Brain, Vol. 125, No. 9, 2044-2057,
September 2002
© 2002 Guarantors of Brain
The dynamic time course of semantic memory impairment in Alzheimer’s disease: clues from hyperpriming and hypopriming effects
1 Inserm E0218–Université de Caen, Laboratoire de Neuropsychologie, CHU Côte de Nacre, 2 Clinique Neurologique, Centre de la Mémoire, CHRU, Hôpital Roger Salengro, Lille, 3 Laboratoire de Psychologie Cognitive et Pathologique, Université de Caen, 4 Ecole Pratique des Hautes Etudes, Université René Descartes, Paris, France
Correspondence to: Prof. Francis Eustache, Inserm E0218–Université de Caen, Laboratoire de Neuropsychologie, CHU Côte de Nacre, 14033 Caen Cedex, France E-mail: neuropsycho{at}chu-caen.fr
Received July 19, 2001. Revised March 7, 2002. Accepted April 14, 2002.
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The nature of semantic memory deficit in Alzheimer’s disease is still a matter of controversy. To clarify this issue, we examined the evolution of semantic memory impairment in 24 Alzheimer’s disease patients by means of a longitudinal study. We used two semantic tasks, one explicit and the other implicit, to evaluate the integrity of the same concepts. The explicit task was a semantic knowledge task composed of naming and questions, involving superordinate and attribute knowledge of concepts. The implicit task, a lexical decision task, assessed semantic priming and allowed a very pure measurement of semantic memory. In this task, related pairs of words had coordinate (e.g. ‘tiger–lion’) or attribute (‘tiger–stripe’) relationships. In the coordinate relation between two words, the semantic priming performances were at first paradoxical: they increased abnormally (hyperpriming) before falling down, whereas in the attribute condition, the priming effects were first normal and then started to decrease in the final sessions (hypopriming). Compared with the semantic knowledge performance, these apparently disconcerting results reflect a coherent pattern of semantic memory degradation in Alzheimer’s disease that is a progressive deterioration starting with specific attribute information. The data reveal in an astonishing yet striking manner the dynamic semantic memory degradation in Alzheimer’s disease through the apparently paradoxical semantic priming effects.
Keywords: Alzheimer’s disease; longitudinal study; semantic memory; semantic priming
Abbreviations: DRS = dementia rating scale; MMSE = mini-mental state examination; SOA = stimulus-onset asynchrony
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Numerous studies have shown that semantic memory impairments may occur relatively early in the course of Alzheimer’s disease. Mildly demented patients with Alzheimer’s disease have been shown to be impaired on tests of object naming (Martin and Fedio, 1983
; Huff et al., 1986; Hodges et al., 1992), verbal fluency (Ober et al., 1986; Troster et al., 1989; Bayles et al., 1990; Hodges and Patterson, 1995; Salmon et al., 1999) or definitions (Hodges et al., 1996; Lambon Ralph et al., 1997). Nevertheless, the nature of the cognitive dysfunction responsible for these semantic processing impairments is still a matter of controversy. Some investigators (Chertkow et al., 1989, 1994; Chertkow and Bub, 1990; Hodges et al., 1992; Martin, 1992; Chan et al., 1993, 1995; Randolph et al., 1993; Binetti et al., 1995) argue that the semantic deficit stems from a gradual breakdown in the hierarchical organization of semantic knowledge (Collins and Quillian, 1969) as Alzheimer’s disease progresses. That is, whereas Alzheimer’s disease patients may retain semantic information about a given concept, they progressively lose knowledge of the concept’s specific attributes that constitute its meaning (Hodges et al., 1992). Another point of view (Ober and Shenaut, 1988; Nebes, 1989, 1992, 1994; Nebes et al., 1989; Bayles et al., 1991; Hartman, 1991) is that the store of semantic memory remains relatively intact in Alzheimer’s disease, and that the deficit is related to a decreased ability to access semantic information. However, many of the tasks used to explore semantic memory require sustained attention, active searching, working memory and overt retrieval in addition to the more basic processes of accessing and using information from semantic memory. Most importantly, these tasks lack sensitivity and do not allow one to establish accurately how and when the structure of semantic memory deteriorates.
Therefore, one of the methods used to investigate the organization of semantic memory with more precision is the single-word semantic priming paradigm. In this study, semantic priming refers specifically to a short-lived phenomenon that is usually considered a measure of semantic knowledge integrity. It does not refer to repetition priming with a delay between study and test that demands the participation of long-term memory in new learning. This paradigm allows one to assess semantic memory implicitly and thus minimizes non-semantic cognitive processes. Semantic priming effects refer to the modification of a stimulus processing behind the presentation of a related stimulus. These effects depend on semantic memory (Tulving, 1995) and require a semantic processing of the prime stimulus and/or a semantic relation between the prime and the target. Generally, in lexical decision or pronunciation tasks, a word (e.g. ‘chair’) is recognized faster if it is preceded by a semantically related word (‘table’) than by an unrelated word (‘horse’) (Meyer and Schvaneveldt, 1971; Fischler, 1977; Neely, 1977). This processing facilitation would depend on the automatic spreading activation through the semantic network (Collins and Loftus, 1975): the presentation of a prime activates its node in memory and this activation automatically spreads to related nodes, thus momentarily increasing their accessibility.
Semantic priming studies have yielded contradictory results in Alzheimer’s disease patients. Some authors have reported less-than-normal priming (hypopriming) in Alzheimer’s disease patients compared with controls (Ober and Shenaut, 1988; Salmon et al., 1988; Silveri et al., 1996), while others have found equivalent priming (Nebes et al., 1984; Ober et al., 1991) or even paradoxical increased priming effects (hyperpriming) in Alzheimer’s disease patients (Chertkow et al., 1989, 1994; Nebes et al., 1989; Balota and Duchek, 1991; Hartman, 1991; Balota et al., 1999; Bell et al., 2001). These diametrically opposed results may reflect not only differences in the methods used, but also clinical heterogeneity in the population samples studied. For example, the severity of dementia, and therefore of semantic deficits, differed from one study to another, and even within individual studies. Interestingly, some authors reported a combination of no priming and intact priming (e.g. Albert and Milberg, 1989; Bushell and Martin, 1997; Glosser et al., 1998) or both intact priming and hyperpriming (e.g. Margolin et al., 1996; Shenaut and Ober, 1996; Giffard et al., 2001).
