Clinical manifestations

Presentation and course

Diagnosing visual agnosia requires diligence and time, as well as the use of directed questions and specialized testing to discover that a patient’s visual struggles are due to impaired recognition. Many patients will not have insight into the fact that their visual problems are due to brain dysfunction and assume that their perceptual issues stem from an ocular disorder. Those with developmental forms of agnosia may have had the deficit forever and seek the attention of a caregiver only when they realize that other people see things that they are not seeing (84).

General visual agnosias

Patients with general visual agnosias often report being unable to focus on viewed objects. They mishandle most daily tasks, being unable to recognize coins or paper currency and not knowing which buttons to press on their telephones. They typically attribute their deficit to their eyes and repeatedly consult providers for replacement spectacles. Many will be labeled as having psychogenic disorders.

Selective visual agnosias

Prosopagnosia. Patients with the apperceptive variant of prosopagnosia may complain that all faces look similar, like pebbles on a beach. When presented with a sequence of faces spanning various ages, they will have difficulty discerning the youngest and the oldest faces. Patients with the associative variant can discern facial features that enable the detection of age and differences between faces, but if shown a series of famous or familiar faces, they will be unable to recognize them. In either form of prosopagnosia, patients will base recognition on other visual cues, such as hairstyle, distinctive facial features, gait, mannerisms, and voice.

Pure alexia. Patients with pure alexia report that they read slowly or that reading is so laborious or error-prone that they have given it up. They may even report that they cannot read their own writing. Print size and distance to printed material do not influence the reading deficit.

Topographagnosia. Patients with agnosia for locales will get lost in familiar surroundings, including their neighborhood or even their own home. Some patients can compensate by relying on devices with a global positioning system.

Prognosis and complications

The prognosis of agnosia depends on its cause. Patients with small strokes may partially recover. Those with carbon monoxide poisoning, herpes encephalitis, or post-temporal lobectomy often have persistent deficits. Unrelenting slow deterioration is the rule with neurodegenerative diseases that lead to dementia, including those with the posterior form (visual variant) of Alzheimer disease.

Developmental agnosias are stable throughout life, although some have fluctuating psychosocial consequences. For instance, those with developmental prosopagnosia might suffer with the impression that they are aloof and unsociable (31) when starting a new job or interacting with a new group of acquaintances.

Clinical vignette

Pure alexia. A 41-year-old mechanical engineer who owned an automotive repair shop awoke with numbness in the right face and leg 1 week after an airplane trip. MRI showed a left medial occipital and fusiform gyrus stroke, attributed to paradoxical emboli through a patent foramen ovale. His memory of events was poor initially but quickly returned to normal. He had experienced trouble naming objects, but this quickly normalized, as did the numbness. He reported persistent symptoms of trouble seeing to the right and problems reading. On examination, he had a right homonymous hemianopia. Reading was very slow, proceeding one letter at a time. He could identify single letters and write well. His partner had researched rehabilitative reading strategies, and he had tried using scrolling text and manually tracing letters. He considered that improvement had been modest.

He knew which faces were familiar to him but initially had trouble recalling their names or identities. Formal neuropsychological testing showed difficulties with verbal memory but normal performance on the visual object and space perception battery. On the Warrington Recognition Memory Test, he scored 37/50 for words and 44/50 for faces, and he had normal car recognition. Measures of his reading of single words showed a markedly elevated word-length effect of 1.5 seconds per letter. He was diagnosed with alexia without agraphia (pure alexia) and right homonymous hemianopia.

Experimental studies showed that his problem with visual processing of language impacted written words and the ability to lip-read from faces even though his ability to recognize the identity of faces was intact (94; 03).

Biological basis

The processes associated with visual perception and recognition involve cerebral networks with simultaneous parallel and sequential processing. After visual information reaches the striate cortex in the occipital lobe, information proceeds to extrastriate areas V2 and V3. Thereafter, visual processing is distributed to various cortical regions by means of two main streams (106).

The first pathway is the dorsal (occipitoparietal, “where”) stream, which is dedicated to spatial processing for action and temporal aspects of vision (78). The second pathway is the ventral (occipitotemporal, “what”) stream, which is dedicated to visual recognition. Both pathways have an intermediate level of processing. For the dorsal stream, intermediate areas include V5 and V5a (or V5/MT complex), which are involved in motion processing (114). For the ventral stream, intermediate areas include V4 and V4a, which are involved in color analysis (115).

Identification of any object, visual attribute, or entity involves a network of regions (16) that can be demonstrated using functional magnetic resonance imaging and lesion analysis. As information moves from the retina to cortical networks, more complex representations of the visual information are extracted and associated with nonvisual sensory perceptions, memories, emotions, and behaviors, including limb and body motor responses.

