Which type of memory is described as knowledge about words concepts and language based knowledge and facts?

Category Specificity

Some patients with disordered semantic memory show disproportionate impairment for certain categories of knowledge (e.g., knowledge about animals or fruits) with relative sparing of other categories (e.g., household utensils) (Warrington and Shallice, 1984). Other patients show the opposite pattern: greater impairment for non-biological compared to biological categories (Warrington and McCarthy, 1983). There are reported disparities too between other classes of information: between knowledge of abstract and concrete words, verbs and nouns, words and numbers. The fact that dissociations are observed in both directions effectively rules out explanations in terms of inherent differences in the relative difficulty of items.

An influential interpretation (the functional-sensory hypothesis) was that category dissociations result from differences in the weightings of properties that define a concept and the sensory processing channels through which information was originally acquired (Warrington and Shallice, 1984). Household objects such as “cup” and “bowl”, it was argued, are distinguished largely by their different functions, whereas fruits such as “apple” and “banana”, are distinguished predominantly on the basis of sensory properties such as colour, shape, texture, taste and smell. A disruption to knowledge about sensory attributes would result in disproportionately severe impairment for animals compared to household objects, whereas disruption to knowledge about functional attributes would produce the reverse. The underlying assumption is that information is organized in the brain by property rather than category.

Within the broad framework of the functional-sensory account, a wide range of dissociations ought theoretically to be possible, reflecting relative degrees of disruption to particular sensory or non-sensory attributes. The conceptual distinction between an apple and orange arguably relies heavily on colour, taste and smell, but less on shape, and not at all on sound, or movement. By contrast, form, sound and movement are relevant dimensions in the conceptual distinction between a cat and mouse. Household objects are manipulable, so are associated with specific motor actions (e.g., grasping, pouring, cutting), whereas other non-living things, such as large pieces of furniture, are typically static so are less conceptually associated with a motor action. Dissociations in performance in semantically-disordered patients can indeed be subtle. Semantic dementia patients have, for example, been reported to show disproportionately impaired knowledge about colour-related compared to form-related words, and about action words referring to face movements and speech acts than actions performed via the hand and arm (Pulvermüller et al., 2010).

The sensory-functional approach is not the only interpretation of category differences that has been advanced. The Domain specificity hypothesis of Caramazza and Shelton (1998) proposed that different categories of knowledge are supported by distinct neural networks, so are regionally segregated in the brain. The finding in semantic dementia that number knowledge (including understanding of quantity and arithmetical skills) can be strikingly preserved, in the face of severe semantic loss, would support the notion of domain specificity. Category differences are however for the most part relative rather than absolute. Patients with disordered semantic memory typically show partial information loss about concepts across multiple domains, a finding that would be predicted by the functional-sensory account because of its emphasis on relative weightings of attributes rather than categories per se.

Irrespective of the relative merits, both functional-sensory and domain-specificity accounts carry the notion of distributed representation of semantic memory across different brain regions.

Semantic memory

Semantic memory is sustained by relatively preserved lateral temporal lobes in AUD. Semantic memory refers to the memory of meaning, understanding, general knowledge about the world, and other concept-based knowledge unrelated to specific experiences. The level of consciousness associated with semantic memory is noetic (giving rise to feelings of familiarity or knowing) because it is independent of encoding context (Tulving, 1985, 2001). Fama et al. (2011) studied remote semantic memory processes in three clinical groups: AUD group, patients infected with the human immunodeficiency virus (HIV), and patients comorbid for both conditions (AUD + HIV group), compared with healthy controls. AUD and HIV groups exhibited performance below healthy controls, but these differences were not statistically significant, whereas AUD + HIV group appeared impaired compared with healthy controls. Although remote semantic memory has been found preserved in AUD patients (Fama et al., 2011), recently detoxified patients may experience difficulties to acquire new semantic information. Pitel et al. (2007b) studied the ability to acquire new semantic concepts including, for each concept, a label, a superordinate category, and three features associated with a picture. The learning protocol comprised eight daily sessions. AUD patients were able to acquire the category and features of the semantic concepts, albeit slowly, but they presented impaired label learning. AUD patients invoked different and inefficient cognitive strategies to attempt to compensate for impaired episodic and working memory. The use of errorless learning may be relevant for AUD patients with cognitive deficits to learn new complex semantic knowledge, and more particularly, new labels (Pitel et al., 2010). Moreover, information acquired with errorless learning was flexible, i.e., it may be generalized and or transferred to other situations. This learning condition allows preventing that patients repeat their errors in the course of the acquisition, learning them instead of the correct answers, and leading to learning impairments.