The aim of the present study was, therefore, to investigate further the structure of semantic memory impairment in a sample of Alzheimer’s disease patients and to assess the relationships between semantic priming effects and semantic memory deficits by means of a longitudinal investigation, in which we were able to follow their subsequent evolution. Our approach relies on the fact that semantic memory is found to deteriorate progressively throughout the course of Alzheimer’s disease (Chan et al., 1997) as a result of the degradation of the neocortical association areas that are presumed to store the semantic representations. Loss of semantic knowledge is often confined to a limited number of items during the early stages of the disorder and then spreads as the disease progresses. This progressive deterioration of semantic information over time bears on the claim that conceptual structure is hierarchically organized (Collins and Quillian, 1969) from superordinate (e.g. ‘a canary is an animal’) to basic-level category (‘a canary, like a swallow or a raven, is a bird’) to specific features (‘a canary is yellow and can fly’). However, this traditional framework of hierarchical organization has been largely challenged. Computational models of semantic memory (Masson, 1995; Plaut, 1995; McRae et al., 1997) assume that conceptual knowledge is represented in a widely distributed network of low-level representational units (semantic features). In such models, concepts are represented by an overlap of features, and domain and category structure is based on similarity, captured in the degree to which semantic properties overlap rather than being distinctly encoded at a separate level (Devlin et al., 1998; Tyler and Moss, 2001). Nevertheless, to make the study more straightforward, the protocol has been realized within the framework of hierarchical conceptual structure, distinguishing several semantic levels.
We made the assumption that the semantic memory deteriorates in a sequential manner, at a subordinate level first and at a superordinate level thereafter (Warrington, 1975). For this reason, we combined two dimensions in the present study: the semantic level (subordinate versus superordinate) and time (longitudinal evaluation). Furthermore, we made the assumption that at the onset of the disease loss of semantic knowledge results in concepts becoming less well defined as their distinguishing attributes are eliminated, and later on there is a weakening of the formerly strong associations between related concepts in the semantic network.
To assess semantic memory in an Alzheimer’s disease group with mild-to-moderate dementia, we compared performances for both an implicit and an explicit task. These two tasks used the same stimuli targets. The implicit semantic memory task, a lexical decision task, assessed the semantic priming effects. In order to explore the hierarchical semantic hypothesis in the case of a conceptual degradation (Warrington and Shallice, 1979), we used two types of word pairs in the lexical decision task. Some word pairs shared a coordinate relationship (e.g. ‘tiger–lion’), i.e. the prime and the target belonged to the same semantic category and shared the same semantic level. Other word pairs were related according to an attribute condition (‘tiger–stripe’), i.e. the target was a specific attribute of the prime concept. The cognitive slowing process is characteristic of Alzheimer’s disease patients and of older subjects in general. In the present study, we controlled for the effects of this slowing with the help of a measurement expressed as a percentage of the priming effects. Moreover, we used the following automaticity criteria to minimize the intervention of attentional mechanisms, such as prelexical expectancy or postlexical semantic matching processes (for a review see Neely, 1991): (i) low proportion of related words (20%); (ii) short stimulus-onset asynchrony (SOA; 250 ms); (iii) low attention to the prime (the subject had just to answer the target); and (iv) the same proportion of word targets and non-word targets (Posner and Snyder, 1975; Neely, 1991). The explicit semantic memory test (a semantic knowledge task) was inspired by Martin (1987) and Desgranges et al. (1996). This task required decisions that involved conscious exploration in semantic memory and was designed to probe for knowledge across the hierarchy of semantic memory from the superordinate to the fine-grained subordinate level.
Our primary goal was to examine the hypothesis that the profile of semantic priming effects evolves in a dynamic manner as the semantic memory deterioration advances (see Chertkow et al., 1990), hence the interest in the longitudinal assessment of these two components. This evolution of priming effects in Alzheimer’s disease could account for the variety of profiles of semantic priming observed in the literature. It was, therefore, necessary to take into account the Alzheimer’s disease semantic deficit variability, which can hide different semantic priming profiles (see Albert and Milberg, 1989); a study that takes into account the performance levels with semantic knowledge could allow a better understanding of the observed priming effects. The hypothesis of this longitudinal study suggests at least three stages in the evolution of priming. (i) When semantic memory is still intact, semantic priming effects in patients and age-matched controls should be equivalent in both coordinate and attribute conditions, because semantic priming depends on semantic memory. (ii) When the specific attributes of concepts begin to be lost—whereas superordinate information is well preserved (e.g. the tiger and the lion are still known to be wild animals, but knowledge about their stripes and mane is lost)—the ability to distinguish between two very similar concepts is impaired. In the case of a coordinate relationship (‘tiger–lion’), priming effects do not only exist—since the words are semantically related through membership in their preserved superordinate class—but they are greater than in a control group (hyperpriming): specific attributes that characterize each concept are lost, hence a confusion, an overlapping between the two coordinate concepts (both are wild animals and also both have fur and are dangerous). Therefore, as suggested by Martin (1992), the semantic priming (‘tiger–lion’) would be treated by the patient as repetition priming (‘wild animal–wild animal’), the magnitude of which is greater than in the former. In the attribute condition (‘tiger–stripe’), at the same semantic deterioration level, the priming effect should decrease because knowledge of the specific attributes (‘stripe’) grows weaker. So, for the patient, the link between the words ‘tiger’ and ‘stripe’ is attenuated. This hypothesis has been confirmed by the study of Giffard and colleagues (Giffard et al., 2001) in a large sample of patients. (iii) As semantic memory deteriorates even more, not only the specific attributes, but also the concepts themselves, are lost. This is the reason why, in the attribute as well as in the coordinate condition, the semantic priming effects should decrease, because not even coordinate links between words will be remembered.
The first time that such a longitudinal investigation was conducted was in a study by Giffard and colleagues (Giffard et al., 2001). In that initial study, 20 normal control subjects and 53 Alzheimer’s disease patients were examined. In the present study, the normal control group’s scores are the same as those in the study of Giffard and colleagues—hence, control subjects have been examined once only—and the 24 patients examined through the four sessions were selected from the 53 Alzheimer’s disease patients. For statistical reasons, only patients who received the four sessions were included in the longitudinal analysis.
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The methodology used was similar to that employed in our previous study (Giffard et al., 2001
) and will, therefore, be described briefly.
Subjects
Twenty-four patients with probable Alzheimer’s disease and 20 elderly normal controls participated in the study. All subjects gave informed consent to the neuropsychological procedure, which was approved by the Ethical Committee of the University of Caen. Diagnosis of patients, who were all right-handed, was made according to the criteria of McKhann and colleagues (McKhann et al., 1984). At the time of the first session, all patients [mean age (± SD) 71 ± 5 years, range 61–78 years; six males and 18 females] underwent a neurological examination, standard laboratory studies, EEG and an extensive routine neuropsychological assessment. No abnormality, other than atrophy, was found on the CT scan or on MRI. The patients had no previous neurological or psychiatric history. The scores for the mini-mental state examination (MMSE; Folstein et al., 1975) and the dementia rating scale (Mattis, 1976) for the whole group are presented in Table 1. At the first session, the Alzheimer’s disease group was divided into two subgroups, A (n = 15) and B (n = 9), on the basis of their semantic knowledge impairment. A detailed explanation of this subdivision is addressed below.
|
Control subjects (mean age 71 ± 6 years, range 63–86 years; eight males and 12 females) received the same examinations as the patients but once only (see Giffard et al., 2001
). They were recruited in clubs for retired people and had no neurological or psychiatric disorders. The score for the MMSE was 27–30, and for the DRS it was 135–143 (means and standard deviations are presented in Table 1). The two groups were paired according to age throughout the four sessions (P = 0.76, P = 0.57, P = 0.31 and P = 0.23, respectively) and to the level of education: the minimum level was equivalent to ‘certificat d’études primaires’, a diploma generally obtained at 
14 years of age, after 8 years of primary education.