Anatomic localization

General visual agnosias.

Form agnosia. Form agnosia is associated with widespread bilateral occipital lesions, particularly damage to medial structures, such as the lingual gyri (66). Deficits are especially likely to occur with widespread but nonfocal damage, as occurs after carbon monoxide poisoning, mercury poisoning, or anoxia (02; 17; 71; 23; 79).

Integrative agnosia. Integrative agnosia occurs following bilateral lesions of the lateral and inferior occipital cortex (67), most commonly from peri-striate infarcts or in the visual variant of Alzheimer disease (62; 56). Damage to the right cerebral hemisphere alone can also cause a milder version with prolonged reaction times for naming objects in line drawings (51).

Transformation agnosia. Knowledge of what objects look like from different viewpoints may be stored in both hemispheres, perhaps with a greater “part-based” representation of features in the left hemisphere and more of a “whole-object” representation in the right hemisphere (45; 68).

Category-specific agnosia. Lesions that cause category-specific associative visual agnosia may be related to anterior occipital and posterior temporal damage (88), possibly bilaterally. This type of agnosia may occur when left-sided lesions disconnect posterior occipital areas from more anterior temporal areas (25) or involve the left parahippocampal, fusiform, and lingual gyri (48).

The selective visual agnosias.

Prosopagnosia. Right occipitotemporal or bilateral occipitotemporal lesions can result in prosopagnosia. There are rare cases with left-sided lesions, but in these instances, some subjects were left-handed and may have had atypical hemispheric dominance (105; 76; 06), and others had subtle right-sided anomalies and perhaps should have been considered as having bilateral damage (113).

A network of face areas has been demonstrated by functional MRI (59; 54), including the occipital face area, the fusiform face area, and the superior temporal sulcus, and there is evidence for an anterior inferotemporal area (69; 104). Beyond this core network, an extended network, including the precuneus and inferior frontal gyrus, is sometimes activated by faces (59).

Lesions that cause the apperceptive variant of prosopagnosia are in the inferior occipitotemporal cortex, with loss of the occipital face area, fusiform face area, or both, on the right or both sides (12; 92; 35). Lesions that cause the amnestic variant localize to the right or bilateral anterior temporal cortex. Some patients have both occipitotemporal and anterior temporal damage. In contrast, a lesion of the superior temporal sulcus may impair facial expression perception but leave face identification relatively intact (49).

There is not yet a consensus about the structural basis of developmental prosopagnosia. These subjects do not have gross structural lesions on clinical scans. Some investigators have found decreased fusiform face area activation (57), but others have not (58; 04). Some investigators have reported decreased cortical thickness of the fusiform gyrus (50). Others have argued more for reduced connectivity or white matter damage either in the vicinity of the fusiform gyrus or in the inferior longitudinal fasciculus, which may carry connections between the occipitotemporal face areas and those in the anterior temporal cortex (55; 96; 97; 74; 116; 117).

Pure alexia. Pure alexia has been explained by two hypotheses. The first hypothesis is that it is a disconnection of visual information from the dominant language hemisphere (39; 52). Damage to the left occipital cortex and to callosal fibers in the splenium would cause a homonymous hemianopia and disconnect the left language regions from visual information coming from the intact right occipital lobe. The second hypothesis is that cortical damage to the left fusiform gyrus causes pure alexia (111; 19). In that hypothesis, there is no need to invoke a callosal disconnection, and the left cerebral hemispheric lesion could spare the visual pathway. Indeed, cases of pure alexia without homonymous hemianopia have been reported.

Topographagnosia. In right-handed people, the bilateral parahippocampal gyri and parieto-occipital junctions are involved in processing scene perception. In functional MRI studies, the region activated during scene perception is known as the parahippocampal place area (43). Dysfunction of this region likely explains the association of right occipitotemporal damage with topographagnosia (26).

Epidemiology

Although the true incidence of visual agnosias is not well documented, they are known to be rare. Their presumed rarity may reflect the necessity for bilateral lesions in most cases, coupled with the fact that lesions must preserve the retinocortical visual pathway. In some cases, these disorders are not identified by patients or caregivers.

Pure alexia is likely the most often identified type of agnosia because it requires only a unilateral left occipitotemporal lesion. Prosopagnosia is another type of agnosia that can occur after a unilateral lesion, but such a lesion typically causes only a temporary deficit (36; 35).

Developmental agnosias are believed to be more common than acquired disorders. Developmental prosopagnosia reportedly has a prevalence of 2% to 3% (22), an estimate that relies on face recognition test scores falling two standard deviations below a group mean (11).

