Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2011 Dec 1.
Published in final edited form as: Cogn Behav Neurol. 2010 Dec;23(4):231–239. doi: 10.1097/WNN.0b013e3181e61cf1

Visual Discrimination Predicts Naming and Semantic Association Accuracy in Alzheimer’s Disease

Stacy M Harnish 1,2,3, Jean Neils-Strunjas 1,4, James Eliassen 1, Jamie Reilly 2,3, Marcus Meinzer 2,3, John Greer Clark 1, Jane Joseph 5
PMCID: PMC3058303  NIHMSID: NIHMS221440  PMID: 21042208

Abstract

Objective

Language impairment is a common symptom of Alzheimer’s disease (AD) and is thought to be related to semantic processing. The present study examines the contribution of another process, namely visual perception, on measures of confrontation naming and semantic association abilities in persons with probable AD.

Methods

Twenty individuals with probable mild-moderate Alzheimer’s disease and twenty age-matched controls completed a battery of neuropsychological measures assessing visual perception, naming, and semantic association ability. Visual discrimination tasks that varied in the degree to which they likely accessed stored structural representations were used to gauge whether structural processing deficits could account for deficits in naming and in semantic association in AD.

Results

Visual discrimination abilities of nameable objects in AD strongly predicted performance on both picture naming and semantic association ability but lacked the same predictive value for controls. Although impaired, performance on visual discrimination tests of abstract shapes and novel faces showed no significant relationship with picture naming and semantic association. These results provide additional evidence to support that structural processing deficits exist in AD and may contribute to object recognition and naming deficits.

Conclusions

Our findings suggest that there is a common deficit in discrimination of pictures using nameable objects, picture naming and semantic association of pictures in AD. Disturbances in structural processing of pictured items may be associated with lexical-semantic impairment in AD due to degraded internal storage of structural knowledge.

Keywords: Alzheimer’s disease, Anomia, Semantic Memory, Visual Perception, Agnosia, Confrontation Naming

INTRODUCTION

Alzheimer’s disease (AD) represents the most common form of dementia in industrialized nations, currently affecting more than five million people in the United States alone.1 Communicative disorders are among the most functionally debilitating aspects of AD.2, 3 Yet, the development of principled language interventions for AD is hindered by a lack of consensus regarding the primary etiology of the associated naming impairment. Although our most common behavioral association with AD is that of impaired episodic memory, it has long been recognized that language disturbances (i.e. anomia) and visual perceptual impairments (i.e. apperceptive agnosia ) are also reliable markers of AD.4 Patients with AD tend to experience inexorable worsening of naming impairment as the disease progresses, the basis for which remains controversial. One theoretical position holds that core knowledge of word and object meaning is relatively intact in early AD but that patients experience deficits in lexical and/or conceptual retrieval.5, 6 An alternate position holds that anomia reflects loss of core semantic knowledge.7, 8 While much of this debate has focused on access versus storage accounts of anomia, comparatively little research has focused on the moderating influence of visual object recognition deficits. We do so here by investigating the degree to which pre-semantic visual processing deficits predict picture naming and semantic association abilities on common diagnostic neuropsychological measures.

Alzheimer’s Disease: Pathology and Cortical Visual Processing

Visual perceptual disturbances often precede the classic episodic memory impairments in AD.9–11 The specific locus of impairment within the visual processing system remains unclear.12 Initially, disturbances were attributed to higher level perceptual processing, as opposed to sensory deficits that impact global visual attributes such as line orientation or form. This conclusion of perceptual over sensory impairment is supported by studies demonstrating relatively intact visual acuity in AD as gauged by perceptual matching and various other visual discrimination tasks (e.g., shape detection, shape discrimination, size discrimination, hue discrimination and dot counting).13 The hypothesis of associative or higher level perceptual processing deficits in AD also receives converging anatomical support from post-mortem investigations demonstrating that the primary visual cortex remains intact relative to more anterior visual association areas (i.e., inferior temporal gyrus and ventral temporal cortex) that tend to show dense concentrations of neuritic plaques.14

There are a number of potential ways that visual perceptual impairment may impact lexical-semantic processing in AD. Visual deficits may impact confrontation naming because of weakened visual input15 or a degraded internal image or perceptual representation16, 17 of the pictured item. Either of these impairments could potentially compromise one’s ability to assimilate an incoming stimulus (e.g., large animal looking object) with a stored structural representation (e.g., global and local visual form of an elephant). Recent research suggests that lower level visual processing in AD may be predictive of the integrity of downstream cognitive processes that support naming and semantic processing.15 Moreover, such sensory deficits may not only contribute to naming, but may also play a role in the integrity of stored structural representations or stored descriptions in long-term visual memory of an object’s 3-D structure. Chronic misperception (or misrepresentation) of visual stimuli could in time degrade stored structural representations, resulting in more insidious disturbances. According to Humphreys & Riddoch18, visual perceptual processes constantly update and recalibrate long-term visual memory, which may deteriorate in the presence of a perceptual deficit.

