Posterior cortical atrophy

Posterior cortical atrophy (PCA) is a neurodegenerative syndrome that is characterised by progressive decline in visuospatial, visuoperceptual, literacy, and praxic skills. The progressive neurodegeneration affecting parietal, occipital, and occipitotemporal cortices that underlies PCA is attributable to Alzheimer's disease in most patients. However, alternative underlying causes, including dementia with Lewy bodies, corticobasal degeneration, and prion disease, have also been identified, and not all patients with PCA have atrophy on clinical imaging. This heterogeneity has led to inconsistencies in diagnosis and terminology and difficulties in comparing studies from different centres, and has restricted the generalisability of findings from clinical trials and investigations of factors that drive phenotypic variability. Important challenges remain, including the identification of factors associated not only with the selective vulnerability of posterior cortical regions but also with the young age of onset of PCA. Greater awareness of the syndrome and agreement over the correspondence between syndrome-level and disease-level classifications are needed to improve diagnostic accuracy, clinical management, and the design of research studies.


Introduction
Posterior cortical atrophy (PCA) is a neurodegenerative condition characterised by a progressive, often striking, and fairly selective decline in visual-processing skills and other functions that depend on parietal, occipital, and occipitotemporal regions of the brain. Age at onset of PCA is typically 50-65 years, and the syndrome is associated with various underlying pathological features. PCA has been recognised for more than two decades and yet, compared with other conditions, is relatively neglected by researchers. Patients often face considerable delays in diagnosis owing to the young age at onset and unusual symptoms at presentation. Furthermore, the term PCA has been applied in consistently, making comparison across studies diffi cult to achieve. Although there have been moves to defi ne neurodegenerative diseases by their underlying pathological features, such progress in relation to PCA is limited currently by scant specifi city in available diagnostic criteria and an absence of clarity with respect to relations between PCA and related syndromic classifi cations, such as aphasic, amnestic, and dysexecutive Alzheimer's disease pheno types and corticobasal syndrome.
In this Review we outline the clinical, psychological, imaging, epidemiological, genetic, and pathological features of PCA. We argue that, within pathological subgroups, characterisation of atypical phenotypes such as PCA will enable identifi cation of biological factors that promote or protect against pathological changes in specifi c brain networks. Problems with and possible solutions to current diagnostic and terminological conundrums are considered, with particular reference to implications for future design of clinical and research trials that include people with PCA. We also aim to increase awareness and improve identifi cation of early and unusual symptoms of PCA and to provide guidance on the provision of support, care, and education for patients, carers, and health-care professionals.

History and defi nitions
The term PCA was fi rst used to describe people with predominant defi cits in higher-order visual processing, a subset of whom also presented with striking atrophy in parieto-occipital areas of the brain. 1 The outlined syndrome accorded with other early reports of patients with similar clinical characteristics. [2][3][4][5][6][7][8][9] Without histopathological data, Benson and colleagues 1 deemed the clinical presentation of PCA to be suffi ciently distinct from that of Alzheimer's or Pick's disease "to warrant classing them separately until defi nitive pathologic information becomes available". Subsequent histopathological studies identifi ed Alzheimer's disease as the most common underlying pathological feature, leading to synonymous use of the terms PCA, biparietal Alzheimer's disease, and visual variant of Alzheimer's disease in some studies. [10][11][12][13][14] The term progressive posterior cortical dysfunction has also been used to describe clinical symptoms in aff ected individuals who do not have clear posterior atrophy. 15 However, PCA is also associated with pathological features that are not linked to Alzheimer's disease, which has led to calls for PCA to be regarded as a distinct nosological entity with its own diagnostic criteria. 16,17

Epidemiology
The prevalence and incidence of PCA are currently unknown; obtaining these data will depend on the adoption of consistent diagnostic criteria. Furthermore, any value is likely to be an underestimate because of poor general awareness of the syndrome. However, Snowden and colleagues 18 noted that, of 523 patients with Alzheimer's disease presenting to one specialist centre for cognitive disorders, 24 (5%) had a visual presentation (also labelled PCA) and a further 13 (3%) had an apraxic presentation.
Age at onset tends to be much earlier in patients with PCA than in those with typical amnestic Alzheimer's disease, with most studies recording age of PCA symptom onset from mid-50s to early 60s, 17,19 although some Review researchers have reported a wider age spread (45-74 years, 20 40-86 years). 21 In terms of sex distribution, no diff erence in prevalence has been noted on the basis of sex, 15,17,19 whereas other groups have reported an overrepresentation in women. 18,[20][21][22][23][24]

