2024 年 23 巻 2 号 p. 33-44
Texture is one of the key perceptions that enable to comprehend the nature of multimodal signals from the environment according to their type and condition and convert them into a logical output in the central nervous system. Neuroimaging and some clinical studies have proven the importance of the ventral visual pathway in visual texture recognition. Although Alzheimer’s disease and dementia with Lewy bodies are believed to cause visual impairments, texture visualization has not been sufficiently investigated. In this paper, the author provides an overview of human visual texture sensation and focuses on texture recognition disorders in patients with cerebrovascular disease and dementia.
Texture perceptualizes the characteristics of surrounding objects and is therefore essential for daily life. For instance, it provides important information for well-being, such as identifying objects, checking if food is spoiled, or determining if mechanical components have degraded. Generally, we can perceive the material and condition of objects based on sensory stimuli inputs, such as visual, auditory, tactile, and olfactory cues. For example, we can visually determine whether an object is made of metal or glass. We can also identify whether vegetables are fresh or slightly old or whether the ground is dry or wet. This ability of humans to assess the structure of objects based on sensory information is referred to as texture (Shitsukan) perception [1]. Texture perception is considered to be a fundamental function of sensory organs achieved through sensory inputs such as vision, hearing, and touch.
Texture recognition encompasses two aspects: material and state recognitions. Material recognition involves transforming visual sensory inputs, such as determining whether the surface of an object is smooth or not, into information regarding the material and properties of the object. Based on this information, we can firmly grasp objects and prevent them from slipping. In other words, our actions are determined by the information provided through material recognition. Carefully walking on a wet road to avoid slipping is an example of utilizing information from material recognition. On the other hand, state recognition gives a clue to the condition of an object. For instance, we are guided by state recognition to avoid spoiled fruits and to select fresh vegetables and fish when shopping.
Texture recognition is a complex cognitive function acquired through learning and is directly associated with movement and emotion. Research on human texture recognition is still restricted, and there are many aspects that are not yet fully understood. However, in recent years, research on this sensation as a part of visual object recognition has been progressing. In this paper, we provide an overview of human visual texture perception and represent our research, the primary subject of which is texture recognition disorders in patients with cerebrovascular disease and dementia.
In our daily lives, we can perceive the structure of various materials, such as metal, glass, wood, stone, leather, and fabric, by just looking at photographs or the objects themselves. Despite the wide range of available materials, we can easily identify what an object is made of. For example, we can make subtle judgments, such as determining whether the flooring in a model room is real solid wood or a vinyl sheet designed to resemble wood or discerning food samples displayed in front of a restaurant, which are incredibly realistic but are actually made of wax or synthetic resin. The question of how we are able to do this is one of the fundamental challenges addressed in research on visual texture recognition.
In addition, one interesting observation related to this is that most of the materials are natural, such as “metal,” “plastic,” “wood,” and “fabric.” Figure 1 presents some examples of material categories. When a verbal name is unavailable, we can connect psychological characteristic with the pattern of a material while categorizing it with the formerly seen sample. In addition, we can judge the material’s properties (e.g., glossy, rough, hard, softly, cold) and characterize them verbally by relevant adjectives and “onomatopoeia” (e.g., “つるつる”, “ざらざら”, “ふわふわ”). These processes can be considered as one aspect of human knowledge (semantic memory) regarding the visual properties of objects.
Visual information processing is thought to be carried out by two pathways: dorsal visual pathway, which revises information about the location and movement of objects, and ventral visual pathway, which translates the data about color and shape [2-3] (Figure 2). In addition to the latter, visual information of objects includes attributes such as color, depth, and texture. Recent neuroimaging studies suggested that the ventral visual pathway is critical for addressing visual attributes, such as texture, color, and shape, and that each specific cortical area is involved in this processing [4-5]. Studies demonstrated that the lateral occipital cortex is responsible for tactile perception, the lingual gyrus and lateral occipital sulcus for color perception, and the lateral occipital region for shape [6-7]. Texture sensation can be stimulated by the various senses, such as vision, touch, and hearing, and brain damage can impair visual understanding of texture.