Consistent with many psycholinguistic models of naming and semantic access19, 20, early sensory processes can potentially impact retrieval of the correct structural, semantic, and lexical representations. Over time visual perceptual deficits could inaccurately update long-term visual memory as a result of weak visual input21–23 and visual distinctive features. This hypothesis about the loss of visual distinctive features has often been invoked to explain the common pattern of category specific naming impairment seen in AD for naming natural kinds (e.g., fruits, animals) relative to manufactured artifacts.24–26 That is, computational investigations of the semantic structure of natural kinds demonstrate a high density of intercorrelated semantic features with high visual similarity (e.g., many animals have tails, fur, four legs, ears, etc) relative to a higher degree of dissimarity among tools.27 The subtle loss of visual distinctive features can impact the ability to distinguish among natural kinds and disproportionately impair naming among these category exemplars.24, 25, 28, 29

There is compelling evidence to suggest that knowledge of distinctive semantic features is vulnerable in AD.28, 30 Alathari and colleagues30 found that AD participants identified fewer features of objects than control subjects and tended to list more features that are shared among category members instead of distinguishing features. They concluded that degradation of stored knowledge resulted in AD participants conceptualizing more shared features among concepts than healthy elderly participants. Again, when distinctive semantic features are related to an object’s structure (i.e. horn distinguishing a rhinoceros from a hippopotamus) we propose that long-term visual perceptual impairment may degrade the internal representation of these structural features. Moreover, we suspect that this degraded knowledge of semantic features may differentially impair visual discrimination of pictured objects with high structural similarity that require additional structural feature knowledge when global form is not sufficient to decide whether they are the same or different. In other words, the visual system may use semantic knowledge as a top-down influence to interpret visual input when bottom-up information contains ambiguities.31, 32

Naming in Alzheimer’s Disease

Several studies have investigated naming errors in AD by classifying errors as visual, semantic or lexical in nature.7, 33 A common finding is that AD patients produce many semantic and/or thematic naming errors (i.e. zebra for horse). The coarse criteria by which errors are divided can potentially overlook interactions among perceptual and lexical-semantic processes. For example, a disturbance in visually perceiving distinctive features may impact an individual’s ability to access the appropriate semantic representation for an object (e.g., Chihuahua is named as Cat), and possibly result in selection of a similar item from the same semantic category.24 In turn, this type of coordinate naming error might reflect the bottom-up loss of a hierarchical semantic category structure when in fact visual perceptual deficits are a contributing factor. Errors are often only considered visual in nature if the incorrect name indicates a gross visual misrepresentation of the picture (i.e. parking garage for harmonica).

Since AD is a multifocal disorder, one must consider the possibility that visual perception and naming are two unrelated areas of concurrent decline. However, we find evidence of a possible link between the two processes in visual perception tasks that require discrimination of real objects. Joseph & Gathers34 found that when healthy individuals underwent fMRI while performing a visual discrimination task between line drawings, they recruited relatively more anterior regions of the fusiform gyrus when the two drawings had high structural similarity and relatively more posterior regions of the fusiform gyrus and inferior occipital cortex when the drawings were lower in structural similarity. Of interest, this posterior-anterior fMRI signal change that shifted with increased structural similarity of pictures occurred for naming as well.35 These results are also consistent with functional imaging work reported by Tyler and colleagues36 showing a posterior-anterior distinction in specificity needed to name items at either the level of superordinate domain (animal or tool) or at a more specific level (dog or cat) with increasing specificity of feature conjunctions as visual processing streams forward toward the perirhinal cortex along the fusiform and inferior temporal gyri (for converging evidence in animal models see Bussey37). Together these findings suggest that fMRI signal in mid-anterior fusiform areas are related to processing of detailed object structure because these regions are sensitive to pictures with high structural similarity and pictures that require high specificity of structural processing, but not to other types of visual similarity (i.e. color38). Damage to this common neural substrate devoted to processing of an objects’ structure may impact performance on picture discrimination, matching and naming tasks.

Visual discrimination tasks that vary in the degree to which they likely access structural and semantic knowledge may provide insight into potential visual perceptual or object recognition deficits in AD that may impact picture naming. Visual discrimination tasks that require matching of novel, complex shapes should not depend on access to a structural description system since neither global form nor local visual processing should evoke a specific stored structural representation, semantic representation, or lexical entry. Similarly, processing of novel faces should not evoke structural, semantic or lexical information to help guide the process of discrimination. On the other hand, visual discrimination tasks that require an individual to determine if two line drawings of real objects are of the same or different object in different views require accurate low level visual perceptual processing and reference to a stored structural representation to help guide the decision.

Aims of the Current Work

Our aim here is to determine whether visual discrimination abilities are predictive of naming and semantic association disturbances in AD. To better gauge the effect of visual perceptual disturbances on pre-semantic processing in AD, we examined correlations between specific measures of visual discrimination that vary in the degree to which they likely access structural and semantic knowledge and measures of naming and semantic association ability. Degraded semantic knowledge is well documented in this population from studies that used non-visual tasks (such as naming to definition, feature listing and priming)30, 39 and likely contributes to errors in picture naming. Thus, our goal was to determine the relative degree to which visual perceptual abilities may be related to picture naming deficits in AD.

We propose that if sensory visual deficits are the primary visual contributor to picture naming impairment, then performance on all tests of visual discrimination, including those that do not involve access to a structural description system (ie. Benton Visual Form Discrimination Test, BVFDT and Benton Facial Recognition Test, BFRT), should show marked impairment that varies with naming abilities. If higher level impairment in access to stored structural representations are the primary cause of picture naming difficulties, then visual perceptual tests that do not rely on these systems (BVFDT, BFRT) may be relatively spared (unless there is concurrent, but unrelated degradation to this system), but performance on tests that may require access to this information (Visual Discrimination Task34) should show impairment that correlates with naming impairment.

Our hypotheses are: (1) Performance on measures of visual discrimination will correlate with performance on measures of lexical-semantic abilities in AD participants. (2) Visual perceptual tasks that likely require access to a structural representation system will account for the most variance in predicting picture naming performance or semantic association performance in AD participants, indicating a common degraded process required to complete each of these tasks. (3) Individuals with AD will be more impaired than elderly control subjects on discriminating between pictures that have high structural similarity over low structural similarity on the Visual Discrimination Task because of a degraded internal structural description system.

