Representational formats of human memory traces

Abstract

Neural representations are internal brain states that constitute the brain’s model of the external world or some of its features. In the presence of sensory input, a representation may reflect various properties of this input. When perceptual information is no longer available, the brain can still activate representations of previously experienced episodes due to the formation of memory traces. In this review, we aim at characterizing the nature of neural memory representations and how they can be assessed with cognitive neuroscience methods, mainly focusing on neuroimaging. We discuss how multivariate analysis techniques such as representational similarity analysis (RSA) and deep neural networks (DNNs) can be leveraged to gain insights into the structure of neural representations and their different representational formats. We provide several examples of recent studies which demonstrate that we are able to not only measure memory representations using RSA but are also able to investigate their multiple formats using DNNs. We demonstrate that in addition to slow generalization during consolidation, memory representations are subject to semantization already during short-term memory, by revealing a shift from visual to semantic format. In addition to perceptual and conceptual formats, we describe the impact of affective evaluations as an additional dimension of episodic memories. Overall, these studies illustrate how the analysis of neural representations may help us gain a deeper understanding of the nature of human memory.

Emotional information facilitates or disrupts memory integration through distinct hippocampal processes of reactivation and connectivity

“…In support of the emotion-facilitated integration, our imaging results showed that emotional information increased hippocampal reactivation of related memories during both encoding and retrieval. Consistent with emerging work using MVPA to study human episodic memory (Heinen et al, 2023;Hennings et al, 2022;Staresina et al, 2012;Xue et al, 2010), our findings of emotionally-charged increases in reactivation reflect trial-specific reinstatement (i.e., pair-specific similarity) controlling for a general category-level representation (i.e., across-pair similarity). Given the benefits of hippocampal pattern completion on integrative encoding (Kuhl et al, 2010;Shohamy and Wagner, 2008;Wong et al, 2019), the increased hippocampal reactivation of first neutral memories during encoding of second emotional associations could facilitate online integration by rapidly linking neural representations of related information.…”

A Lasting Impression: Exploring the Meaningfulness of a Singular Moment

“…Categorically, a lasting impression may be made into an object for scientific inquiry to be studied as episodic memory. For example, we can research episodic memories as neural representations, recognizable patterns of brain states that reflect the previously experienced episodes (Heinen et al, 2023). This kind of focus, however, passes over experiential aspects of memory.…”

Context and semantic object properties interact to support recognition memory

“…Our results suggest that it is not only the degree to which a concept shares features that defines the structured representational space, but also the correlated structure across those features. Memory traces for items which have more shared conceptual features will become more similar to one another through semantisation of the trace, leading to poorer memory performance (Heinen et al, 2023), and increased false memories when items are more closely semantically related (Roediger et al, 2001). This may mean that the reason conceptually distinctive objects are better remembered is because this measure increases for items in less dense areas of semantic space -a space represented in the ventral visual and frontal cortices, and memory traces in less dense representational spaces benefit from reduced interference, in turn supporting the retrieval of that item.…”