Neuropsychological features
The neuropsychological defi cits cited most frequently in individuals with PCA are visuospatial and visuoperceptual impairments, alexia, and features of Bálint's syndrome (simultanagnosia, oculomotor apraxia, optic ataxia, environmental agnosia) and Gerstmann's syn drome (acalculia, agraphia, fi nger agnosia, left-right disorientation; fi gure 1). 15,17,[19][20][21][25][26][27][28][29][30][31] Defi cits in working memory and limb apraxia have also been noted. 20 Findings of longitudinal studies indicate that anterograde memory, executive functions, and linguistic skills, which are sometimes preserved strikingly in earlier stages of disease, gradually deteriorate in some patients with PCA as they progress to a more global dementia state. 10,19,23 Although higher-order visual problems, such as diffi culties with object and space perception, are reported more often than are basic visual impairments, many such problems are at least partly due to defi cits in more basic visual processing-eg, form, motion, colour, and point localisation. In a detailed comparison of basic and higherorder perception, 32 all patients with PCA showed impairment in at least one basic visual process, emphasising the vulnerability of fundamental aspects of vision associated with occipital cortical dysfunction. This study also indicated specifi c correlations of basic visual processing with higher-order visuospatial and visuoperceptual skills, but not with non-visual parietal functions (such as calculation and spelling), suggesting specifi c involvement of visual networks in PCA. 32 Basic and higher-order visual defi cits, in combination, have predictable outcomes on the performance of PCA patients on general neuropsychological testseg, performance IQ in patients is often up to 30-40 points lower than verbal IQ. Performance on cognitive tasks with any relevant visual component (eg, visual memory recall, trail-making test, Stroop test) is vulnerable to impairment and misinterpretation, and thus accurate assessment requires selection of tasks that minimise visual demands (eg, auditory-verbal memory tasks, naming from verbal description).
Many patients with PCA also have unusual symptoms, referred to as positive perceptual phenomena, which include abnormally prolonged colour after-images, 33 reverse size phenomena, 34,35 perception of movement of static stimuli, 35 and in one individual, 35 180° upside-down reversal of vision. Reading skills can be limited by several processes, including visual disorientation (getting lost on the page), reverse size phenomena (accurately perceiving small but not large print), and visual crowding (impaired identifi cation of constituent letters of a word owing to excessive integration of visual features from surrounding letters). [36][37][38] PCA can also lead to primary peripheral dyslexias. 39,40 Anecdotally, individuals with PCA frequently report heightened sensitivity to glare from shiny surfaces and can experience a range of localised sensation and pain phenomena and disturbances of balance and bodily orientation, which could potentially be linked to deranged visuovestibular interactions.
Should PCA be deemed a unitary clinicoanatomical syndrome or, rather, a collection of related but distinct syndromic subtypes? By extrapolation from basic neuroscientifi c evidence of distinct cortical streams that process diff erent types of visual information, 41,42 researchers have suggested that separate parietal (dorsal), occipitotemporal (ventral), and primary visual (striate cortex, caudal) forms of PCA exist. 12,13 However, these claims are based on fi ndings from individual case reports. Subsequent studies of neuropsychological case series have failed to provide evidence to support a pure ventralstream syndrome 19 and, rather, have indicated considerable overlap in neuropsychological profi les and patterns of cortical thinning in patients with behaviourally defi ned predominantly dorsal-stream or ventral-stream

Review
impairments. 32 Thus, these fi ndings suggest that phenotypic diff erences might most appropriately be judged to represent points on a continuum of variation within PCA.