MATERIALS AND METHODS

Participants

Participants (N=40) included healthy community dwelling elderly controls (n=20) and individuals with probable AD (n=20) recruited from extended care facilities, adult day programs, and an AD support group. Patients and controls were similar in their distributions of gender and age. The age range of the elderly control group and the AD group were 69–87 [mean=82] and 67–96 [mean=82], respectively. The AD group included individuals who fit the criteria for mild-moderate AD and had a score between the range of 11–26 on the Mini Mental State Exam.40

Inclusion criteria were natural or correctable visual acuity equal or better than 20/40; Mini Mental State Exam40 scores of 25–30 for control subjects and 10–24 for cognitively impaired subjects in the AD group (consistent with mild to moderate cognitive impairment); ability to demonstrate understanding of completing a visual discrimination practice test, whereby same-different judgments of line drawings must be made; no electroconvulsive therapy treatment within 2 years of recruitment or ≥10 lifetime electroconvulsive treatments at any time; no clinical signs of stroke or a Hachinski Ischemic Scale score ≥5; no history of neurological dysfunction as indicated by significant head trauma, migraine, seizures or developmental delay; and no history of visual or ocular problems including, but not limited to, artificial lens implants, color blindness or macular degeneration. Two AD participants scored above the upper limit for MMSE (26), but had a formal diagnosis of AD and were therefore included in the study.

Potential subjects were screened for eligibility using a screening checklist and by completing a form related to medical history. Caregivers or family members completed the medical history forms for AD subjects. Testing took approximately 1.5 to 2 hours for elderly and AD subjects across 1 to 3 testing sessions depending on subject fatigue. Subjects were either tested in their homes, adult day center, or a comfortable place for them such as a labor union hall. All subjects participated in the following neuropsychological tests to document abilities related to cognition, semantic knowledge, naming and visual perception.

  1. The Mini Mental State Exam (MMSE)40 was used to screen for cognitive impairment. Subjects were asked a series of questions related to orientation, attention, recall, repetition, comprehension, reading and writing.

  2. The Pyramids & Palm Trees test (P&PT)41 assessed semantic knowledge by asking the participant to match pictures based on meaning (e.g. [anchor] matches with [ship] but not with [canoe]).

  3. The Boston Naming Test (BNT)42 assessed confrontation naming by asking the participant to name drawings of objects (e.g. mushroom).

  4. The Benton Visual Form Discrimination Test (BVFDT)43 required the individual to match patterns based on complex shapes and spatial construction.

  5. The Benton Facial Recognition Test (BFRT)43 required the participant to match pictures of faces based on facial features alone.

  6. The Visual Discrimination Task is a task used in previous research to investigate the functional organization of the human occipitotemporal cortex34, an area important for object recognition.The visual discrimination task involved the simultaneous presentation of two black and white line drawings of animals (e.g. mammals, reptiles, birds) or fruits/vegetables that were of the same or different referents in different poses or views. See Figure 1. Structural similarity of the objects was determined based on judgments of healthy young adults in other experiments.44, 45 The pictures were black line drawings on white 5-inch squares presented side by side. The line drawings were approximately 4 inches in length or width depending on the shape of the drawing. The white squares were presented on a gray rectangle projected on a black computer screen.

Figure 1.

Figure 1

Open in a new tab

Query and picture presentation for the visual discrimination task

The participants were told to indicate verbally whether the two pictures were of the same or different object by saying “same” or “different”. The examiner then indicated the participants’ response on the laptop computer. All responses were recorded by E-Prime 2.0 software (Psychology Software Tools, www.pstnet.com) to obtain response choices and accuracy. A query screen that read “Same or Different” in 48 point font appeared for 2 seconds prior to each picture pair. Each trial was 6 seconds in duration including a 2 second “Same or Different” instruction and a 4 second presentation of the picture pair. Participants could respond at any time during the 4-second presentation. After control subjects participated in the study, it was determined that individuals with AD would not be able to complete the task in the 4-second time limit. As we were not investigating naming latency, we decided to allow additional time for visual discrimination in AD subjects in order to provide better insight into behavioral performance. Therefore, AD subjects were given unlimited time to discriminate between the picture pairs. This precaution was also included to prevent potential frustration associated with increased processing time for individuals with AD. Individuals with AD were also given the verbal prompt “Are these the same things or different things?” to elicit a response if they did not automatically respond to the visual query and picture pairs.

The paradigm was presented in a block design with each block consisting of either 2 high similarity pairs of pictures (HIGH), or 2 low similarity pairs of pictures (LOW). See Figure 2 for examples of each. In addition, one pair of pictures in each block was of the same referent (SAME), but in different views or poses. Therefore, each block had three pairs of pictures, one of which was for the same referent in different poses and two of which were of different objects that were either high similarity or low similarity. For example, four blocks of presentation may have consisted of the following pairs of pictures: HIGH-SAME-HIGH, LOW-LOW-SAME, HIGH-HIGH-SAME, SAME-LOW-LOW.

Figure 2.

Figure 2

Open in a new tab

Sample low and high similarity picture pairs for the visual discrimination task

Caption: Low similarity comparisons (e.g. frog and dolphin) are less structurally similar than high similarity comparisons (e.g. elephant and buffalo)

The subjects were instructed to identify if the pairs of pictures were of the same thing or different things. Because the stimuli were presented simultaneously, no memory load was involved. Also, no feedback was given to subjects as to the correctness of responses, thereby reducing potential frustration associated with providing incorrect responses. The subjects participated in two runs of 30 pairs of pictures each. There were a total of 60 pairs of pictures, 20 of the same referent, 20 low similarity, and 20 high similarity.

