Shared neural codes of recognition memory

Abstract

Research on recognition memory has identified several electrophysiological indicators of successful memory retrieval, known as old-new effects. These effects have been observed in different sensory domains using various stimulus types, but little attention has been given to their similarity or distinctiveness and the underlying processes they share. Dual process theories of memory retrieval have linked the early mid-frontal effect to familiarity and the late parietal component to recollection. Here, a data-driven approach was taken to investigate the temporal evolution of shared information content between different memory conditions using openly available EEG data from healthy human participants of both sexes, taken from six experiments. A test dataset involving personally highly familiar and unfamiliar faces was used. The results show that neural familiarity signals for face stimuli were highly generalized starting from around 200 ms following stimulus onset. When training was performed on non-face datasets, an early (around 200-300 ms) to late (post-400 ms) differentiation was observed over most regions of interest. Successful cross-classification for non-face stimuli (music and object/scene associations) was most pronounced in late period. Additionally, a striking dissociation was observed between familiar and remembered objects, with shared signals present only in the late window for correctly remembered objects, while cross-classification for familiar objects was successful in the early period as well. These findings suggest that late neural signals of memory retrieval generalize across sensory modalities and stimulus types, and the dissociation between familiar and remembered objects may provide insight into the underlying processes.

Your place or mine? The neural dynamics of personally familiar scene recognition suggests category independent familiarity encoding

“…Thus, these alternatives need to be explored in specific future fMRI studies. Furthermore, complete generalizability would require the comparison of several other non-face object categories and familiar stimuli from other modalities, such as auditory stimuli as well (Ambrus, 2022). Future studies should systemically investigate familiarity processing in a variation of non-face stimulus categories in the visual domain, to corroborate evidence from our study as well.…”

“…It is thus very likely that this effect extends to other experimental paradigms, different qualities of familiarity, and even different stimulus categories. Indeed, a recent cross-category classification study by Ambrus (2022) showed that the same processing dynamics observed for familiar faces also apply to other categories and experimental paradigms, such as familiar and remembered objects, remembered (vs. forgotten) object-scene associations, and even to music, reflecting a generalization across sensory modalities. Familiarity effects were observed for all categories examined in the analysis at a relatively later time range (post 300-400 ms), while a shared and sustained effect was observed earlier (at 200 ms) only for personally familiar faces and experimentally familiarized objects later judged as subjectively familiar, but not for those judged as remembered.…”

“…Instead, all evidence indicates that familiarity, as previously established, reflects a general and stimulus-independent process triggered by previously encountered exemplars of an object category, which can be seen as early as 200 ms after stimulus onset. It is likely that this process is related not just to a familiarity judgement per se, but, especially at later stages, to specific recall processes involving episodic memory and recollection (Ambrus 2022).…”

“…While recognition memory, including familiarity and recollection (Rugg and Yonelinas 2003;Dimsdale-Zucker et al 2022) has been explored for various types of stimuli, there has been limited research on whether the underlying neural processes of familiarity differ or are similar across stimuli categories, and if they generalize across different types of stimuli (Kwon et al 2022). Recently, Ambrus (2022) conducted a cross-experiment classification analysis based on data from prior studies and found a significant overlap in neural signals of recognition memory processes between 400-600 ms, regardless of sensory modality, stimulus type, or memory age. Although these findings provided evidence for shared recognition effects across different types of stimuli, a systematic, direct comparison of the temporal dynamics of memory processes for face and non-face stimuli is still lacking.…”