Differential diagnosis

Confusing conditions

Vision loss. Vision loss due to ocular or retro-ocular visual pathway disorders is an important cause of an inability to recognize by sight. Although homonymous visual field loss can accompany several selective visual agnosias, the recognition deficit should not be attributed to the field loss. A macular-splitting homonymous hemianopia can reduce reading speed if it involves the central 5 degrees, but such a deficit should be called “hemianopic dyslexia” rather than pure alexia (118; 103). Hemianopic dyslexia has a modest word-length effect of, at most, 120 ms per letter. By contrast, the word-length effect may reach 1 second or more per letter in pure alexia (94). As a precaution, however, patients with impaired reading should not be diagnosed with a recognition disorder until perimetry with special attention to the central 10 degrees of the visual field has excluded small homonymous paracentral scotomas typically caused by infarction of the occipital pole.

Anomia. Naming deficits due to language dysfunction can be mistaken for visual agnosias. Patients with anomia should be able to describe what they see and how to use the object and show its use with gestures. But they cannot name it.

Associated or underlying disorders

The classic “ventral visual syndrome” is the association of prosopagnosia with a right or bilateral superior homonymous quadrantanopia (77; 32), central dyschromatopsia (the inability to distinguish hues) (81), and topographagnosia (28). This combination of signs is caused by right or bilateral inferior occipitotemporal lesions. When the lesions are bilateral, pure alexia may be present as well.

Prosopagnosia can also be associated with phonagnosia, which is an impaired recognition of familiar voices, and amusia, also known as tone deafness (73; 30).

In developmental prosopagnosia, the patient may fall within the autism spectrum, and face processing can be impaired (107; 10; 65; 101).

Diagnostic workup

Basic visual examination. The examination begins by testing visual acuity and visual fields to ensure that the patient has adequate basic visual functions to allow for visual recognition. If there is impaired recognition of letters or shapes, patients might be unable to identify symbols on the Snellen visual acuity chart. The examiner must then rely on testing visual acuity with the letter E displayed in various orientations (“tumbling E”). The patient indicates the E’s orientation by positioning the fingers of one hand.

The visual field must also be tested formally to rule out loss of the central 5 to 15 degrees, which could impair the identification of objects, faces, or letters.

Cognitive examination. The Mini-Mental State Examination and Montreal Cognitive Assessment aid in screening for cognitive deficits. However, one often needs a formal neuropsychological assessment to identify language deficits or loss of semantic knowledge that does not rely on vision. A thorough evaluation of memory functions can also be completed in a formal neuropsychological assessment, which is particularly important when assessing whether associative visual agnosia is present.

Assessment of general visual agnosias. A diagnostic evaluation for the presence of general visual agnosia can begin by showing patients various items in the room and asking them to name the items, describe what they are, or mime how they are used. Identifying objects in line drawings may be more sensitive because they provide fewer visual cues than real objects. Difficulties on these tests should be compared to the ability to identify the same objects by hearing or touch. If the identification is impaired using nonvisual presentations, the patient has a multimodal problem rather than isolated visual agnosia.

To evaluate for the presence of form agnosia, a patient should be asked to identify simple shapes like squares, triangles, and circles. If they cannot identify these items, a form of agnosia may be present (42; 80). To differentiate a loss of semantic knowledge that does not depend on vision, a patient can be asked about the differences between shapes (“what is the difference between a square and a circle?”). Some patients with form agnosia can identify objects in drawings if they can trace their outlines, translating a visual percept into a kinesthetic one (01; 71).

If patients can identify simple shapes, one can use an overlapping figures test, a robust method of detecting integrative agnosia (62; 56). Patients with this type of visual agnosia also have trouble recognizing the impossibility of images like Escher drawings. Doing so requires an appreciation that the local elements do not integrate into a correct three-dimensional structure (40).

To test for knowledge of object structure, patients may be asked to describe what specific objects look like or to draw them from memory. Their visual access to semantic stores can be probed by showing them several objects and asking them which ones are related and which are not. For example, patients can be asked, “Which is the odd one out?” while viewing pictures of a hammer, wrench, and eggbeater. To ensure that the problem is specific to vision, this performance should be compared to the ability to do the same task with the names of objects rather than pictures.

Other neuropsychological batteries can supplement these bedside tests to distinguish between associative and apperceptive visual agnosias. The Visual Object and Space Perception Battery (VOSP) assesses various elements of object recognition and visuospatial perception (110). In contrast, the Birmingham Object Recognition Battery (BORB) assesses low-level, intermediate, and high-level processing in object recognition (91). Other useful tests include the Cortical Vision Screening Test (CORVIST) (87; 86) and the Leuven Perceptual Organization Screening Test (L-POST) (102).