Each subject underwent training on the Visual Discrimination Task prior to beginning the task. During the training, the examiner presented two pictures simultaneously and asked if the pictures were of the same or different object. The participant was reminded that two pictures of the same object may be in different poses or views. In this case, they are still of the same object and the participant should say, “same.” Five pairs of objects were presented and subjects were given feedback as to the correctness of their response during training. For example, the examiner might say, “Yes, that is correct, this picture is of a dog and this picture is of a dog. So, ‘same’ is the correct response.” Or “This picture is of a dog and this picture is of a pig. So, ‘different’ is the correct response.”

RESULTS

AD patients performed worse than elderly controls on all five measures as revealed by MANOVA: BNT [F(1, 38)=18.40, p<0.001], P&PT [F(1, 38)=14.35, p<0.001], BVFDT [F(1, 38)=74.29, p<0.001], BFRT[F(1,38)=14.10, p<0.001] and Visual Discrimination Task [F(1,38)=30.88, p<0.001]. See Table 1 for AD and elderly control mean scores on neuropsychological measures.

Table 1.

Mean Scores on Neuropsychological Tests by Group

GROUP (max score) BNT* (60) P&PT* (52) VisDis* (60) BVFDT* (32) BFRT* (54) MMSE* (30)
Normative scores (n=51)
44.7 mean
9.6 SD		 (n=13)
49–52
51 mean 		(n=85)
23–32
29.73 mean 		(n=286)
34–54
45.4 mean
3.96 SD 		(n=96)
22–30
27 mean
0.9 SD
ELDERLY CONTROL (n=20) 43–59
53.45 mean
5.38 SD		 44–52
49.25 mean
2.59 SD		 48–59
56.5 mean
2.89 SD		 24–32
28.55 mean
2.48 SD		 36–51
45.85 mean
3.88 SD		 25–30
27.15 mean
1.79 SD
AD (n=20) 9–58
38.85 mean
14.24 SD		 24–51
43.10 mean
6.78 SD		 39–57
48.90 mean
5.39 SD		 13–32
20.40 mean
3.42 SD		 29–51
40.00 mean
5.79 SD		 14–26
18.80 mean
3.61 SD
Open in a new tab

Normative data for the tests: MMSE (Folstein et al., 1975), P&PT (Howard & Patterson, 1992), BVFDT and BFRT (Benton et al., 1983), BNT (Welch et al., 1996).

*

Indicates a significant difference between elderly control and AD groups

Performance on measures of visual discrimination will correlate with performance on measures of lexical-semantic abilities in AD participants.

AD subjects’ scores were entered into a correlation analysis to determine whether there were significant correlations between performance on any of the visual perceptual measures with the BNT and/or the P&PT. Results indicated that performance on the Visual Discrimination Task that included pictures of animals, vegetables, and fruit was significantly correlated with performance on the tests of lexical semantic ability (BNT and PT&P; Figure 3, Figure 4). The Visual Discrimination Task also correlated with the other tests of visual perception (BVFDT and BFRT), but this correlation did not remain significant after applying Bonferroni’s correction for multiple comparisons. See Table 2 for correlations.

Figure 3.

Figure 3

Open in a new tab

Correlation between AD participants’ scores on the Visual Discrimination Task and the BNT

Figure 4.

Figure 4

Open in a new tab

Correlation between AD participants’ scores on the Visual Discrimination Task and the P&PT.

Table 2.

Correlations for AD participants

BNT P&PT VisDis BVFDT BFRT
BNT 1 .724**
.000		 .609**
.004		 .260
.268		 .351
.130
P&PT .724**
.000		 1 .781**
.000		 .340
.142		 .502*
.024
VisDis .609**
.004		 .781**
.000		 1 .444*
.050		 .538*
.014
BVFDT .260
.268		 .340
.142		 .444*
.050		 1 .526*
.017
BFRT .351
.130		 .502*
.024		 .538*
.014		 .526*
.017		 1
Open in a new tab

Significance levels are below each R value in the chart. Only correlations significant at the 0.01 level remained significant when Bonferroni’s correction for multiple comparisons was applied.

**

Pearson correlations significant at the 0.01 level (2 tailed).

*

Pearson correlations significant at the 0.05 level (2 tailed).

The same correlation analyses for the elderly control subjects yielded no significant correlations among performance on any of the measures. (See Table 3).

Table 3.

Correlations for elderly control subjects

BNT P&PT VisDis BVFDT BFRT
BNT 1 .354
.126		 .391
.088		 −.248
.291		 −.014
.952
P&PT .354
.126		 1 .333
.151		 .149
.530		 .061
.797
VisDis .391
.088		 .333
.151		 1 .114
.633		 .180
.447
BVFDT −.248
.291		 .149
.530		 .144
.633		 1 .118
.619
BFRT −.014
.952		 .061
.797		 .180
.447		 .118
.619		 1
Open in a new tab

Significance levels are below each R value on the chart.

No significant correlations at the 0.05 level (2 tailed).

Visual perceptual tasks that likely require access to a structural representation system will account for the most variance in predicting picture naming performance or semantic association performance in AD participants.

Scores on the BVFDT, BFRT, and Visual Discrimination Task from the subjects with Alzheimer’s disease were each entered into a stepwise regression to determine the extent to which they predicted performance on the BNT. The overall model was significant [R=0.61, R2= 0.37, F(1,19)=10.59, p<0.005] in predicting performance on the BNT, however performance on the Visual Discrimination Task [β=0.61, p=0.004] was the only independent predictor that accounted for significant variance (R2= 0.37), whereas the BVFDT [β=−0.01, p=0.95] and the BFRT [β=0.03, p=0.89] did not significantly contribute to the model.