The patients were tested with the lexical decision and the semantic knowledge tasks four times, with 6 months between each test session. Thus, the interval between the initial and the fourth evaluation spanned 18 months. At each evaluation, patients were tested individually by the same trained psychometrist in a quiet room. The normal control subjects were tested once with the same protocol.
Lexical decision task
Briefly, there were 30 related pairs of words: 20 pairs of words semantically related and of the same semantic level (coordinate relation, e.g. ‘tiger–lion’) and 10 pairs of words in which the target was a specific attribute of the prime (attribute relation, e.g. ‘zebra–stripe’). These 30 related pairs concerned living and non-living things [seven animals, seven plants, two body parts and 14 objects (furniture, buildings, clothes, kitchen utensils and means of transport)]. It was important to offset the number of pairs concerning these two domains given the controversy regarding the impact of semantic degradation on living versus non-living things in Alzheimer’s disease (Caramazza and Shelton, 1998; Devlin et al., 1998; Garrard et al., 2001; Tyler and Moss, 2001). The 30 related pairs were included in a list of 300 pairs. All the primes were words. In order to minimize the intervention of postlexical attentional processes, the likelihood of encountering a word versus a non-word in the target position was 50%. Among the pairs in which the target was a word, 20% were semantically related (coordinate or attribute condition) and 80% shared no semantic or associative link, which helps prevent the subject’s expectancy about the nature of the target. The non-words, which were all pronounceable, were created by replacing one letter per syllable of a concrete word.
Stimuli were presented using the software Superlab 1.68 (Cedrus Corporation, Phoenix, AZ, USA) which allows response times to be measured accurately to 1 ms. During a trial, the subject saw on the screen a fixation point lasting 500 ms, followed by a prime word for 200 ms. Thereafter, the screen remained empty for 50 ms. The SOA was 250 ms, which was too short for the subject to anticipate the nature of the target. Subsequently, the target stimulus appeared and remained visible until a response was made. The screen was left empty for 1500 ms and another trial started. In order to favour the automaticity of the task, each subject was instructed to respond for the target only: if he/she recognized, in the series of letters, a French word, he/she had to press the ‘yes’ key as fast as possible with his/her dominant hand. If the series had no meaning for him, he/she pressed the ‘no’ key with his/her other hand.
Priming effects, based on differences of response times between unrelated and related conditions, are expressed as a percentage for each subject (priming effect divided by mean response times for the unrelated condition x100); this approach helps to prevent a slowing effect on the priming effects. The lexical decision task was always conducted first in the study.
Semantic knowledge task
Drawn from a protocol first described by Martin (1987) and Desgranges and colleagues (Desgranges et al., 1996, 1998), the semantic knowledge task was composed of naming, categorical knowledge and attribute knowledge of concepts. The 30 concepts assessed in this task belonged to four semantic categories (animals, plants, body parts and objects) and corresponded to the related targets in the lexical decision task.
First, we asked the subject to name 30 drawings (corresponding to the 30 concepts). If he/she failed, a recognition task of the noun was carried out: the correct noun and three others from the same semantic category were presented to the subject one after the other. Then, the subject was asked to answer a series of yes/no or multiple-choice questions for each of the 30 items. The first questions concerned the knowledge of the superordinate category (‘Does it occur naturally or is it man-made?’). The second questions concerned category membership (‘Is it an animal, a plant, a body part or an object?’). The third questions referred to the subcategory (‘Is it a domestic or a wild animal?’). Finally, there were three questions concerning specific attributes, which were either functional (‘Is it edible?’) or perceptive (‘Does it have a mane?’).
Statistical analyses
We used non-paired Student’s t-tests and ANOVA (analysis of variance), with Bonferroni correction for multiple tests, for cross-sectional analyses when comparing the control group with the Alzheimer’s disease group for each of the four sessions, and when comparing the dementia scores of the Alzheimer’s disease subgroup A with the Alzheimer’s disease subgroup B. ANOVAs with repeated measures were carried out in order to compare the performances of each session for the dementia severity index, the lexical decision and the semantic knowledge tasks in the two Alzheimer’s disease subgroups. To test the sphericity of the distribution of data, we used Mauchly’s test (Mauchly, 1940) and then corrected the degrees of freedom on the basis of the 
value of Greenhouse and Geisser (1959) or Huynh and Feldt (1976). Corrected degrees of freedom were truncated to the nearest integer. To investigate the relationships between semantic priming effects and response times and between semantic priming effects and dementia severity index, we used Pearson correlations.
|
Results |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The performance of the control and Alzheimer’s disease subjects with respect to dementia severity index and lexical decision and semantic knowledge tasks through the four sessions is shown in Table 1. Results of inter-group (controls versus Alzheimer’s disease group in each session) analyses for cross-sectional comparisons for each test are also shown.
Dementia severity index
In each session, the MMSE and DRS scores of the Alzheimer’s disease group were significantly worse (P < 0.0001) than those of the normal control group. In the Alzheimer’s disease group, one-way ANOVAs with repeated measures (four periods) on each dementia index indicated a significant period effect on the MMSE [F(3,67) = 5.17, P = 0.005] and DRS scores [F(3,62) = 4.68, P = 0.006]. On both tests, post hoc analyses (Bonferroni/Dunn’s test) showed significant differences between the 1st and 4th session, and between the 2nd and 4th session.
Lexical decision task
Accuracy data and response times
In keeping with other studies on semantic priming effects in Alzheimer’s disease (Chertkow et al., 1994; Giffard et al., 2001), we will report only results for ‘yes’ responses. The errors of the patients, which were in most cases immediately self-corrected, were at times 2, 3 and 4 of the study, more than those of the controls. Likewise, as in the study of Giffard and colleagues, in order to ensure that the performances were not influenced by extreme scores in each condition, response latencies >3 SD above each participant’s mean were treated as outliers, and the mean was calculated again.
At each time of testing, mean response times for the correct responses were submitted to a two-way ANOVA with repeated measures: two groups (controls and Alzheimer’s disease) x three conditions (coordinate, attribute, unrelated). In each session, the results showed a significant effect of group (P = 0.0002, P = 0.0006, P = 0.002 and P = 0.002, respectively), indicating that the response times of the Alzheimer’s disease patients were globally longer than the control group. The analyses also showed a significant effect of condition (P < 0.0001 in the four sessions): in the two groups, the responses in the experimental conditions were faster than in the unrelated condition. At the first and second times of testing, the interaction (group x condition) was significant (P = 0.001 and P = 0.002, respectively), and is explained by the fact that the response times of the patients were longer in the attribute than in the coordinate condition, whereas the controls showed the opposite profile.