Assessment of selective visual agnosias. Most of the visual processing necessary for object perception is relatively intact in patients with selective visual agnosias. One or more of the above tests should be completed before assessing for the selective agnosias.

Prosopagnosia. Assessment in the clinic can begin by using photographic albums available on smartphones to assess for familiarity. If the telephone of a family member or spouse is used, the correctness of the patient’s responses can be confirmed.

The Famous Faces test is a useful tool, bearing in mind the patient’s age and culture. Alternatively, short-term familiarity with faces can be assessed by showing several faces in a learning phase, presenting those faces along with others in a recall phase, and asking the patient to indicate which face in the larger set is recalled from the initial set. Formal tests include the Cambridge Face Memory Test (41) and the Warrington Recognition Memory Test (108). With the latter, poor performance on the face half of the test can be contrasted with better performance on the word half. The diagnosis of prosopagnosia is based on a lack of facial recognition but intact voice recognition.

Pure alexia. Pure alexia can be assessed by having the patient read single words and paragraphs in a magazine or book and displaying slowed reading of longer-length words. (13). If patients cannot read words, they can be assessed for reading of numbers and single letters to determine the severity of the deficit (98).

Patients with pure alexia can write much better than they can read. However, some patients have “surface dysgraphia,” an impaired ability to spell and write words with irregular spelling (yacht, colonel). Such a finding denotes a deficit in accessing an internal lexicon (95). Patients with both reading and writing impairment have “alexia with agraphia,” a language domain deficit rather than a visual perceptual deficit. In that circumstance, speech and auditory comprehension should be tested to ensure that poor reading is not simply part of an aphasia. Even without aphasia, reading and writing deficits may be part of a dementia.

Topographagnosia. An initial assessment includes the patient’s ability to recognize familiar landmarks, scenes, or buildings. A collection of pictures from a local city or from famous landmarks (the Eiffel Tower, Niagara Falls) can be used. The photographic library on a family member’s smartphone may serve.

Management

Occupational therapists can help patients learn strategies to compensate in part for their deficits. However, there are very few studies to guide the approach. Mobile devices with recognition software have been steadily improving in accuracy and speed, but most patients find these applications and devices cumbersome.

General visual agnosia. Patients with general visual agnosia should be encouraged to use tactile and auditory cues to help recognize objects. The value of putting verbal labels on items in the environment depends on whether reading is spared. Organizing the environment to maintain the same location for items used frequently can reduce the need to search for items using touch. This approach will also allow a person to predict which objects they will encounter in a given place, making recognition easier. Patients should be advised not to drive and counseled about safety concerns in operating heavy machinery.

Prosopagnosia. Patients with prosopagnosia often benefit from learning which individuals will be present in a room before they enter. Patients can also be instructed to depend on voice recognition or on companions aware of their deficits. Patients should be encouraged to explain their deficits to friends and acquaintances when they are comfortable doing so. Nametags can be helpful in the right setting.

Topographagnosia. The widespread availability of devices with GPS capability has made route-finding much easier for patients with topographagnosia, although navigating inside buildings can be difficult. Advanced planning and excursions to familiar locations to discover compensatory methods are also helpful.

Alexia. Substitutive strategies include audiobooks and text-to-speech software applications available on many electronic devices or within audible text pages.

Outcomes

There is no clinical consensus on effective rehabilitative treatments or strategies for visual agnosias. The efficacy of restorative approaches is limited, and a recent systematic review highlights the paucity of rehabilitative investigations and their less-than-impressive results (60).

There is a long history of case reports describing attempts to improve face recognition in prosopagnosia (14; 38; 34). More recently, studies involving small cohorts of patients with prosopagnosia have shown that perceptual learning can lead to modest improvements in face-shape perception (37; 34; 29). Only anecdotal evidence supports the notion that this finding translates into benefits in daily life. One trial found benefits related to oxytocin treatment (14).

None of the treatments used for pure alexia has been accepted as a standard of care. These include increasing the spacing between words or phrases (15; 75), oral articulation during reading (27), repetitive oral reading of the text (15), and finger tracing of letters (75; 83).

There is sparse information about therapies for topographagnosia. Over-learning of familiar routes may help the navigation of some patients in daily life, although this ability may not generalize to new situations.

In the appropriate clinical context, assistive technology has considerable potential to support patients with visual agnosias, particularly those with stable deficits and minimal cognitive impairment. There have been few advances in the practical use of technologies for visual agnosia during the past decade (112). The use of assistive technologies for those with vision loss due to ocular disease is closer to becoming an everyday reality with recent progress in the field (89). This progress gives hope that what is gained and learned using assistive devices for ocular visual impairment can one day be translated into research and development of devices for visual brain dysfunction.