Scores on the BVFDT, BFRT, and Visual Discrimination Task by the subjects with Alzheimer’s disease were entered into a stepwise regression to determine if they predicted performance on the P&PT. The overall model was significant [R=0.78, R2=0.61, F(1,19)=28.12, p<0.001] in predicting performance on the P&PT, however performance on the Visual Discrimination Task [β=0.78, p<0.001] was the only independent predictor that accounted for significant variance (R2=.61), whereas the BVFDT [β=−0.09, p=0.96] and the BFRT [β=0.11, p=0.53] did not significantly contribute to the model.

Individuals with AD will be more impaired than elderly control subjects on discriminating between pictures that have high structural similarity over low structural similarity on the Visual Discrimination Task.

The differences between Visual Discrimination accuracy for high similarity, low similarity, and same pairs of pictures were assessed across the AD and control subjects in a mixed ANOVA to determine if the AD or healthy elderly participants’ performance on the task varied depending on structural similarity. Group membership was the between subject factor and similarity of pictures was the within subjects factor. There was a significant main effect of structural similarity of pictures (low similarity, high similarity and same) [Greenhouse-Geisser Corrected, F(1.38, 52.44) = 12.24, p <0.001] and group (AD or control) [F(1, 38) = 33.43, p <0.001], but the similarity by group interaction [F(1.38,52.44)=2.12, p=0.15] was not statistically significant.

Subsequent independent sample t-tests indicated that there were significant differences between groups for low similarity pictures [t(19.97)= 3.08, p<0.05], high similarity pictures [t(32.41)=2.64, p<0.05], and same pictures [t(24.57)=4.78, p<0.001].

DISCUSSION

Our results indicate significant relationships between visual discrimination abilities of pictured objects with performance on tests of picture naming and visual semantic association. Thus, we find support that the ability to visually discriminate line drawings of nameable objects may play a role in performance on picture tests typically used to assess lexical semantic abilities in persons with AD. These findings are consistent with studies showing that incorrect responses on confrontation naming tasks are likely to be visually similar to the target46 and as less perceptual information is available, AD subjects tend to make more naming errors than control subjects.47 Moreover, the opportunity to use nonverbal sensory information such as touch increases AD subjects’ abilities to name objects.48

In this study, performance on all tests of visual discrimination and lexical-semantic abilities was significantly worse in AD participants than in elderly control participants indicating that the AD subjects were impaired in visual perceptual, semantic and naming abilities. AD participants’ performance on one visual perceptual task, the Visual Discrimination Task, correlated with both the P&PT and the BNT, indicating a relationship with naming and semantic association abilities. Of the three visual perceptual tasks, performance on the Visual Discrimination Task was the only significant predictor of performance on the BNT and the P&PT.

Three visual discrimination tasks that vary in the degree to which they likely access structural and semantic knowledge were utilized in order to provide insight into the level at which impairments in the visual perceptual system impact picture naming. Of the three tasks, we suspect that the BVFDT that requires matching of novel, complex shapes that are not easily named should not depend on access to a structural description system since global and local processing should not evoke a specific stored structural, semantic, or lexical entry. Similarly, processing of novel faces in the BFRT should not evoke structural, semantic or lexical information to help guide the process of discrimination, as it is unlikely that matching a previous mental image to the percept would assist with the task. On the other hand, discrimination of animals, fruits and vegetables on the Visual Discrimination Task likely would have been supported by access to information about the known structure of objects to complete the task, since objects were matched across different views or poses.

We proposed that if low level visual deficits were the main contributor to naming impairment, then the BFRT and BVFDT, which do not likely involve access to a structural description system, should have shown marked impairment that varied with naming abilities. Individuals with AD were significantly impaired on the BFRT and the BVFDT, but this impairment did not correlate with naming impairment, implying that although low-level visual deficits may have existed, they were not the main source of picture naming difficulties.

We also proposed that if higher level impairment in access to stored structural representations were the primary contributor to picture naming difficulties, then visual perceptual tests that do not rely on these systems (BVFDT, BFRT) may be relatively spared, but performance on tests that may require access to this information to complete the task (Visual Discrimination Task) should show impairment that correlates with naming impairment. Results indicated that the BVFDT, BFRT and Visual Discrimination Task were all impaired in AD subjects, but only performance on the Visual Discrimination Task showed a relationship with naming abilities. One possible explanation for this finding is that both low level visual deficits and higher level structural representations of objects were impaired in the AD subjects, but only the deficits in higher level structural representations produced a large enough effect to impact naming.

This explanation is supported by the finding that AD participants performed significantly worse than elderly control subjects on each of the three types of pictures presented in the Visual Discrimination Task. One might argue that the significant difference in performance between groups indicates the contribution of deficits in low level perceptual processing to difficulty in discriminating low similarity pictures, as well as problems with pre-semantic structural descriptions required for discriminating high similarity pictures. It has been hypothesized that low spatial frequency information such as global form may be sufficient to distinguish between line drawings with low similarity ratings, whereas line drawing that are high in similarity may need fine detailed visual discrimination which is processed in more anterior aspects of the ventral visual stream.49 Moreover, discrimination of high similarity line drawings may require access to stored structural representations in order to make a decision. Both low similarity and high similarity pictures were more difficult for AD participants than for healthy controls, possibly indicating concurrent deficits in low level visual processing as well as higher level structural representations.