Semantic priming effects
Coordinate condition. At times 1 and 2, non-paired Student’s t-tests indicated that the Alzheimer’s disease patients showed a larger amount of coordinate semantic priming than did the control subjects (10.6% versus 6.9% in the two sessions, respectively). The third and fourth assessments showed no more significant difference between the priming effects in the two groups of subjects. In the Alzheimer’s disease group, a one-way ANOVA with repeated measures (four sessions) indicated a significant period effect on the coordinate semantic priming effects [F(3,66) = 4.21, P = 0.009].
Attribute condition. The priming effects in the Alzheimer’s disease patients and in the control subjects differed significantly at the fourth session only, where the Alzheimer’s disease group showed significantly lesser attribute priming effects than control subjects (5.1% versus 9.6%, respectively). In the Alzheimer’s disease group, a one-way ANOVA with repeated measures (four sessions) showed no significant period effect [F(3,69) = 2.30, P = 0.085].
Correlations with response times and with dementia severity index. In the Alzheimer’s disease group, to confirm the absence of slowing effects on the semantic priming effects, we used Pearson correlations between semantic priming effects in the two related conditions and mean response times in the unrelated condition. In the coordinate condition, the results showed the absence of a significant relationship, except at the fourth assessment (r = 0.54, P = 0.006). In the attribute condition, and for each session, there was no significant correlation. In the control group also, no significant correlations were found between response times and semantic priming effects in both coordinate (r = 0.32, P = 0.17) and attribute conditions (r = 0.18, P = 0.45).
In the Alzheimer’s disease group, significant negative correlations were observed between priming effects in coordinate condition and DRS score at the first (r = –0.41, P = 0.05) and second sessions (r = –0.47, P = 0.02), with the more severely affected patients showing the largest priming effects. No significant correlations were found between priming effects in attribute condition and DRS score.
Semantic knowledge task
Non-paired Student’s t-tests indicated that through the four sessions of the longitudinal evaluation, the performances of the patients were inferior to those of the control subjects in naming and in the attribute knowledge test, but not in the categorical knowledge test. In the Alzheimer’s disease group, one-way ANOVAs with repeated measures (four sessions) were realized to test a period effect. The analyses indicated no significant effect of the period on naming [F(3,65) = 0.45, P = 0.71] and categorical knowledge scores [F(3,69) = 0.23, P = 0.88]. On the contrary, there was a significant period effect on the attribute knowledge performance [F(2,47) = 5.79, P = 0.006].
Alzheimer’s disease patients: division into two subgroups (subgroup A and subgroup B)
At the first session, the Alzheimer’s disease group as a whole had heterogeneous semantic performances. So, depending on the scores obtained at the first session, we divided the group of patients into two subgroups on the basis of their results to the questions that focused on the specific knowledge in the semantic knowledge task. The first subgroup was composed of 15 patients with scores that did not exceed >1 SD of the mean of the control group (patients without semantic deficits, subgroup A). The second subgroup was composed of nine patients with performances that exceeded >1 SD of the control group (patients with semantic deficits, subgroup B). We opted for this threshold of 1 SD below the mean of the control group, because the semantic knowledge task was not very sensitive, with a ceiling effect easily reached in controls. This subdivision of the Alzheimer’s disease group allows a more detailed understanding of the effect of the semantic deficits on semantic priming effects. During the follow-up period, we used the same subdivision to test and detect a contrast in the evolution of semantic priming effects and semantic memory deficits in each subgroup.
Demographic data
An ANOVA performed at the first session between the three groups of subjects (one control group and the two Alzheimer’s disease subgroups) showed no significant differences in terms of age [F(2,41) = 0.20, P = 0.82], education level [F(2,41) = 1.68, P = 0.20] or sex distribution [F(2,41) = 0.57, P = 0.57]. The performances of the two Alzheimer’s disease subgroups in MMSE and DRS tasks through the four sessions are shown in Table 2. Results of inter-subgroup (subgroup A versus subgroup B) analyses for cross-sectional comparisons (non-paired Student t-tests) and results of intra-subgroup analyses for longitudinal comparisons (one-way ANOVAs with repeated measures) are also shown.
|
Semantic knowledge task
The performances on the semantic knowledge task of the three groups of subjects are shown in Fig. 1. In the categorical knowledge subtest, to compare the scores between the three groups of subjects (control, subgroup A and subgroup B) through the four sessions, the performances were submitted to four one-factor ANOVAs. The analyses showed no group effect in any session. A two-way ANOVA with repeated measures of two Alzheimer’s disease subgroups (subgroups A and B) x sessions 1–4 showed neither effect of group [F(1, 22) = 0.667, P = 0.42] nor effect of period [F(3,66) = 0.19, P = 0.90], and the interaction group x period was not significant [F(3,66) = 1.78, P = 0.16].
|
On the contrary, in the attribute knowledge test, group effect was significant in each session (P < 0.0001). At time 1, the subgroup B had significantly worse attribute knowledge performances than the subgroup A and control group (because the two Alzheimer’s disease subgroups were divided on this basis). Post hoc comparisons (Bonferroni/Dunn’s test) showed that at times 2, 3 and 4 of the study, both subgroup A and subgroup B had worse scores than the control group, and differences between the two Alzheimer’s disease subgroups were no more significant. A two-way ANOVA with repeated measures of two Alzheimer’s disease subgroups (subgroups A and B) x sessions 1–4 showed a significant effect of group [F(1,22) = 10.23, P = 0.004] and a significant effect of session [F(2,43) = 5.52, P = 0.002], but no significant interaction between group and session [F(2,43) = 0.78, P = 0.50]. In subgroup A, post hoc analysis (Bonferroni/Dunn’s test) indicated a significant difference between attribute knowledge scores at time 1 and time 4 (P = 0.0047).
Lexical decision task
The mean response times of the three groups of subjects in the lexical decision task in each of the test sessions are shown in Table 3. Four two-way ANOVAs comparing response times in the four sessions [three groups (control, subgroup A, subgroup B) x three conditions (coordinate, attribute, unrelated)] showed significant effects of group and condition in the four analyses. The interactions group x condition were also significant, but only in the first and second sessions (P = 0.0025 and P = 0.01, respectively). Post hoc analyses (Bonferroni/Dunn’s test) showed that in the three conditions and for each session, response times of the two subgroups of patients were similar and significantly greater than those obtained by the controls. In the two subgroups of patients, a repeated measures ANOVA of Alzheimer’s disease subgroup (subgroup A and subgroup B) x condition of response times (coordinate, attribute, unrelated) x session 1–4 revealed a significant main effect of condition of response times [F(1,31) = 49.56, P = 0.0002], but no significant main effect of group [F(1,22) = 0.40, P = 0.53] and session [F(1,32) = 2.37, P = 0.12]. There were no significant group x condition of response times [F(1,31) = 0.87, P = 0.39], group x session [F(1,32) = 3.26, P = 0.07], condition of response times x session [F(6,132) = 1.26, P = 0.28], and group x condition of response times x session [F(6,132) = 0.51, P = 0.80] interaction effects.