Clinical studies have reported disorders of visual texture perception in the cases of localized brain damage caused by stroke, head trauma, encephalitis, etc. [8-13].
Suzuki [13] gives the term “visual texture agnosia” to visual texture perception impairment and correlates it with visual agnosia, where the shape of an object can be recognized through tactile sensation but not vision. Patients with this disorder had problems identifying texture and correlating it with materials shown in photographs or doing the reverse tasks. Although this category of patients had normal visual acuity, visual field, and contrast sensitivity, some of them experienced stereopsis deficit, dysfunction of motion-based form recognition [8], or cerebral achromatopsia [6]. On the other hand, Cavina-Pratesi et al. [11] demonstrated a dissociation between texture/color perception and form perception by comparing cases with impaired texture/color recognition and cases with visual form agnosia.
Lesions promoting visual texture agnosia are located in the ventral visual stream, which is composed of the right collateral sulcus (CoS), lingual gyrus, and posterior fusiform gyrus; the left occipitotemporal subcortical areas close to the inferior temporal sulcus and the CoS [10]; the bilateral CoS, fusiform, and parahippocampal gyri, temporal pole (also the right primary visual cortex), and inferior/middle temporal gyri [6, 11]; and left fusiform gyrus and right lateral temporal area [13]. There have been cases of severe traumatic multiple hemorrhages, where chronic-phase brain wave analysis revealed a functional decline in the occipitoparietal regions bilaterally [8].
In the following paragraphs, we report a case of visual texture agnosia caused by cerebrovascular disease, where recognition of forms and texton is preserved, whereas that of visual textures is primarily impaired.
An 80-year-old right-handed man who was a company executive was admitted to our hospital with symptoms of weakness in the right limbs and mild dysarthric speech. His medical history included hypertension, atrial fibrillation, and cerebral embolism in the area of the left posterior cerebral artery. On admission, he was conscious, and oriented in time and place. Neurological symptoms included right upper quadrantanopia, and his visual acuity was normal from the two sides. His general cognitive abilities and language functions were maintained, except for writing kanji. During hospitalization that lasted about 1 month, he received speech therapy for dysgraphia. Magnetic resonance imaging (MRI) was performed using axial multi-echo with coronal reconstructions. The FLAIR-weighted sequence showed the inclusion of the left thalamus, parahippocampal gyrus, a portion of the fusiform and lingual gyri into the lesions, and part of the ventral portion of the temporal pole (Figure 3).
A meticulous assessment of neuropsychological status was conducted 12 months after the disease onset. All the inspections were performed after obtaining informed consent according to the Declaration of Helsinki. The patient had a score of 107 for verbal IQ and 75 for performance IQ on the Wechsler Adult Intelligence Scale 3rd Edition [14]. He particularly experienced difficulty in searching targets, which affected his performance IQ. The Japanese version of Western Aphasia Battery [15] and the Standard Language Test of Aphasia (SLTA) [16], a standardized language test for Japanese, were also used for evaluation. The patient had no other disturbances in oral language abilities, and his aphasia quotient was 94.5 (normal > 93.8). No aphasia was observed in the SLTA, and he was able to correctly name all the indicated objects and to maintain auditory and reading comprehension for complex grammar and long sentences. There were no conditions such as visual agnosia or image agnosia. However, he exhibited agraphia (difficulty in writing kanji) and mild reading impairment (dyslexia) specifically with regard to kana characters.
