lang.
(2007) Item, context and relational episodic encoding in humans
Curr. Opin. Neurobiol.
(2006)- et al.
Awake reactivation of emotional memory traces through hippocampal–neocortical interactions
NeuroImage
(2016) - et al.
Cellular and systems reconsolidation in the hippocampus
Neuron
(2002) - et al.
Emotional processing in anterior cingulate and medial prefrontal cortex
Trends Cogn. Sci.
(2011) Different patterns of famous people recognition disorders in patients with right and left anterior temporal lesions: a systematic review
Neuropsychologia
(2007)- et al.
Neurobiology of schemas and schema-mediated memory
Trends Cogn. Sci.
(2017) - et al.
Neural systems for recognition of familiar faces
Neuropsychologia
(2007)
Neural response to the visual familiarity of faces
Brain Res. Bull.
Value, pleasure and choice in the ventral prefrontal cortex
Trends Cogn. Sci.
Post-learning hippocampal dynamics promote preferential retention of rewarding events
Neuron
Relational knowledge: the foundation of higher cognition
Trends Cogn. Sci.
Reverberation of recent visual experience in spontaneous cortical waves
Neuron
Let's face it: it's a cortical network
NeuroImage
Dynamic neural systems enable adaptive, flexible memories
Neurosci. Biobehav. Rev.
An automated method for neuroanatomic and cytoarchitectonic atlas-based interrogation of fMRI data sets
NeuroImage
Memory consolidation, retrograde amnesia and the hippocampal complex
Curr. Opin. Neurobiol.
The role of the ventromedial prefrontal cortex in memory consolidation
Behav. Brain Res.
Play it again: reactivation of waking experience and memory
Trends Neurosci.
The what and How of prefrontal cortical organization
Trends Neurosci.
Upgrading the sleeping brain with targeted memory reactivation
Trends Cogn. Sci.
Interplay of hippocampus and prefrontal cortex in memory
Curr. Biol.
Personal semantics: at the crossroads of semantic and episodic memory
Trends Cogn. Sci.
Ventromedial prefrontal-subcortical systems and the generation of affective meaning
Trends Cogn. Sci.
Valuation and decision-making in frontal cortex: one or many serial or parallel systems?
Curr. Opin. Neurobiol.
Hippocampal–medial prefrontal circuit supports memory updating during learning and post-encoding rest
Neurobiology of Learning and Memory, Hippocampal Interactions with Brain Networks that Influence Learning & Memory
Rewarded outcomes enhance reactivation of experience in the hippocampus
Neuron
Retrograde amnesia and memory consolidation: a neurobiological perspective
Curr. Opin. Neurobiol.
Memory trace reactivation in hippocampal and neocortical neuronal ensembles
Curr. Opin. Neurobiol.
Resting state functional connectivity changes induced by prior brain state are not network specific
NeuroImage
Enhanced brain correlations during rest are related to memory for recent experiences
Neuron
Automated anatomical labeling of activations in spm using a macroscopic anatomical parcellation of the mni mri single-subject brain
NeuroImage
Early visual learning induces long-lasting connectivity changes during rest in the human brain
NeuroImage
Functional connectivity during light sleep is correlated with memory performance for face-location associations
NeuroImage
How schema and novelty augment memory formation
Trends Neurosci.
Changes in the brain intrinsic organization in both on-task state and post-task resting state
NeuroImage
Famous faces activate contextual associations in the parahippocampal cortex
Cereb. Cortex
Controlling the false discovery rate: a practical and powerful approach to multiple testing
J. R. Stat. Soc. Ser. B Methodol.
Ventromedial prefrontal cortex supports affective future simulation by integrating distributed knowledge
Proc. Natl. Acad. Sci.
Constructing realistic engrams: poststimulus activity of hippocampus and dorsal striatum predicts subsequent episodic memory
J. Neurosci.
Hippocampal immediate poststimulus activity in the encoding of consecutive naturalistic episodes
J. Exp. Psychol. Gen.
Overlap among spatial memories triggers repulsion of hippocampal representations
Curr. Biol.
Hippocampal contributions to episodic encoding: insights from relational and item-based learning
J. Neurophysiol.
Memory consolidation by replay of stimulus-specific neural activity
J. Neurosci.
Image sequence reactivation in awake V4 networks
Proc. Natl. Acad. Sci.
The medial temporal lobe and recognition memory
Annu. Rev. Neurosci.
Fast-forward playback of recent memory sequences in prefrontal cortex during sleep
Science
Effective connectivity within the distributed cortical network for face perception
Cereb. Cortex
Cited by (28)
Amplitude of low-frequency fluctuation after taste exposure revealed by resting-state fMRI
2023, Physiology and BehaviorThe influence of prior knowledge on the formation of detailed and durable memories
2021, Journal of Memory and LanguageA predictive account of how novelty influences declarative memory
2021, Neurobiology of Learning and MemoryCausal Contribution of Awake Post-encoding Processes to Episodic Memory Consolidation
2020, Current BiologyCitation Excerpt :Here, we sought to target and test the role of post-encoding processes in human memory consolidation using a combined transcranial magnetic stimulation (TMS) and fMRI approach. We took advantage of prior work showing that representational regions in occipitotemporal cortex spontaneously reactivate representations of recently encoded stimuli and display post-encoding interactions with the hippocampus that are related to later memory [13, 15, 20, 43–45]. To interfere with these processes, we applied inhibitory continuous theta-burst TMS (cTBS) [46] to lateral occipital cortex (LOC) after participants incidentally encoded object-face pairs and before an extended post-encoding awake rest period (Figure 1).
Awake Reactivation of Prior Experiences Consolidates Memories and Biases Cognition
2019, Trends in Cognitive SciencesCitation Excerpt :Moreover, hippocampal–cortical resting connectivity immediately after learning is predictive of the reorganization of cortical memory representations across shared experiences one week later [81]. On the flip side, other work has shown that the presence of prior knowledge during new learning is also associated with increased post-encoding hippocampal–cortical connectivity [73,82], consistent with a role of prior experience in modulating post-encoding systems-level interactions. Together, this work provides compelling evidence that post-encoding reactivation not only serves to strengthen recent experiences in memory, but also reflects the selection of relevant experiences and the promotion of memory integration both behaviorally and via the emergence of integrated, or schematic, cortical representations [81].