|
The coordinate and attribute priming effects of the two subgroups of patients (subgroups A and B) through the four sessions of testing are shown in Fig. 2. In the coordinate condition, to compare the semantic priming effects between the three groups of subjects (control, subgroup A, subgroup B) through the four sessions, the performances were submitted to four, one-factor ANOVAs. The analyses showed significant differences between the three groups at the first and second sessions [F(2,41) = 5.24, P = 0.009 and F(2,41) = 3.86, P = 0.029, respectively], but not at the third and fourth assessments. At time 1, post hoc analysis indicated that subgroup B obtained significantly greater semantic priming effects than controls (hyperpriming) (P = 0.0025). In the second session of testing, while subgroup A began to show semantic deficits, this Alzheimer’s disease subgroup revealed a significant hyperpriming effect (P = 0.011 between subgroup A and the control group); the difference between subgroup B and controls was no longer significant (P = 0.09). At the third and fourth assessments, there were no longer significant differences between the three groups of subjects. In the two Alzheimer’s disease subgroups, in order to detect a period effect between the performances, the semantic priming effects were submitted to two-way ANOVAs with repeated measures. The analysis of the two Alzheimer’s disease subgroups (subgroups A and B) x sessions 1–4 showed no effect of group [F(1,22) = 0.78, P = 0.39], but a significant effect of period [F(3,63) = 4.14, P = 0.01]. The interaction group x session [F(3,63) = 0.993, P = 0.40] was not significant. In subgroup A, post hoc analyses showed significant differences between time 2 and time 3 (P = 0.0199), and between the second and the final assessment (P = 0.034). The post hoc analysis in the subgroup B indicated a significant difference between the first and the final session (P = 0.032).
|
In order to detect changing patterns of priming effects in the coordinate condition between the two subgroups of patients between sessions one and two, we conducted a two-way ANOVA with repeated measures (two groups x two sessions). There was no significant group effect [F(1,22) = 0.63, P = 0.43] and no significant session effect [F(1,22) = 0.27, P = 0.61], but the interaction group x session [F(1,22) = 4.65, P = 0.042] was significant.
In the attribute condition, the analyses of variance showed significant differences between the three groups of subjects, but only at the final testing period [F(2,41) = 4.75, P = 0.014]. Post hoc analyses indicated that, at this session, the priming effects of the two Alzheimer’s disease subgroups were significantly smaller than those of the control group (subgroup A, P = 0.0148; subgroup B, P = 0.0148). Nevertheless, at the third session, the scores of subgroup B tended to be smaller than that observed for the priming effects in the control group (P = 0.045). In the two Alzheimer’s disease subgroups, a two-way ANOVA with repeated measures of two groups (subgroups A and B) x sessions 1–4 indicated that there was neither a significant group effect [F(1,22) = 0.006, P = 0.94] nor a session effect [F(3, 66) = 2.56, P = 0.06], and the interaction group x session was not significant [F(3,66) = 0.621, P = 0.60].
|
Discussion |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
The results of the present longitudinal study show changing patterns of semantic priming effects over the course of Alzheimer’s disease. Moreover, the evolution of the priming effect does not always follow a linear curve. The experimental procedure allowed us to reveal a dissociation according to the condition (coordinate versus attribute): the priming effect observed in the attribute condition declines in a linear progression over the course of the disease, whereas in the coordinate condition the priming performance increases abnormally (hyperpriming) before falling. Furthermore, at the initial assessment of the study, the subdivision of the Alzheimer’s disease group according to the severity of the specific attribute deficits revealed that subgroup A (patients without semantic deficits at the first session, but with semantic deficits from the second session) shows exactly the same longitudinal profiles of evolution as that of subgroup B (patients with semantic deficits from the beginning of the evaluation), but with one session interval.
These data corroborate the findings of a transversal study performed by Giffard and colleagues (Giffard et al., 2001) in that hyperpriming is observed in the coordinate condition when semantic memory begins to deteriorate. In addition to this initial work, the present longitudinal study suggests that the hyperpriming phenomenon is not a set manifestation of semantic deficits, but just occurs when the semantic attributes start to be lost. Thereafter, in the coordinate condition, this paradoxical effect disappears: semantic priming effects decrease progressively along with semantic memory deterioration, which concerns not only specific attributes but actual concepts themselves. Furthermore, contrary to the present work, the transversal study was not able to reveal a hypopriming effect in the attribute condition since this semantic priming decrease appears later as the semantic memory deteriorates.
Our very controlled methodology, which was always the same during the longitudinal follow-up, allows us to consider that these patterns of semantic priming reflect changes in semantic memory deficits exclusively. Therefore, our finding calls into question many interpretations in studies showing normal priming effects or hyperpriming in Alzheimer’s disease. Some researchers (Nebes et al., 1984; Ober and Shenaut, 1995) consider that normal priming effects in Alzheimer’s disease reflect intact semantic knowledge. Nevertheless, this intuitively plausible hypothesis is right in part, but does not take into account the complexity of the conceptual structure (Moss et al., 1995; Nakamura et al., 2000; Giffard et al., 2001). Partial semantic degradation can still allow apparently normal priming effects: damage to stored representation may result in loss of some, but not all, of the specific attribute information. In this case, semantic priming effects, supported by the remaining intact features only, can be observed. At time interval 4 of our longitudinal investigation, in subgroup B we observed equivalent semantic priming as controls in the coordinate condition, whereas performance was inferior in the attribute condition (hypopriming). If knowledge of specific attributes had remained intact over the sessions, we would have observed equivalent semantic priming as that of controls in both conditions.
Concerning the paradoxical hyperpriming phenomenon, conflicting hypotheses have been advanced to explain this effect. According to Nebes and colleagues (Nebes et al., 1989), hyperpriming means that the semantic memory is intact and would just reflect an artefact of general slowing in Alzheimer’s disease: the slower the patient responds, the larger semantic priming effects he/she shows. The results of our longitudinal investigation contradict the explanation of Nebes and colleagues: we did not observe in the Alzheimer’s disease group a significant effect of session on response times in the lexical decision task, whereas the semantic priming effects changed significantly over time. We have also observed equivalent priming effects even though the patients responded significantly more slowly than the controls. Moreover, Pearson correlations indicated the absence of significant relationship between the magnitude of priming effects and response times, except at the fourth assessment in the coordinate condition (but the patients did not show hyperpriming at this time). Furthermore, in our study, semantic priming effects are expressed as a percentage of the unrelated condition response times, which minimizes any effect of slowing on the size of the priming effect.
According to others (Hartman, 1991; Ober et al., 1991; Silveri et al., 1996; Bell et al., 2001) hyperpriming may occur only in some experimental conditions that incite the subject to develop attentional strategies (pre-lexical expectancy or post-lexical semantic matching processes). These attentional strategies are deficient in Alzheimer’s disease. Ober and Shenaut (1995) observed in a meta-analysis that hyperpriming mainly occurs in paradigms, bringing into play these attentional processes (long SOAs, high proportions of related word pairs, high proportions of non-words). Thus, the hyperpriming effect would be the result of Alzheimer’s disease strategic deficits in semantic priming tasks. Attentional processes involve divided attention and working memory. Patients with Alzheimer’s disease have considerable difficulty in dividing attention between concurrently running cognitive operations. Thus, when semantic priming tasks involve long SOAs and a high proportion of related pairs, the patients (like normal subjects) are attempting to divide their attention among generate, search and decision processes. However, these multiple cognitive operations ‘hamper’ their impaired working memory, especially for the pairs in which words are unrelated and in which the potential targets will have to be inhibited. This will probably create for the patient a doubt about the decision asked (‘yes’ the target is a word or ‘yes’ the target is related to the prime). This confusion is expressed by a magnified difference between the response times in unrelated targets and in targets related to the prime (due mainly to increased slowing for unrelated pairs) for Alzheimer’s disease patients compared with controls.