An alternate explanation for the relationship between performance on visual discrimination and naming is that there are two separate cognitive processes that declined simultaneously. This alternative is plausible given the diffuse neuropathology in AD and the wide range of cognitive decline in our participants as evidenced by MMSE scores, but we believe there may be a causal link for the following reasons. Other tests of visual perceptual abilities (BFRT and BVFDT) did not produce significant correlations with naming. The major distinction was that the Visual Discrimination Task utilized nameable objects, whereas the other visual perceptual tasks did not. The BNT and P&PT also use nameable items, which may elicit a mental image or stored structural representation prior to semantic access. We propose that the integrity of this structural representation system may be the causal link between visual processing deficits and language deficits in the AD participants. Moreover, the finding of a common neural substrate between discrimination and naming of line drawings34 indicates that damage could impact both processes. Specifically fMRI signal found in mid-anterior fusiform areas show a posterior-anterior shift corresponding to low versus high structural similarity, respectively.

Since the Visual Discrimination Task used real, nameable objects, it may have been a strong predictor of naming and semantic association abilities by tapping into a visual process that occurs at a stage prior to naming, such as comparing visual percepts to stored structural representations. Degradation of structural knowledge could account for the deficits in the Visual Discrimination Task, BNT and P&PT, but another deficit, such as low level visual processing would have to have been present to produce the disturbances in the BFRT and the BFVDT. In line with the structural degradation theory, Done and Hajilou16 found that AD participants needed a more visually complete object representation before they could correctly identify pictures of familiar objects, but performance on visually degraded words was similar for AD participants and control subjects. They concluded that in early stages of AD, there is some degradation of structural knowledge thought to be pre-semantic representations of objects within the visual perceptual system in addition to degradation of semantic representations. Furthermore, analysis of AD naming errors for visually intact pictures showed predominantly semantic level deficits that were characterized by substantial categorical or circumlocutory errors, but few pure visual confusion errors. Hence, the degradation of structural knowledge may not result in obvious visual perceptual errors but instead impact the ability to access appropriate semantic representations and therefore cause semantic errors in naming.

Conclusions

Our results indicate that as Done & Hajilou16 suggest, perceptual and semantic processing that occurs with naming may be informed by outputs of the other stages. Low-level visual perceptual errors, such as the inability to perceive visual features, or higher-level errors in mapping perceived features onto a stored structural representation could impact selection of the correct semantic description downstream. This visual perceptual deficit could be manifest as difficulties in picture naming or semantic association impairment. The findings of the present study are consistent with the position of Rogers, Ivanoiu, Patterson, & Hodges50 who postulated that performance on individual semantic tests may reflect the abnormal functioning of other cognitive faculties that provide input to the semantic system. Our findings suggest that there is a common deficit in discrimination of pictures using nameable objects, picture naming and semantic association of pictures in AD. Future studies may investigate whether a treatment approach targeting visual discrimination abilities improves lexical-semantic performance in this population.

Acknowledgments

This work was supported by the Indiana Alzheimer’s Disease Center (P30 AG10133), the National Institutes of Health (R01 MH063817, T32 DC008768, K23 DC010197) and the German Foundation for Science (DFG, ME 3161/2-1). We appreciate the participation of the Alois Alzheimer’s Center, Cincinnati, OH.