Semantic variant Primary Progressive Aphasia (svPPA)

svPPA is characterized by progressive loss of semantic memory which moderates information of words, objects, and concepts. There is characteristically a history of “forgetting the names” of items and faces. Other language skills, such as producing speech and repeating phrases and sentences spoken by others, are not affected. Even though the patients continue to speak fluently, their speech becomes vague and difficult to understand. As the disease progresses, similar behaviors seen in bvFTD, such as disinhibition and rigid food preferences, may be added to classic picture (Catani et al., 2013).

Introduction

Compared with other memory subtypes (e.g., semantic memory or procedural memory), the development of episodic memory during childhood is particularly intricate, unfolding progressively over three distinct periods: infantile amnesia (from birth to age 2), followed by childhood amnesia (from age 2 to age 6), and the emergence of adult-like episodic memory (from age 6 years onward) (Fig. 44.1). The phenomena of infantile and childhood amnesia have been known since the end of the 19th century, stemming from the observation that adults cannot recall detailed memories from their infancy and early childhood (e.g., Miles, 1895). They were thus initially studied from the observation of adults’ first memories. However, during the past three decades, research in the growing field of memory development brought two key findings to light. First, infantile and childhood amnesia are not explained by an absence of long-term memory abilities during the first years of life, as episodic memory abilities are present early in ontogeny, albeit in an immature form. Second, the boundaries of infantile amnesia (around 2 years old) and childhood amnesia (around 6 years old), as defined from adulthood, also correspond to important qualitative transitions in infants’ and children’s memory competence. Therefore, episodic memories are formed early in ontogeny, but they are rapidly forgotten over time, and important shifts in memory competence occur around 2 and 6 years of age (see Bauer, 2015a,b; Bauer, Larkina, & Deocampo, 2011, for discussion).

Importantly, early life experiences, despite not being remembered at adulthood, are not without consequence to one’s cognitive and psychological development. It is well documented that early adverse experiences (such as neglect or abuse) can have long-lasting detrimental effects on cognition and behavior. Adverse experiences have been linked to an increase in the relative risk of experiencing anxiety, depression, or personality disorders at adulthood (Alberini & Travaglia, 2017; Heim & Nemeroff, 2001; Perry & Sullivan, 2014). They have also been related to structural and functional anomalies in the hippocampus, a structure pivotal for long-term memory (Frodl, Reinhold, Koutsouleris, Reiser, & Meisenzahl, 2010; Lambert et al., 2017). The study of the early development of episodic memory is hence crucial for identifying the early risk factors of various neuropsychiatric disorders and for treating their later manifestations.

During the past decades, several theories have been proposed to explain infantile and childhood amnesia, linking the rapid forgetting of early memories and the protracted maturation of the hippocampus (e.g., Bauer, 2015a; Josselyn & Frankland, 2012; Li, Callaghan, & Richardson, 2014; Olson & Newcombe, 2014). However, they diverge by focusing on distinct aspects of hippocampal maturation. Recently, a conceptual framework based upon the concept of critical periods (Alberini & Travaglia, 2017; Ramsaran, Schlichting, & Frankland, 2019) was proposed to reconcile the conflicting aspects of these theories. Critical periods are developmental windows during which increased plasticity drives cerebral maturation in an experience-dependent manner (Hensch, 2005; Ramsaran et al., 2019).

In this chapter, we describe normal episodic memory development during the infantile and childhood amnesia periods, with a focus on the relation between memory development and hippocampal maturation. Then, we discuss how the concept of critical periods could explain pivotal aspects of infantile and childhood amnesia and how it could bring new insights to the study of disorders caused by adverse experiences during infancy and childhood.

Semantic Memory

In our model of memory processes, episodic memory and semantic memory are connected to each other by two arrows (Fig. 1). The first one reflects the fact that information is encoded in semantic memory before being encoding in episodic memory, whereas the second one reflects the process of memory semanticization of most episodic memories over time.

