The role of the episodic buffer in working memory for language processing

Abstract

A body of work has accumulated to show that the cognitive process of binding information from different mnemonic and sensory sources as well as in different linguistic modalities can be fractionated from general executive functions in working memory both functionally and neurally. This process has been defined in terms of the episodic buffer (Baddeley in Trends Cogn Sci 4(11):417-423, 2000). This paper considers behavioural, neuropsychological and neuroimaging data that elucidate the role of the episodic buffer in language processing. We argue that the episodic buffer seems to be truly multimodal in function and that while formation of unitary multidimensional representations in the episodic buffer seems to engage posterior neural networks, maintenance of such representations is supported by frontal networks. Although, the episodic buffer is not necessarily supported by executive processes and seems to be supported by different neural networks, it may operate in tandem with the central executive during effortful language processing. There is also evidence to suggest engagement of the phonological loop during buffer processing. The hippocampus seems to play a role in formation but not maintenance of representations in the episodic buffer of working memory.

Predictive processing increases intelligibility of acoustically distorted speech: Behavioral and neural correlates

“…According to this interpretation, the impact of memory in speech comprehension progressively increases when the quality of the acoustic signal decreases, until at the extreme forms of distortion, speech comprehension becomes a cued memory task. Related ideas have been presented in the Ease of Language Understanding (ELU) model that describes how and when working memory is involved in disambiguating acoustically distorted speech, and how it interacts with long‐term memory (LTM) during this process (Baddeley, 2000; Rönnberg, Rudner, Lunner, & Zekveld, 2010; Rönnberg et al., 2013; Rudner & Rönnberg, 2008). This model suggests that in easy listening conditions, sensory information matches with the corresponding syllabic phonological representation in semantic LTM, resulting in rapid memory retrieval.…”

Working memory for meaningless manual gestures.

“…In particular, working memory for sign language specifically recruits bilateral superior parietal regions associated with visuospatial processing (Bavelier, Newman, Working memory for manual gestures Mukherjerr, Hauser, Kemeny, Braun & Boutla, 2008;Rönnberg, Rudner & Ingvar, 2004;Rudner, Fransson, Nyberg, Ingvar & Rönnberg, 2007). This suggests that there are fundamental differences in the working memory processing of oral and gesture-based languages, despite considerable functional similarities Rudner et al, 2009;Rudner & Rönnberg, 2008). It is likely that these differences are related to sensorimotor mechanisms supporting the two language modalities and the ways in which they are represented (Emmorey, McCullough, Mehta & Grabowski, 2014).…”

Preexisting semantic representation improves working memory performance in the visuospatial domain

“…The Ease of Language Understanding (ELU) model of WM proposes that WM in the service of communication is multimodal. Input to the system can be in any language modality, transmitted by any or several sensory modalities, and enters an episodic buffer (Rudner & Rönnberg, 2008b) whose function is Rapid Automatic Multimodal Binding of PHOnology (RAMBPHO). When the input can be smoothly matched to existing representations in long-term memory, language understanding is implicit and experienced as effortless.…”