The subject was able to correctly identify all the differences in the difficult-to-verbalize fractal shapes, and he could distinguish subtle differences in morphology. Figure 4 presents an example of the stimuli used in the black and white fractal task. The face perception abilities of the subject were intact, as demonstrated by the ability to match and discriminate between different faces, judge sex and age, and recognize familiar faces. However, the ability to name famous faces was impaired. Correct classification of hues on the 100-Hue test proved that the color perception was intact. However, the subject had difficulty in color naming and identifying colors from their names, as well as in associating colors with objects or their names. We have already reported in detail about the color anomia and impaired color semantic memory in this case (Oishi et al., 2021) [17].
1) Visual texture perception by textons
An examination was conducted to discriminate the area differences formed by those of textons. The tasks included judging the size of the region based on the angles, colors, and thickness of textons, and the size of regions distinguished by color. As a result, all tasks had normal accuracy and reaction times, and the extraction of the region by texture that popped out was preserved.
2) Material judgment
The study used six uniformly shaped materials (metal, ceramic, glass, wood, cloth, and leather) to evaluate visual differentiation and material identification. Rod-shaped, real object stimuli were transmitted for the real material identification task. The surfaces of 18 objects (200 mm ×ϕ25 mm) were made from six material categories (bark, fabric, leather, ceramic, glass, and metal) (three examples per category) [18]. The results indicated that the patient was able to distinguish different materials; however, he had problems with material identification except for certain geometric patterns on cloth or the bark of a tree (Figure 5A). In addition, the study used 10 images of each material (metal, ceramic, glass, wood, cloth, and leather) to conduct similar evaluation. During the study, the patient struggled to identify materials and had more difficulty in identifying them when presented with image stimuli compared to when real material stimuli were used (Figure 5B). Furthermore, the patient had difficulty describing surface characteristics such as smoothness and roughness.
3) Matching of material with object
Matching of material with object tests was conducted in two formats: actual visual images of materials and text-based material names. First, a patient was shown a line drawing of an object and images of four different materials and was asked to select the material’s image that matched the presented object. For example, the appropriate choice would be metal if a line drawing of fork was presented (Figure 6). The study used 24-line drawings of objects (e.g., key, frying pan, tweezers, tea cup, and tie) and images of eight materials (metal, ceramic, glass, stone, fur, bake, leather, and fabric). Second, the patient was instructed to point to the name of the material that matched the presented object drawing from a list that had the names of the eight materials. A line drawing of an object was shown, and the subject was asked to provide a relevant material name for it.
The patient correctly identified the names of all the objects but failed to choose proper visual images of materials for one-third of the items (15/24). He made 7 incorrect responses over 24 trials in the task of matching material’s names with objects. Overall, his performance on tasks on the association of material and object was unsatisfactory.
4) Visual judgments of perceptual qualities
Based on the previous study [19], the author assessed visual judgments of perceptual qualities in the case. The patient rated the materials for the latter (e.g., smoothness, roughness, hardness) when viewing photos of various materials. The patient was asked to evaluate the perceived quality of the presented material images and enter their responses into a table on a scale of one to five. Before starting the tasks, the perceptual quality of the materials was defined, and the characteristics of the five-point scale (i.e., correspondence of low and high values) was specified. The following six qualities were assessed with the following definitions:
•Smoothness: How smooth is this material? Low values indicate a nonsmooth state, whereas high values indicate a very smooth state.
•Roughness: If you were to touch the material, how rough would it feel? Low values indicate that the surface does not feel rough, whereas a high value indicates that it feels very rough.
•Softness: If you were to touch the material, how soft would it feel? Low values indicate that the surface is not very soft; high values indicate that it would feel very soft.
•Hardness: If you were to touch the material, how hard would it feel? Low values indicate that the surface is not perceived as very hard, whereas high values demonstrate a very firm texture.
•Coldness: How cold do you expect the surface to feel when touched? Low values indicate that the material is not very cold, whilea high value indicates that it feels very cold.
•Warmness: How warm do you expect the surface to feel when you touch it? Low value indicates that the material does not feel very warm, whereas a high value indicates that it feels very warm.