Early in sobriety, AUD patients present with impairment in novel semantic encoding and more specifically a deficit in acquiring new labels information. Learning new categories and features can be achieved by AUD patients but at a slower rate than controls. Compared with controls, patients invoke different and inefficient cognitive strategies to attempt to compensate for impaired episodic and working memory. Active and strategic learning appears therefore to be not up to par with controls, leading to the shallower (more fragile) encoding of the new information in episodic memory. AUD patients invoked specific cognitive processes, such as working memory abilities, in addition to those invoked by controls. Previous neuroimaging studies already showed that AUD patients recruit higher-level cognitive systems to perform even simple tasks at control levels. Such alteration of trial and error new semantic learning may be compensated by the use of rehabilitation techniques. Errorless learning enabled AUD patients to normalize new label learning performance and resulted in as flexible information as trial and error.

Regarding the semantic network in AUD, a high level of anxiety has been shown to be associated with significant difficulties in inhibiting alcohol-related verbal stimuli. The authors interpreted the results as the effect of anxiety that may facilitate the activation of alcohol related concepts in semantic memory. In AUD patients, the alcohol-related network may have a particular status since behavioral and electrophysiological data revealed that patients process information linked to alcohol cues more rapidly than neutral cues. Such pattern is in accordance with the attention and memory bias toward alcohol-related information frequently reported in this population. Indeed, the use of alcohol-related stimuli in go/no-go or memory tasks interacts with performance.

Episodic and Semantic Memory: Distinctions and Interactions

Work with K.C. and other amnesic patients provided evidence that episodic and semantic memory are distinct within declarative memory, supported by dissociable brain structures. These distinctions were confirmed with the use of more sensitive, real-world measures, such as the Autobiographical Interview (Levine et al., 2002). With this measure, participants are asked to recall personal experiences from childhood and adolescence and through adulthood. Details of the memories are classified by trained scorers as event-specific (internal or episodic details) or as semantic, metacognitive, or unrelated to the event itself (external details). For example, a participant might retrieve episodic details associated with delivering a speech at a wedding, including the layout of the room, the facial expression of the groom, and the feelings that they experienced when no one seemed to appreciate their inside joke. They might also recall a semantic detail that the person who officiated the wedding had been ordained in Spain. However, some of the event-specific details become “semanticized,” having lost their episodic nature over time by virtue of being frequently retrieved, such as retelling the story of how the father of the groom fell asleep at the table (Cermak and O'Connor, 1983).

Work with amnesic patients continues to be influential in showing the ways in which episodic memory is distinct from semantic memory but also the intersection between these and other forms of memory, as the last example illustrates. Other evidence of a close connection between episodic and semantic memory comes from individuals with semantic dementia (i.e., temporal-variant frontotemporal dementia). These individuals appear to show an opposite pattern to individuals with hippocampal damage: impaired remote semantic memory but relatively spared episodic memory (McKinnon et al., 2006; Westmacott et al., 2001). An exception is preserved remote semantic memories that are personally significant or meaningful; the concept of an armchair might be meaningless to a person with semantic dementia unless it is an armchair that belongs to that person (Snowden et al., 1996).

Interactions between episodic and semantic memory do not only apply to remote memory. Even patient H.M., who had extensive MTL tissue removed bilaterally, was able to acquire a small amount of personal and general semantic information over the decades after his injury. He learned that he had a memory impairment and underwent brain surgery (Corkin, 2013). He had also acquired knowledge about a limited number of celebrities (O'Kane et al., 2004) and the basic floor plan of a house that he lived in for several years after his surgery (Corkin, 2002). Currently, it is unclear whether the lateral temporal cortex, the MTL, and/or residual hippocampus were responsible for autobiographical fact updating in H.M., as this form of semantic knowledge is related to the self but is not accompanied by the type of re-experiencing thought to be associated with episodic memory (Renoult et al., 2012; Grilli and Verfaellie, 2014). The dichotomy between episodic and semantic memory is increasingly blurry (Renoult et al., 2019), with amnesic patients continuing to provide valuable insight into the relationship between these and other forms of memory.