In our study, the adjustment of the protocol (SOA = 250 ms, 20% of related words, response on the target only, same proportion of target words and target non-words) did not incite the subjects to engage expectancy or post-lexical processes. Therefore, the hyperpriming effect observed in our study was not the result of the intervention of such strategies. Moreover, with this same protocol, we have observed inferior, equivalent and superior semantic priming effects than with a control group.
Our results show that when semantic memory is still entirely preserved (subgroup A at the first session), semantic priming effects are similar in patients and their controls in both coordinate and attribute conditions, because semantic priming depends on semantic memory. Thereafter, a hyperpriming effect is encountered in the coordinate condition at the beginning of the semantic deficit only (subgroup B at the first session and subgroup A at the second session). These results agree with those of Chertkow et al. (1989) concerning words sharing a coordinate relation, where a hyperpriming effect is observed for the degraded items only. In our study, the hyperpriming effect reflects a deterioration of semantic memory and, more specifically, a storage deficit for specific attribute information: from the onset of the dementia, semantic representations deteriorate progressively, affecting the specific attributes first, with preservation of general knowledge (Martin and Fedio, 1983). The distinction between coordinate concepts is therefore more and more difficult because their specific attributes, which allow them to be distinguished, are lost. Without being exactly the same, hyperpriming could be close to a repetition priming (in which the prime and the target are the same) in which the intensity is greater than semantic priming (Martin, 1992). Hyperpriming does not appear to be an artefact of overall slowing or the result of attentional processes intervention, but rather, as argued here and by Giffard et al. (2001), it may well be a direct consequence of degraded semantic representations.
However, according to our hypothesis, semantic priming effects in both conditions should have decreased at the same time as general knowledge deterioration. In the categorical knowledge test, we did not observe in either of the Alzheimer’s disease subgroups, a significant difference compared with control subjects at any session. This absence of difference may be explained by the fact that semantic memory deficits are measured with a task composed of yes/no and multiple choice questions and for which the answers are very straightforward, and therefore a ceiling effect is quickly reached. This task would therefore not be sensitive enough to detect general knowledge deficits. However, we can suppose that as semantic memory deteriorates even more, semantic priming effects in both coordinate and attribute conditions would continue to decrease (see Fig. 2, dotted lines).
Otherwise, we made the assumption that at the beginning of the semantic deterioration, while patients showed hyperpriming in the coordinate condition, we should have observed simultaneously a semantic priming decrease in the attribute condition, when in fact the performances were still normal (Fig. 3). To explain this pattern, we can suppose that in Alzheimer’s disease, features of concepts would not be lost in an all-or-none manner, but the loss could be progressive and incomplete at the start of the disease. Computational models based upon distributed networks, in which each concept is represented by an overlap of features (e.g. the concept ‘tiger’ is represented by the features ‘animal’, ‘wild’, ‘four legs’, ‘fur’ and ‘stripes’), could allow a better understanding of our results than the traditional framework of hierarchical organization (Collins and Quillian, 1969) does. Such connectionist models assume that category structure is based on similarity, captured in the degree to which semantic properties overlap. Thus, ‘tiger’ and ‘lion’ belong to the same category (animal) and are semantically close because they share a large number of category-relevant properties (‘wild’, ‘four legs’, ‘fur’). However, they can be distinguished by some specific features (for example, the stripes of the tiger and the mane of the lion). Therefore, there would exist two kinds of attributes: those that are common to lots of concepts and that also tend to co-occur across exemplars, and those that are specific to one concept and that usually occur in isolation. In Alzheimer’s disease, common features would be preserved longer than distinctive features (Devlin et al., 1998).
|
Following such a conception, the hyperpriming effect observed in the coordinate condition could just be caused by a loss of those distinctive features. The representations of the tiger and the lion would therefore only be characterized by attributes like ‘four legs’, ‘wild’ and ‘fur’, which are shared by the two concepts. Thus, these two concepts would become synonyms at that point of the semantic deterioration. In the attribute condition, semantic priming effects would still be normal because the pairs in this condition are mainly composed of common features. Thereafter, as semantic memory deteriorates even more, not only the distinctive attributes, but also those shared by the two concepts, would be progressively altered. These two concepts would then become even less close. This could explain why, in the coordinate and attribute conditions, we observed a decrease in semantic priming effects.
The present study demonstrates that the presence of semantic priming effects in Alzheimer’s disease does not necessarily prove integrity of the semantic memory, and that on the contrary, considering models of semantic memory we can interpret the paradoxical evolution of semantic priming effects as a progressive deterioration of semantic memory. Following hierarchical models, increased semantic priming (hyperpriming) reflects the loss of specific information represented at lower hierarchical levels in spite of the integrity of general knowledge represented at a higher level. Considering computational models, hyperpriming reflects progressive loss of semantic features, to start with by distinctive features and thereafter by shared properties. Therefore, the results of this longitudinal study suggest that the profile of semantic priming effects in Alzheimer’s disease evolves in a dynamic manner and is dependent on the level of semantic memory deterioration. The longitudinal profiles observed might explain in part the conflicting results noted in the literature. This method of investigation leads to a very precise and accurate measurement of semantic deficits in patients with Alzheimer’s disease. Overall, our findings are an example of a way in which a disease may temporarily result in a ‘supra-normal’ performance on specific tests, through a modular loss of function within an organized cognitive system.
|
Acknowledgements |
|---|
We wish to thank Drs O. Letortu and S. Schaeffer for the recruitment of the patients, Dr J. Segui, Professor F. Viader, Professor J. C. Baron and Mrs J. Lambert for their methodological advice, Professor S. Faure, Drs P. Piolino and S. Rossi for their statistical advice, and Dr A. R. Young for revising the English style. This work was supported by the France-Alzheimer association and by a grant from the Ministère de l’Education Nationale de la Recherche et de la Technologie.
|
References |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsDiscussionReferences |
|---|
Albert M, Milberg W. Semantic processing in patients with Alzheimer’s disease. Brain Lang 1989; 37: 163–71.[Web of Science][Medline]
Balota DA, Duchek JM. Semantic priming effects, lexical repetition effects, and contextual disambiguation effects in healthy aged individuals and individuals with senile dementia of the Alzheimer type. Brain Lang 1991; 40: 181–201.[Web of Science][Medline]
Balota DA, Watson JM, Duchek JM, Ferraro FR. Cross-modal semantic and homograph priming in healthy young, healthy old, and in Alzheimer’s disease individuals. J Int Neuropsychol Soc 1999; 5: 626–40.[Web of Science][Medline]
Bayles KA, Tomoeda CK, Trosset MW. Naming and categorical knowledge in Alzheimer’s disease: the process of semantic memory deterioration. Brain Lang 1990; 39: 498–510.[Web of Science][Medline]
Bayles KA, Tomoeda CK, Kaszniak AW, Trosset MW. Alzheimer’s disease effects on semantic memory: loss of structure or impaired processing? J Cogn Neurosci 1991; 3: 166–82.