The results of the experiment are presented in Figure 7, where some obvious errors are displayed. For example, in the evaluation of roughness, the patient gave a high rating to “glass,” which should normally have a low value. In addition, when asked to evaluate softness, he gave a high rating for “ceramic,” which typically has a low value.
To summarize, this case had impaired visual recognition of texture. Furthermore, there were disorders in the visual and linguistic associations between visual texture and objects, as well as in the judgments of visual perceptual qualities. This suggests that many aspects of his ability to revise semantic information regarding visual texture are affected. This case illustrates semantic memory impairment related to object material. Contrarily, discrimination of form differences, and recognition of object shape were well preserved. The lesion involved the medial occipital lobe, including the fusiform and lingual gyri, on the left side.
Cavina-Pratesi et al. [6] reported a case of preserved shape recognition but impaired texture and color perception, which is similar to the symptoms in the present case. Furthermore, their study used functional magnetic resonance imaging (fMRI) and suggested that bilateral parietal sulcus is crucial for texture and color recognition. Given that the lesion in the present case, and also in the previous [6], involves the left medial part of the occipital lobe, including left CoS, it is suggested that the left intraparietal sulcus is more essential for visual texture recognition.
The author introduces several previous research reports on visual texture perception in patients with degenerative cognitive disorders. Many studies on texture perception in healthy individuals have been conducted; however, studies focusing on texture recognition on individuals with dementia are scarce.
In the previous studies on visual texture perception in Alzheimer’s disease (AD) [20], texture discrimination tasks using textons and luminance were assigned to patients with AD. The texture discrimination task utilizes textures represented by textons, characterized by features such as line gradients, density, and intersections, based on Julesz’s texton theory [21-22]. Different textons naturally attract visual attention, and in these texture discrimination tasks, no significant decline was observed compared with the control group. However, in the tasks requiring the identification of forms expressed by the differences in texture within complex backgrounds, a decline was detected, which was attributed to the impairments in the search system.
A study conducted by Oishi et al. [18] investigated the visual texture recognition abilities of patients with AD and dementia with Lewy bodies (DLB). The subjects were 25 patients with probable DLB, 53 with probable AD, and 32 healthy controls. Visual texture recognition of real materials/images and visuoperceptual functions, including contrast sensitivity, color perception, stereopsis, shape detection, and position in space, were investigated. As a result, visual texture recognition function decreased in both patients with DLB and those with AD compared with the healthy older adults, with the DLB group showing greater impairment. In DLB, a decline in basic visual perception function was observed, and material identification was severely impaired. However, it was demonstrated that visual texture recognition impairment occurred independently of the impairment in basic visual perception function. In the analysis stratified by severity, patients with DLB exhibited impairment in both real and image material identifications from the mild cognitive impairment (MCI) stage. In patients with AD, disturbed image material identification was detected from the MCI stage, whereas impairment in real material identification occurred at mild to moderate stages.
The author represents research on the association between impairments in object recognition and texture perception. In our surroundings, various objects exist under different environmental conditions, and the shape of an observed object changes as the observer’s viewpoint changes. In addition, under varying lighting conditions, the object’s shading changes, and the contours become less distinct, often misinterpreting information about form. In the object’s perception under these circumstances, it is believed that cues, such as color and texture, are utilized, and the recognition of objects from different perspectives is necessary by comparing them with semantic memory.