Language

The problems with language in Alzheimer's disease are usually attributable to the combination of an anomia plus an impairment of semantic memory. From a cognitive testing standpoint, two aspects of this impairment are typically found. First, there is an anomia, or impairment in naming uncommon objects. In very mild and mild Alzheimer's disease this impairment may be observed in tests such as the Boston Naming Test (Balthazar et al., 2008). In this test, line drawings of items are presented, with very common items shown at the beginning of the test (e.g., comb) and less common items shown toward the end of the test (e.g., trellis). Although the official Boston Naming Test consists of 60 items, 30- and 15-item versions have been successfully used. Note that only when patients reach the moderate stage of Alzheimer's disease do they typically show difficulty in naming the two items on the MMSE.

The second aspect of language that is typically impaired in Alzheimer's disease is the ability to generate an adequate number of words in 1 minute in certain semantic categories, such as “animals,” “fruits,” and “vegetables.” In fact, many patients with very mild and mild disease show impairment on these tests of word generation to categories, but perform normally on tests of word generation to letters. See Appendix A: Cognitive test and questionnaire forms, instructions, and normative data for additional discussion of these tests.

Task-switching versus TSSI organization in higher-order “visual” cortices

It is important to note that many of the studies reporting task-switching plasticity toward higher-order cognitive tasks, such as verbal memory, semantic and syntactic processing of language, or mathematical reasoning, in the early visual cortices of congenitally blind adults, also reported extensive crossmodal recruitment for these tasks beyond these early regions, across higher-order “visual” regions.109–112,116 These are the same regions for which TSSI recruitment has been shown. Then, how can these divergent findings be integrated together into a unified framework on sensory reorganization following blindness? Unfortunately, there are not many studies addressing this crucial question. A recent investigation by Kim et al.122 tested the (re)-organization properties of the visual word form area (VWFA), a region in the “ventral” visual stream repeatedly described as TSSI and responsive to symbol-to-phoneme conversions52,76,80 but that was also shown to be recruited by less specific linguistic tasks.109–112 Specifically, Kim and colleagues showed that the VWFA was responsive to both Braille letters and the grammatical complexity of auditory sentences in congenitally blind adults, whereas in sighted adults it was activated only during reading of print and not auditory sentences.122 The authors interpreted these results as evidence suggesting that the deprived visual cortex lost its selectivity to specific computations, supporting Bedny's proposal that the deprived visual cortex is pluripotent with the ability to take over a wide range of functions.115 In other words, Bedny proposes that brain specializations are constrained neither to a specific sensory modality (i.e., the natural selection account) nor to specific sensory-independent computations (i.e., the TSSI account), but rather that they are only constrained by preexistent connectivity patterns and by experience during critical periods early in development.115 Recently, she further refined her proposal by suggesting that the strongest weight to cortical repurposing is provided by experiences during critical periods rather than by connectivity biases.123 Specifically, Kanjlia and colleagues tested congenitally blind, late blind, and sighted controls in mathematics and language-related tasks manipulating cognitive load (i.e., all tasks that have been shown to recruit the deprived visual cortex).123 The authors also acquired resting-state data on the same participants.123 Their results indicated that, while resting-state functional connectivity between the deprived visual cortex and the rest of the brain was similar in the two blind groups, regional specialization for mathematics and language as well as load-dependent activity across the deprived visual cortices was observed only in congenital blindness.123 The authors concluded that there are critical periods for the repurposing of the visual pluripotent cortex, i.e., that experiences early in development play a crucial role in determining the properties of cortical specializations.123

However, there are numerous studies showing TSSI organization in the deprived higher-order visual cortices.80 Thus, which of the two organizational principles, namely TSSI or task-switching, is more dominant in shaping the organization of these cortices? A recent study from our laboratory provides initial results in answering this crucial question.124 Similarly to the work by Kim and colleagues,122 Sigalev et al. used fMRI to examine the (re)-organization properties of VWFA in congenitally blind,124after training on reading letters via an SSD. After SSD training, in congenitally blind participants, the VWFA responded only to SSD-presented words and not during an auditory semantic task.124 These results suggest that, with the appropriate training, TSSI organization may overcome task-switching plasticity.124 These findings are not conclusive in this matter as the authors did not test VWFA recruitment by semantics before the SSD training. Nonetheless, this study suggests the interesting working hypothesis that there might be indeed some predispositions to specific sensory-independent computations in the higher-order visual cortices as suggested by the TSSI account for brain (re)-organization. Furthermore, it suggests that such predispositions might be somewhat (re)-awakened or strengthened by task-specific training, even if such training is relatively short compared to the lifelong experience following task-switching in a given region, and even if the training is undertaken during adulthood. Future studies could investigate this issue more systematically, for instance, by performing longitudinal studies during which blind participants are scanned in both task-switching and TSSI-eliciting tasks before and after SSD task-specific learning.