Bell EE, Chenery HJ, Ingram JCL. Semantic priming in Alzheimer’s dementia: evidence for dissociation of automatic and attentional processes. Brain Lang 2001; 76: 130–44.[Web of Science][Medline]
Binetti G, Magni E, Cappa SF, Padovani A, Bianchetti A, Trabucchi M. Semantic memory in Alzheimer’s disease: an analysis of category fluency. J Clin Exp Neuropsychol 1995; 17: 82–9.[Web of Science][Medline]
Bushell CM, Martin A. Automatic semantic priming of nouns and verbs in patients with Alzheimer’s disease. Neuropsychologia 1997; 35: 1059–67.[Web of Science][Medline]
Caramazza A, Shelton JR. Domain-specific knowledge systems in the brain: the animate-inanimate distinction. [Review]. J Cogn Neurosci 1998; 10: 1–34.[Web of Science][Medline]
Chan AS, Butters N, Paulsen JS, Salmon DP, Swenson MR, Maloney LT. An assessment of the semantic network in patients with Alzheimer’s disease. J Cogn Neurosci 1993; 5: 254–61.[Web of Science]
Chan AS, Salmon DP, Butters N, Johnson SA. Semantic network abnormality predicts rate of cognitive decline in patients with probable Alzheimer’s disease. J Int Neuropsychol Soc 1995; 1: 297–303.[Medline]
Chan AS, Butters N, Salmon DP. The deterioration of semantic networks in patients with Alzheimer’s disease: a cross-sectional study. Neuropsychologia 1997; 35: 241–8.[Web of Science][Medline]
Chertkow H, Bub D. Semantic memory loss in dementia of Alzheimer’s type: what do various measures measure? Brain 1990; 113: 397–417.
Chertkow H, Bub D, Seidenberg M. Priming and semantic memory loss in Alzheimer’s disease. Brain Lang 1989; 36: 420–46.[Web of Science][Medline]
Chertkow H, Bruemmer A, Merling A. Evolution in semantic priming changes in dementia of the Alzheimer’s type. J Clin Exp Neuropsychol 1990; 12: 18.
Chertkow H, Bub D, Bergman H, Bruemmer A, Merling A, Rothfleisch J. Increased semantic priming in patients with dementia of the Alzheimer’s type. J Clin Exp Neuropsychol 1994; 16: 608–22.[Web of Science][Medline]
Collins AM, Loftus EF. A spreading-activation theory of semantic processing. Psychol Rev 1975; 82: 407–28.[Web of Science]
Collins AM, Quillian MR. Retrieval time from semantic memory. J Verb Learn Verb Behav 1969; 8: 240–7.
Desgranges B, Eustache F, Rioux P, de la Sayette V, Lechevalier B. Memory disorders in Alzheimer’s disease and the organization of human memory. Cortex 1996; 32: 387–412.[Web of Science][Medline]
Desgranges B, Baron JC, de la Sayette V, Petit-Taboué MC, Benali K, Landeau B, et al. The neural substrates of memory systems impairment in Alzheimer’s disease. A PET study of resting brain glucose utilization. Brain 1998; 121: 611–31.
Devlin JT, Gonnerman LM, Andersen ES, Seidenberg MS. Category-specific semantic deficits in focal and widespread brain damage: a computational account. J Cogn Neurosci 1998; 10: 77–94.[Web of Science][Medline]
Fischler I. Associative facilitation without expectancy in a lexical decision task. J Exp Psychol Hum Percept Perform 1977; 3: 18–26.[Web of Science]
Folstein MF, Folstein SE, McHugh PR. Mini-mental state. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–98.[Web of Science][Medline]
Garrard P, Lambon Ralph MA, Watson PC, Powis J, Patterson K, Hodges JR. Longitudinal profiles of semantic impairment for living and nonliving concepts in dementia of Alzheimer’s type. J Cogn Neurosci 2001; 13: 892–909.[Web of Science][Medline]
Giffard B, Desgranges B, Nore-Mary F, Lalevée C, de la Sayette V, Pasquier F, et al. The nature of semantic memory deficits in Alzheimer’s disease. New insights from hyperpriming effects. Brain 2001; 124: 1522–32.
Glosser G, Friedman RB, Grugan PK, Lee JH, Grossman M. Lexical semantic and associative priming in Alzheimer’s disease. Neuropsychology 1998; 12: 218–24.[Web of Science][Medline]
Greenhouse SW, Geisser S. On methods in the analysis of profile data. Psychometrika 1959; 24: 95–112.[Web of Science]
Hartman M. The use of semantic knowledge in Alzheimer’s disease: evidence for impairments of attention. Neuropsychologia 1991; 29: 213–28.[Web of Science][Medline]
Hodges JR, Patterson K. Is semantic memory consistently impaired early in the course of Alzheimer’s disease? Neuroanatomical and diagnostic implications. Neuropsychologia 1995; 33: 441–59.[Web of Science][Medline]
Hodges JR, Salmon DP, Butters N. Semantic memory impairment in Alzheimer’s disease: failure of access or degraded knowledge? Neuropsychologia 1992; 30: 301–14.[Web of Science][Medline]
Hodges JR, Patterson K, Graham N, Dawson K. Naming and knowing in dementia of Alzheimer’s type. Brain Lang 1996; 54: 302–25.[Web of Science][Medline]
Huff FJ, Corkin S, Growdon JH. Semantic impairment and anomia in Alzheimer’s disease. Brain Lang 1986; 28: 235–49.[Web of Science][Medline]
Huynh H, Feldt LS. Estimation of the Box correction for degrees of freedom from sample data in the randomized block and split plot designs. J Educ Stat 1976; 1: 69–82.
Lambon Ralph MA, Patterson K, Hodges JR. The relationship between naming and semantic knowledge for different categories in dementia of Alzheimer’s type. Neuropsychologia 1997; 35: 1251–60.[Web of Science][Medline]
Margolin DI, Pate DS, Friedrich FJ. Lexical priming by pictures and words in normal aging and in dementia of the Alzheimer’s type. Brain Lang 1996; 54: 275–301.[Web of Science][Medline]
Martin A. Representation of semantic and spatial knowledge in Alzheimer’s patients: implications for models of preserved learning in amnesia. J Clin Exp Neuropsychol 1987; 9: 191–224.[Web of Science][Medline]
Martin A. Semantic knowledge in patients with Alzheimer’s disease: evidence for degraded representations. In: Bäckman L, editor. Memory functioning in dementia. Amsterdam: Elsevier Science; 1992. p. 119–34.