Oishi et al. [23] demonstrated the association between visual texture recognition impairment and object recognition in patients with degenerative cognitive disorders. This study aimed to determine how viewpoints, textures, and visual cognitive functions impact object recognition and cause visual misidentifications in patients with DLB or AD. A total of 37 patients with probable DLB, 58 with probable AD, and 32 healthy controls underwent neuropsychological status evaluation and performed object identification tasks using four conditions (noncanonical view + Gaussian blurring, noncanonical view + clear texture, canonical view + Gaussian blurring, canonical view + clear texture). An example of stimuli used in the object identification test is presented in Figure 8. The association between object identification and other visuoperceptual functions was investigated. As a result, patients with DLB, and AD showed greater impairment in object recognition under noncanonical viewing with blurry texture conditions, with the DLB patients demonstrating significantly worse results than the AD patients. Patients with DLB and AD had visual misidentifications during object identification tasks under noncanonical viewing. In patients with DLB, the number of misinterpretations was significantly correlated with the scores of visual texture recognition. The present study displayed significantly impaired object recognition in patients with DLB under the influence of both viewpoint and visual texture and in those with AD under the influence of viewpoint. Visual discrepancy in object recognition could be connected with impaired visual texture recognition in DLB.
Furthermore, to investigate the texture perception of recognition ability, Oishi et al. [24] conducted a study using vegetable freshness perception. In dementia, various symptoms appear in daily life in addition to memory loss. Not only do individuals forget what they have bought and end up purchasing the same food again, they also struggle judging the condition of food, resulting in cooking using spoiled ingredients or refusing to discard them. As the condition progresses, there may be instances of pica, when individuals mistakenly ingest nonfood items such as erasers or soap. These symptoms are believed to be caused by various factors, and one potential factor is visual perceptual impairment, which impairs the recognition of the object’s shapes and textures.
This study aimed to assess vegetable freshness perception in patients with AD and DLB and to detect the connection between vegetable freshness perception and various visuoperceptual functions. We enrolled 37 patients with probable DLB, 58 with probable AD, and 32 healthy controls. The parameters included vegetable brightness perception, contrast sensitivity, color perception, and stereopsis. The freshness of vegetables can be evaluated by processing vegetable freshness perception. We, therefore, used the photographs of vegetables (carrot and Japanese mustard spinach [komatsuna]) previously degraded by heat in a thermostatic chamber from 0 h to 66 h as stimuli [25]. We used this method because it is difficult to obtain vegetables with the same level of freshness; our method also prevented the subjects from using other senses, such as smell, when assessing the freshness of the samples. We created two stimulation pairs at different levels of freshness for each vegetable. As a control task for the vegetable freshness determination test, we performed a luminance determination test. Using a photographic stimulation of komatsuna, we created 20 pairs of photographs with differing luminance levels and asked the subjects to select the brighter photograph out of each pair (20 trials). Patients with DLB displayed disproportionate deficits in vegetable freshness and vegetable luminance perceptions compared with those with AD and controls. The groups with two different types of freshness perceptions had significant differences in contrast sensitivity and visual texture recognition, where significantly impaired vegetable freshness perception was detected in patients with DLB. Vegetable freshness perception may be linked to visual texture recognition in patients with DLB.
The above summary provides an overview of research on visual texture perception deficits in degenerative cognitive disorders. Unlike the previous studies, this study used real stimuli in the texture perception tasks, allowing us to elucidate the texture recognition impairments that are more closely related to daily life experiences of individuals with cognitive disorders. It is believed that complex texture recognition is impaired at the early stages of both DLB and AD. Visual misperceptions commonly observed in DLB patients, such as seeing clothes, may also be influenced by perspective and texture. Considering the potential impact of texture recognition deficits on the occurrence of visual misperceptions in individuals with cognitive disorders, it is important to consider environmental modifications and provide family guidance from the early stages of dementia.
Texture is one of the important elements in human recognition of objects. As discussed in this paper, although knowledge about visual texture perception is gradually increasing from the studies using local brain lesion cases or brain imaging methods, clinical research on visual texture recognition in patients with cerebral vascular disease and degenerative dementia is limited. Accumulation of further data in future research trials is expected.
The author would like to thank Hidehiko Komatsu and Katsunori Okajima for providing image data and Kyoko Suzuki for her helpful comments and support.
This work was supported by a Grant-in-Aid for Scientific Research (C) [grant number 21K11229] from MEXT to YO, Japan.
The authors have no conflicting interests to declare.