References

  • 1.Heeringa SG, Weir DR, Ofstedal MB, et al. Prevalence of dementia in the United States: The aging, demographics, and memory study. Neuroepidemiology. 2007;29:125–132. doi: 10.1159/000109998. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Gonzalez Rothi L, Fuller R, Leon SA, et al. Errorless practice as a possible adjuvant to donepazil in Alzheimer's disease. Journal of the International Neuropsychological Society. 2009;15:311–322. doi: 10.1017/s1355617709090201. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Grandmaison E, Simard M. A critical review of memory stimulation programs in Alzheimer's disease. Journal of Neuropsychiatry & Clinical Neurosciences. 2003;15(2):130–144. doi: 10.1176/jnp.15.2.130. [DOI] [PubMed] [Google Scholar]
  • 4.Giannakopoulos P, Gold G, Duc M, Michel JP, Hof PR, Bouras C. Neuroanatomic correlates of visual agnosia in Alzheimer's disease: A clinicopathologic study. Neurology. 1999;52(1):71–77. doi: 10.1212/wnl.52.1.71. [DOI] [PubMed] [Google Scholar]
  • 5.Bayles KA, Tomoeda CK. Confrontation naming impairment in dementia. Brain and Language. 1983 May;19(1):98–114. doi: 10.1016/0093-934x(83)90057-3. [DOI] [PubMed] [Google Scholar]
  • 6.Ober BA, Shenaut GK. Well-organized conceptual domains in Alzheimer's disease. Journal of the International Neuropsychological Society. 1999 Nov;5(7):676–684. doi: 10.1017/s1355617799577102. [DOI] [PubMed] [Google Scholar]
  • 7.Hodges JR, Patterson K, Graham N, Dawson K. Naming and knowing in dementia of Alzheimer's type. Brain and Language. 1996 Aug;54(2):302–325. doi: 10.1006/brln.1996.0077. [DOI] [PubMed] [Google Scholar]
  • 8.Garrard P, Lambon Ralph MA, Patterson K, Pratt KH, Hodges JR. Semantic feature knowledge and picture naming in dementia of Alzheimer's type: a new approach. Brain and Language. 2005 Apr;93(1):79–94. doi: 10.1016/j.bandl.2004.08.003. [DOI] [PubMed] [Google Scholar]
  • 9.Hof PR, Archin N, Osmand AP, et al. Posterior cortical atrophy in Alzheimer's disease: analysis of a new case and reevaluation of an historical report. Acta Neuropathologica. 1993;86:215–223. doi: 10.1007/BF00304135. [DOI] [PubMed] [Google Scholar]
  • 10.Hof PR, Bouras C, Constantinidis J, Morrison JH. Selective disconnection of specific visual association pathways in cases of Alzheimer's disease presenting with Balint's syndrome. Journal of Neuropathology and Experimental Neurology. 1990;49:168–184. doi: 10.1097/00005072-199003000-00008. [DOI] [PubMed] [Google Scholar]
  • 11.Benson DF, Davis RJ, Snyder BD. Posterior cortical atrophy. Archives of Neurology. 1988;45:789–793. doi: 10.1001/archneur.1988.00520310107024. [DOI] [PubMed] [Google Scholar]
  • 12.Gilmore G. Stimulus encoding in Alzheimer's disease: a multichannel view. In: Allen P, Bashore T, editors. Age differences in word and language processing. Amsterdam, The Netherlands: Elsevier Science; 1995. pp. 199–219. [Google Scholar]
  • 13.Katz B, Rimmer S. Ophthalmologic manifestations of Alzheimer's disease. Survey of Ophthalmology. 1989 Jul-Aug;34(1):31–43. doi: 10.1016/0039-6257(89)90127-6. [DOI] [PubMed] [Google Scholar]
  • 14.Lewis DA, Campbell MJ, Terry RD, Morrison JH. Laminar and regional distributions of neurofibrillary tangles and neuritic plaques in Alzheimer's disease: a quantitative study of visual and auditory cortices. The Journal of Neuroscience. 1987 Jun;7(6):1799–1808. doi: 10.1523/JNEUROSCI.07-06-01799.1987. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Adlington RL, Laws KR, Gale TM. Visual processing in Alzheimer's disease: surface detail and colour fail to aid object identification. Neuropsychologia. 2009 Oct;47(12):2574–2583. doi: 10.1016/j.neuropsychologia.2009.05.004. [DOI] [PubMed] [Google Scholar]
  • 16.Done DJ, Hajilou BB. Loss of high-level perceptual knowledge of object structure in DAT. Neuropsychologia. 2005;43(1):60–68. doi: 10.1016/j.neuropsychologia.2004.06.004. [DOI] [PubMed] [Google Scholar]
  • 17.Hajilou BB, Done DJ. Evidence for a dissociation of structural and semantic knowledge in dementia of the Alzheimer type (DAT) Neuropsychologia. 2007;45:810–816. doi: 10.1016/j.neuropsychologia.2006.08.008. [DOI] [PubMed] [Google Scholar]
  • 18.Humphreys G, Riddoch MJ. Features, objects, actions: The cognitive neuropsychology of visual object processing, 1984–2004. Cognitive Neuropsychology. 2006;23(1):156–183. doi: 10.1080/02643290542000030. [DOI] [PubMed] [Google Scholar]
  • 19.Humphreys G, Riddoch MJ, Quinlan PT. Cascade processes in picture identification. Cognitive Neuropsychology. 1988;5:67–103. [Google Scholar]
  • 20.Humphreys GW, Forde EM. Hierarchies, similarity, and interactivity in object recognition: "category-specific" neuropsychological deficits. Behavioral and Brain Sciences. 2001 Jun;24(3):453–476. discussion 476–509. [PubMed] [Google Scholar]
  • 21.Cronin-Golomb A, Gilmore G, Neargarder SA, Morrison SR, Laudate TM. Enhanced stimulus strength improves visual cognition in aging and Alzheimer's disease. Cortex. 2007;43:952–966. doi: 10.1016/s0010-9452(08)70693-2. [DOI] [PubMed] [Google Scholar]
  • 22.Gilmore G, Cronin-Golomb A, Neargarder SA, Morrison SR. Enhanced stimulus contrast normalizes visual processing of rapidly presented letters in Alzheimer's desease. Vision Research. 2005;45:1013–1020. doi: 10.1016/j.visres.2004.10.017. [DOI] [PubMed] [Google Scholar]
  • 23.Gilmore G, Groth KE, Thomas CW. Stimulus contrast and word reading speed in Alzheimer's disease. Experimental Aging Research. 2005;31:16–33. doi: 10.1080/03610730590882828. [DOI] [PubMed] [Google Scholar]
  • 24.Gonnerman LM, Andersen ES, Devlin JT, Kempler D, Seidenberg MS. Double dissociation of semantic categories in Alzheimer's disease. Brain and Language. 1997 Apr;57(2):254–279. doi: 10.1006/brln.1997.1752. [DOI] [PubMed] [Google Scholar]