Episodic and Semantic Autobiographical Memory

Autobiographical memory can be broken down into episodic and semantic subcomponents. This distinction is rooted in the work of Tulving (2002) who defined semantic memory as the ability to retrieve general facts that are abstract from the context in which they are learned (e.g., I remember that I was student at that university). On the other hand, episodic memory is the ability to retrieve specific events situated in time and space (e.g., I remember that lecture in which that professor explained memory function). Critically, episodic retrieval triggers what Tulving coins “mental time travel” (i.e., the ability to mentally project oneself in time to relive the past or even imagine the future). According to Tulving (2002), this ability triggers a state of “autonoetic consciousness” (i.e., the phenomenological sense of experiencing oneself being present at the time when the remembered event originally took place). Autonoetic consciousness can be contrasted with noetic consciousness, which characterizes semantic memory. Compared with autonoetic consciousness, noetic consciousness involves a more abstract sense of time as well as a general awareness of knowing and understanding the world and the self. The distinction between autonoetic and noetic consciousness also involves differences regarding several phenomenological features such as mental imagery. Mental imagery is the ability to create mental representations of the retrieved events and to manipulate these representations in the mind (Kosslyn et al., 2001). Mental imagery is intimately associated with the phenomenological experience of episodic memory: the more vivid the image, the stronger the autonoetic experience (Rubin, 2006; Vannucci et al., 2016). Mental imagery thus contributes to the specificity of autobiographical retrieval (Dewhurst and Conway, 1994; Williams et al., 1999), realness (Mazzoni and Memon, 2003), and vividness of autobiographical retrieval (D'Argembeau and Van der Linden, 2006; Rubin et al., 2003). The contribution of mental imagery to autobiographical construction was highlighted by Conway and Pleydell-Pearce (2000), who suggested that mental imagery provides rich contextual, perceptual, and/or sensorial information that acts as a powerful cue in the retrieval of specific and vivid autobiographical events.

Overall, autobiographical memory can be fundamentally divided into episodic and semantic subcomponents. While semantic autobiographical memories involve retrieval of general knowledge, episodic autobiographical memories involve retrieval of specific events and a stronger subjective experience involving significant time travel as well as vivid mental images of the retrieved events.

8.1.8.2 Category Learning in Alzheimer’s Disease (AD)

Alzheimer's disease is a progressive neurodegenerative disease that damages the hippocampus, posterolateral temporal-parietal cortex, and dorsolateral prefrontal cortex. These individuals commonly experience difficulties in retrieving semantic memory (Smith & Grossman, 2008). As a result, AD patients consistently show deficits in rule learning, and are impaired in some prototype distortion (dot pattern) tasks (Kéri et al., 1999). Koenig, Smith, Moore, Glosser, and Grossman (2007) ran AD patients, corticobasal degeneration (CBD) and age-matched controls in rule-based and similarity-based categories. AD patients were impaired at rule-based categorization, but achieved the same performance level as healthy controls for the similarity-based categories. This deficit in rule-based categorization was correlated with measures of executive functioning, suggesting a deficit in the cognitive resources necessary for rule-based category learning. It was also correlated with a measure of semantic memory. AD patients were unable to apply a rule even when they were reminded what the rule was. CBD patients were also impaired at rule-based categorization, but they were even more impaired at the similarity-based categories. In addition to AD, patients with frontotemporal dementia (which, as the name implies, affects primarily the frontal and temporal lobes) also displayed deficits in learning rule-based categories (Koenig, Smith, & Grossman, 2006).

The impact of Alzheimer’s and other degenerative diseases on procedural learning is less clear but AD appears to have little if any effect on procedural learning and memory (Eldridge, Masterman, & Knowlton, 2002). In this particular experiment, a probabilistic classification task was used, and AD patients’ performance did not differ significantly from that of healthy controls. They were, however, significantly impaired in an explicit memory test, suggesting that they may indeed have used the procedural system to perform this task.