According to the findings proposed by Braak and Braak (1991), the authors demonstrated the development of AD neuropathology over six stages, they first observed the intracellular neurofribrilliary changes mainly confined to the
transentorhinal region, a transition zone between the proper entorhinal region and temporal isocortex, while only mild involvement of the hippocampus (CA1) at the stage I and II (i.e., the “transentorhinal stages). During the stage III and IV (i.e., the limbic stage), both entorhinal and transentorhinal regions are affected with mild to moderate involvement of hippocampus, but the involvement of isocorticlal regions
remains low. At the stage V and VI (i.e., the isocortical stages), isocortex is
devastating affected. The clinical manifestations for AD have been proved reliable and highly correlated with neuropathological evolution and brain structure changes (Jack et al., 2004; McKhann et al., 2011). Accordingly, the identification of the earlier neuropathological changes in transentorhinal stages based on related cognitive function assessments may provide opportunities for early detection in the high-risk group who may develop AD in the future.
Poor episodic memory performance is currently considered as the first clinical sign of AD and associated with lesions in entorhinal cortex (ERC) and hippocampus (HP) (Killiany et al., 2002). However, neuropathological changes in AD first affects neither in the ERC nor HP regions, but in their neighboring “transentorhinal cortex”.
This region corresponds to Brodmann’s area (BA) 35 (Brodmann, 2007; Insausti et al., 1998) and Van Hoesen and Pandya’s field 35a (Van Hoesen & Pandya, 1975) which is known as the medial part of the perirhinal cortex (PRC) (Taylor & Probst, 2008). The pathological changes in transentorhinal stages I-II may silently begin years or decades before the diagnosis of AD (Braak & Braak, 1991). Therefore, many studies have investigated the functional-anatomical associations between the PRC and different kinds of cognitive functioning, suggesting that the cognitive functions associated with the PRC may be promising for the early detection of AD.
With respect to the PRC, lying in the anteromedial temporal lobe (BA 35 and 36), play a critical role in memory processing, stimulus-stimulus association, semantic memory, and naming of objects (Davies et al., 2004; Murray & Bussey, 1999; Murray
& Richmond, 2001). Davies and colleagues demonstrated that PRC volume loss is associated with semantic memory but not with episodic memory.
Since the researcher proposed the episodic and semantic knowledge as the
distinct types of declarative memory (Tulving, 1972). Episodic memory refers to the capacity to recall the contents of specific events happened in the past or the particular space-time context. Semantic memory refers to the capacity to recall the conceptual knowledge about the world or living culture. In other words, semantic memory is culturally shared and also plays a role on using the stored attributes to help individuals to understand the meanings of daily sensory experiences (Juliana V. Baldo &
Shimamura, 1998; Hodges et al., 1992). That is to say that semantic memory represents verbal function aspects of long-term memory (Sumiyoshi et al., 2009).
Some studies indicated that both episodic memory and semantic memory rely on the integrity of temporal lobe while the former one additionally relies on frontal lobe (P.
C. Chen & Chang, 2016; Knowlton & Squire, 1995; Shimamura & Squire, 1987;
Larry R. Squire & Zola, 1998). Under this view, new information first presented as an event and through several rehearsals the new information could be organized into a
concept without the original context as semantic memory, which suggests that the episodic memory may be a gateway of semantic memory. In other words, once episodic memory impaired, semantic memory should also impaired. However, a different view is that episodic memory is not the critical component for semantic memory formation. For example, Tulving and Squire (1991) based on observation of amnestic patient that can still learn some semantic knowledge after several repetition so they proposed that new information can via perceptual systems to enter the semantic memory. Suzuki and Amaral (1994) support this notion and suggest that PRC is the area receives inputs from visual object processing stream and also from other unimodal and poly-modal sensory areas. In addition, Vargha-Khadem et al. (1997) demonstrated that semantic memory is relatively preserved in amnestic patient refined to hippocampus lesion. Furthermore, recent research have demonstrated that neuropsychological markers, such as semantic object memory impairments, associated with medial perirhinal and entorhinal cortex functioning were impaired about twelve years before diagnosis of AD (Hirni et al., 2016).
With respect to semantic memory performance in AD, a number of previous studies have demonstrated that semantic knowledge is impaired in AD patients (Bayles et al., 1989; Chan et al., 1993; Chan et al., 1997; Chan et al., 1995; Chan et al., 2001; Hodges et al., 1992; Nebes, 1989; Salmon, Butters, et al., 1999).
Furthermore, recent researches support the notion that impairment of semantic memory can occur in the prodromal stage of AD (Adlam et al., 2006; H. T. Chang et al., 2015; Gainotti et al., 2014; Hodges & Patterson, 1995; Joubert et al., 2010). For instance, Gainotti et al. (2014) reviewed neuropsychological predictors to evaluate the best candidates of conversion from MCI to AD, and the authors found that AD
converters showed impairment of episodic memory as well as semantic memory.
Hodges and colleagues (1992) assessed semantic knowledge in AD by using a battery of neuropsychological tests, and the results showed that AD patient were impaired on the category fluency task, picture naming, verbal word-picture matching, picture sorting and generation of verbal definitions. Furthermore, superordinate knowledge in sorting and definition tests are relatively preserved in patients with AD, while a disproportionate reduction in the lower-ordered categories exemplars generation was observed, reflecting a “bottom-up” breakdown in the semantic network (Tröster et al., 1989). Across different tests, there was an item-specific consistency between the errors made by each patient, which suggested a breakdown in the storage structure of semantic knowledge. Contrary to this view, Nebes (1989) argued that the stimulus contents (e.g., object naming and naming to definition) and requirements for the tests significantly influenced the magnitude and nature of semantic deficits in AD. For example, AD patients appear most impaired on those tasks that require them to
perform a directed and intentional search of their semantic memory, such as verbal fluency, object learning, word finding, word-stem completion, and generation of associates. That is to say that the performance of semantic memory may vary in the retrieval efforts which the task demands. Therefore, the author suggested that the AD patients have a retrieval deficit in assessing semantic knowledge while the
structure of semantic knowledge remains intact. In order to evaluate the controversy between the structural or retrieval deficits hypotheses, Chan and colleagues (1993) compared the cognitive maps, generated by multidimensional scaling technique and ADDTREE clustering analysis, between normal elderly controls, AD and
Huntington’s Disease (HD) for investigating the semantic network based on the
animal naming data in 60 sec of category fluency task. The results indicated that the semantic organization of AD is disrupted and characterized by abnormal association of animal names and forming uninterpretable clusters, rather than a deficiency in retrieving semantic information. Recently, a longitudinal study also supported that semantic memory degradation in the course of AD is due to the loss of information rather than the difficulty in assessing semantic information (Mårdh et al., 2013). In addition, Joubert et al. (2010) investigated the neural correlates in semantic memory impairment in aMCI and early AD, semantic deficits in both aMCI and AD were associated with atrophy in anterior temporal lobe and inferior prefrontal regions.
They also suggested that semantic impairment in aMCI may result from semantic knowledge breakdown and the difficulties in selecting, retrieving, and manipulating the information of semantic knowledge. Adlam et al. (2006) evaluated semantic knowledge in MCI and mild AD and found that MCI was impaired on the category fluency and object knowledge tests while mild AD showed additional impairment on object naming, comprehension, and semantic association. The results suggested that semantic memory impairments could occur in the early AD and preclinical stage and also proved that the category fluency task is suitable for reflecting the breakdown of semantic memory. Furthermore, Cerhan et al. (2002) demonstrated that category fluency task has better diagnostic utility in AD than the letter fluency task.