  • 25.Farah MJ, McClelland JL. A computational model of semantic memory impairment: Modality specific and emergent category specificity. Journal of Experimental Psychology: General. 1991;120(4):339–357. [PubMed] [Google Scholar]
  • 26.Grossman M, McMillian C, Moore P, et al. What's in a name: voxel-based morphometric analyses of MRI and naming difficulty in Alzheimer's disease, frontotemporal dementia and corticobasal degeneration. Brain. 2004;127(3):628–649. doi: 10.1093/brain/awh075. [DOI] [PubMed] [Google Scholar]
  • 27.McRae K, de Sa VR, Seidenberg MS. On the nature and scope of featural representations of word meaning. Journal of Experimental Psychology: General. 1997 Jun;126(2):99–130. doi: 10.1037//0096-3445.126.2.99. [DOI] [PubMed] [Google Scholar]
  • 28.Grossman M, Koenig P, Troiani V, Work M, Moore P. How necessary are the stripes of a tiger? Diagnostic and characteristic features in an fMRI study of word meaning. Neuropsychologia. 2007;45:1055–1064. doi: 10.1016/j.neuropsychologia.2006.09.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Warrington EK, Shallice T. Category specific semantic impairments. Brain. 1984 Sep;107 ( Pt 3):829–854. doi: 10.1093/brain/107.3.829. [DOI] [PubMed] [Google Scholar]
  • 30.Alathari L, Trinh Ngo C, Dopkins S. Loss of distinctive features and a broader pattern of priming in Alzheimer's disease. Neuropsychology. 2004 Oct;18(4):603–612. doi: 10.1037/0894-4105.18.4.603. [DOI] [PubMed] [Google Scholar]
  • 31.Hedge J. Time course of visual perception: Coarse-to-fine processing and beyond. Progress in neurobiology. 2008;84:405–439. doi: 10.1016/j.pneurobio.2007.09.001. [DOI] [PubMed] [Google Scholar]
  • 32.Bar M. A cortical mechanism for triggering top-down facilitation in visual object recognition. Journal of Cognitive Neuroscience. 2003 May 15;15(4):600–609. doi: 10.1162/089892903321662976. [DOI] [PubMed] [Google Scholar]
  • 33.Tippett LJ, Blackwood K, Farah MJ. Visual object and face processing in mild-to-moderate Alzheimer's disease: from segmentation to imagination. Neuropsychologia. 2003;41(4):453–468. doi: 10.1016/s0028-3932(02)00140-9. [DOI] [PubMed] [Google Scholar]
  • 34.Joseph JE, Gathers AD. Effects of structural similarity on neural substrates for object recognition. Cognitive, Affective, and Behavioral Neuroscience. 2003 Mar;3(1):1–16. doi: 10.3758/cabn.3.1.1. [DOI] [PubMed] [Google Scholar]
  • 35.Joseph JE, Jones KM, Zeffiro TA, Eden GF. FMRI correlates of structural similarity in object naming. Society of Neuroscience Abstracts. 2000;26(2):1502. [Google Scholar]
  • 36.Tyler LK, Stamatakis EA, Bright P, et al. Processing objects at different levels of specificity. Journal of Cognitive Neuroscience. 2004 Apr;16(3):351–362. doi: 10.1162/089892904322926692. [DOI] [PubMed] [Google Scholar]
  • 37.Bussey TJ, Saksida LM. The organization of visual object representations: a connectionist model of effects of lesions in perirhinal cortex. European Journal of Neuroscience. 2002;15:355–364. doi: 10.1046/j.0953-816x.2001.01850.x. [DOI] [PubMed] [Google Scholar]
  • 38.Joseph JE, Steinmetz NA. Resolving shape and color similarity in the human ventral processing stream. Paper presented at: 9th International Conference on Functional Mapping of the Human Brain; New York. 2003. [Google Scholar]
  • 39.Salmon DP, 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. Journal of the International Neuropsychological Society. 1999;5(7):692–703. doi: 10.1017/s1355617799577126. [DOI] [PubMed] [Google Scholar]
  • 40.Folstein MF, Folstein SE, McHugh P. Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. Journal of Psychiatry Research. 1975;12:189–198. doi: 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
  • 41.Howard D, Patterson K. Pyramids and Palm Trees: A test of semantic access from pictures and words. Bury St. Edmunds: Thames Valley Test Company; 1992. [Google Scholar]
  • 42.Kaplan E, Goodglass H, Weintraub S. Boston Naming Test. 2. Philadelphia: Lea and Febiger; 1983. [Google Scholar]
  • 43.Benton AL, Hamsher KD, Varney NR, Spreen O. Contributions to neuropsychological assessment. New York: Oxford University Press; 1983. [Google Scholar]
  • 44.Joseph JE. Color processing in object verification. Acta Psychologica. 1997;97:95–127. doi: 10.1016/s0001-6918(97)00026-7. [DOI] [PubMed] [Google Scholar]
  • 45.Joseph JE, Profitt DR. Semantic versus perceptual influences of color in object recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition. 1996;22(2):407–429. doi: 10.1037//0278-7393.22.2.407. [DOI] [PubMed] [Google Scholar]
  • 46.Rochford G. A study of naming errors in dysphasic and in demented patients. Neuropsychologia. 1971 Dec;9(4):437–443. doi: 10.1016/0028-3932(71)90008-x. [DOI] [PubMed] [Google Scholar]
  • 47.Kirshner HS, Webb WG, Kelly MP. The naming disorder of dementia. Neuropsychologia. 1984;22(1):23–30. doi: 10.1016/0028-3932(84)90004-6. [DOI] [PubMed] [Google Scholar]
  • 48.Barker MG, Lawson JS. Nominal aphasia in dementia. The British Journal of Psychiatry. 1968 Nov;114(516):1351–1356. doi: 10.1192/bjp.114.516.1351. [DOI] [PubMed] [Google Scholar]
  • 49.Joseph JE, Farley AB. Cortical regions associated with different aspects of object recognition performance. Cognitive, Affective, and Behavioral Neuroscience. 2004 Sep;4(3):364–378. doi: 10.3758/cabn.4.3.364. [DOI] [PubMed] [Google Scholar]
  • 50.Rogers TT, Ivanoiu A, Patterson K, Hodges JR. Semantic memory in Alzheimer's disease and the frontotemporal dementias. Neuropsychology. 2006;20(3):319–335. doi: 10.1037/0894-4105.20.3.319. [DOI] [PubMed] [Google Scholar]

RESOURCES