Martin A, Fedio P. Word production and comprehension in Alzheimer’s disease: the breakdown of semantic knowledge. Brain Lang 1983; 19: 124–41.[Web of Science][Medline]
Masson MEJ. A distributed memory model of semantic priming. J Exp Psychol Learn Mem Cogn 1995; 21: 3–23.
Mattis S. Mental status examination for organic mental syndrome in the elderly patient. In: Bellak L, Karasu TB, editors. Geriatric psychiatry: a handbook for psychiatrists and primary care physicians. New York: Grune and Stratton; 1976. p. 77–121.
Mauchly JW. Significance test for sphericity of a normal n-variate distribution. Ann Math Stat 1940; 11: 204–9.
McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the NINCDS-ADRDA Work Group under the auspices of Department of Health and Human Services Task Force on Alzheimer’s Disease. Neurology 1984; 34: 939–44.
McRae K, de Sa VR, Seidenberg MS. On the nature and scope of featural representations of word meaning. J Exp Psychol Gen 1997; 126: 99–130.[Web of Science][Medline]
Meyer DE, Schvaneveldt RW. Facilitation in recognizing pairs of words: evidence of a dependence between retrieval operations. J Exp Psychol 1971; 90: 227–34.[Web of Science][Medline]
Moss HE, Tyler LK, Hodges JR, Patterson K. Exploring the loss of semantic memory in semantic dementia: evidence from a primed monitoring study. Neuropsychology 1995; 9: 16–26.
Nakamura H, Nakanishi M, Hamanaka T, Nakaaki S, Yoshida S. Semantic priming in patients with Alzheimer and semantic dementia. Cortex 2000; 36: 151–62.[Web of Science][Medline]
Nebes RD. Semantic memory in Alzheimer’s disease. [Review]. Psychol Bull 1989; 106: 377–94.[Web of Science][Medline]
Nebes RD. Cognitive dysfunction in Alzheimer’s disease. In: Craik FIM, Salthouse TA, editors. The handbook of aging and cognition. Hillsdale (NJ): Lawrence Erlbaum; 1992. p. 373–446.
Nebes RD. Contextual facilitation of lexical processing in Alzheimer’s disease: intralexical priming or sentence-level priming? J Clin Exp Neuropsychol 1994; 16: 489–97.[Web of Science][Medline]
Nebes RD, Martin DC, Horn LC. Sparing of semantic memory in Alzheimer’s disease. J Abnorm Psychol 1984; 93: 321–30.[Web of Science][Medline]
Nebes RD, Brady CB, Huff FJ. Automatic and attentional mechanisms of semantic priming in Alzheimer’s disease. J Clin Exp Neuropsychol 1989; 11: 219–30.[Web of Science][Medline]
Neely JH. Semantic priming and retrieval from lexical memory: roles of inhibitionless spreading activation and limited-capacity attention. J Exp Psychol Gen 1977; 106: 226–54.[Web of Science]
Neely JH. Semantic priming effects in visual word recognition: a selective review of current findings and theories. In: Besner D, Humphreys GW, editors. Basic processes in reading: visual word recognition. Hillsdale (NJ): Lawrence Erlbaum; 1991. p. 264–336.
Ober BA, Shenaut GK. Lexical decision and priming in Alzheimer’s disease. Neuropsychologia 1988; 26: 273–86.[Web of Science][Medline]
Ober BA, Shenaut GK. Semantic priming in Alzheimer’s disease: meta-analysis and theoretical evaluation. In: Allen PA, Bashore TR, editors. Age differences in word and language processing. Amsterdam: Elsevier; 1995. p. 247–71.
Ober BA, Dronkers NF, Koss E, Delis DC, Friedland RP. Retrieval from semantic memory in Alzheimer-type dementia. J Clin Exp Neuropsychol 1986; 8: 75–92.[Web of Science][Medline]
Ober BA, Shenaut GK, Jagust WJ, Stillman RC. Automatic semantic priming with various category relations in Alzheimer’s disease and normal aging. Psychol Aging 1991; 6: 647–60.[Web of Science][Medline]
Plaut DC. Double dissociation without modularity: evidence from connectionist neuropsychology. J Clin Exp Neuropsychol 1995; 17: 291–321.[Web of Science][Medline]
Posner MI, Snyder CRR. Attention and cognitive control. In: Solso RL, editor. Information processing and cognition: the Loyola Symposium. Hillsdale (NJ): Lawrence Erlbaum; 1975. p. 55–85.
Randolph C, Braun AR, Goldberg TE, Chase TN. Semantic fluency in Alzheimer’s, Parkinson’s and Huntington’s disease: dissociation of storage and retrieval failures. Neuropsychology 1993; 7: 82–8.
Salmon DP, Shimamura AP, Butters N, Smith S. Lexical and semantic priming deficits in patients with Alzheimer’s disease. J Clin Exp Neuropsychol 1988; 10: 477–94.[Web of Science][Medline]
Salmon DP, Heindel WC, Lange KL. Differential decline in word generation from phonemic and semantic categories during the course of Alzheimer’s disease: implications for the integrity of semantic memory. J Int Neuropsychol Soc 1999; 5: 692–703.[Web of Science][Medline]
Shenaut GK, Ober BA. Methodological control of semantic priming in Alzheimer’s disease. Psychol Aging 1996; 11: 443–8.[Web of Science][Medline]
Silveri MC, Monteleone D, Burani C, Tabossi P. Automatic semantic facilitation in Alzheimer’s disease. J Clin Exp Neuropsychol 1996; 18: 371–82.[Web of Science][Medline]
Troster AI, Salmon DP, McCullough D, Butters N. A comparison of the category fluency deficits associated with Alzheimer’s and Huntington’s disease. Brain Lang 1989; 37: 500–13.[Web of Science][Medline]
Tulving E. Organization of memory: quo vadis? In: Gazzaniga MS, editor. The cognitive neurosciences. Cambridge (MA): MIT Press; 1995. p. 839–47.
Tyler LK, Moss HE. Towards a distributed account of conceptual knowledge. Trends Cogn Sci 2001; 5: 244–52.[Web of Science][Medline]
Warrington EK. The selective impairment of semantic memory. Q J Exp Psychol 1975; 27: 635–57.[Web of Science][Medline]
Warrington EK, Shallice T. Semantic access dyslexia. Brain 1979; 102: 43–63.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Bosch-Domenech, R. Nagel, and J. V. Sanchez-Andres Prosocial Capabilities in Alzheimer's Patients J Gerontol B Psychol Sci Soc Sci, May 25, 2009; (2009) gbp034v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Hodges Alzheimer's centennial legacy: origins, landmarks and the current status of knowledge concerning cognitive aspects. Brain, November 1, 2006; 129(Pt 11): 2811 - 2822. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Manenti, C. Repetto, S. Bentrovato, A. Marcone, E. Bates, and S. F. Cappa The effects of ageing and Alzheimer's disease on semantic and gender priming Brain, October 1, 2004; 127(10): 2299 - 2306. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||