Clinical features
Clinical presentation of PCA is aff ected by several factors. These include time taken before an individual presents to medical services or is referred to a cognitive specialist; the specifi c pattern of defi cits; in some people, the underlying pathological features; and the patient's psychological response to their symptoms. The relative rarity of PCA, the sometimes unusual nature of its symptoms, and the fairly young age at onset can lead to misdiagnosis of many patients as depressed, anxious, or even malingering in early stages of the disease. Early anxiety is (at least anecdotally) a common feature, perhaps indicating that patients with PCA typically have some insight into the possibility that they have a medical problem, even if its nature is unclear. Furthermore, even to experienced cognitive neurologists, a patient's initial history can be more suggestive of anxiety, until examination shows impairment that relates to function of the parietal lobe, the occipital lobe, or both of these areas. Patients are usually referred fi rst to opticians and ophthalmologists in the belief that an ocular abnormality is causing their visual symptoms, sometimes leading to unnecessary medical procedures such as cataract surgery.
The symptoms reported by a patient with PCA probably indicate broadly their individual pattern of neuropsychological impairment. Visual symptoms are perhaps more likely to be mentioned than other posterior defi cits, with individuals describing diffi culties reading lines of text, judging distances (often leading to repeat minor car accidents or diffi culties parking), identifying static objects within the visual fi eld, or having problems with stairs and escalators (panel). Visual symptoms such as light sensitivity or visual distortions can be mistaken for migraine. Careful history taking could reveal some of the more unusual visual phenomena described above, including the presence of prolonged after-images or visual crowding. Individuals might volunteer diffi culties in using common objects, suggesting dyspraxia, or describe progressive diffi culty with calculations or spelling. The presence of other neurological symptoms, including visual hallucinations (reported in up to 25% of patients with PCA) 19,21,43 and rapid eye movement (REM) sleep behaviour disorder, could be suggestive of underlying dementia with Lewy bodies. Very occasionally, the patient's history might be consistent with occipital lobe seizures.
Careful bedside testing can elicit signs of disproportionate parietal or occipital dysfunction, including (but not restricted to) visual disorientation, diffi culties resolving degraded stimuli, ideational or ideomotor dyspraxia (or both), dyscalculia, and problems with spelling (fi gure 2A). The physical examination in most cases of PCA is unremarkable; however, if severe cortical visual impairments are present, interpretation of visual acuity and visual fi elds can be diffi cult, with hemianopia sometimes misdiagnosed owing to the presence of higher-order visual attentional impairments. Finger myoclonus is also not uncommon. Snowden and colleagues 18 reported similar frequencies of extrapyramidal signs (41%), myoclonus (24%), and grasp refl ex (26%) in individuals with PCA compared with those seen in patients with typical Alzheimer's disease. Nonetheless, features of clear symmetrical motor parkinsonism (suggestive of Lewy body pathology) or prominent asymmetric myoclonus and dystonia (suggestive of corticobasal degeneration) can give important clues to the underlying cause of PCA, although pathological data on which these clinical observations can be confi rmed are currently sparse.

Neuroimaging
Increasingly advanced image analysis techniques have been used to localise and quantify (typically group) diff erences in patterns of atrophy in patients with PCA, compared either with controls or with patients with typical Alzheimer's disease. Cross-sectional voxel-based morphometry has shown widespread diff erences in grey matter between patients with PCA and healthy controls, with the most signifi cant reductions found in regions of the occipital and parietal lobes followed by areas in the Panel: Case study A 62-year-old right-handed woman presented with 4 years of progressive visuospatial dysfunction. Her fi rst symptom was diffi culty seeing when driving at night. In the following years she frequently dented her car when parking, tended to bump into doors on her right side, and had trouble locating items even when they were directly in front of her. She reported problems reading, trouble distinguishing between currency bank notes, and diffi culty deciding whether to push or pull a door to open it. When she watched television, images seemed to move slowly. She was referred to the cognitive neurology clinic by an ophthalmologist, who had ruled out primary ocular disease. On neurological assessment she was fully oriented and was an excellent historian. On testing of visual fi elds, she was inconsistent in counting fi ngers in the right hemifi eld. Pupillary responses and extraocular movements were normal, although she was slow initiating saccades and had diffi culty reaching for items under visual guidance. Her physical neurological examination was otherwise normal. On cognitive testing, her MMSE score was 26/30, and she showed severe impairment when copying intersecting pentagons and the Benson fi gure (fi gure 2A). She was able to name colours correctly but showed moderate diffi culty matching faces. She had severe diffi culty reading, which was improved by spelling words out loud, and had mild defi cits in confrontation naming (improved with cues) and category fl uency. Verbal memory, phonemic fl uency, and attention were intact. She could not complete many tasks because of visual dysfunction. Brain MRI showed striking atrophy in bilateral parietal, posterior temporal, and lateral occipital cortex Review temporal lobe. 24,30 Direct comparison between individuals with PCA and typical Alzheimer's disease, using both voxel-based morphometry and cortical thickness measures, has shown greater right parietal and less left medial temporal and hippocampal atrophy in patients with PCA. It is noteworthy that in several studies, researchers report asymmetric atrophy patterns in PCA (right greater than left), but these diff erences could be due to selection biases in the diagnosis and recruitment of patients with prominent visual dysfunction. Limited data from diff usion tensor imaging studies also suggest that PCA reduces the integrity of white matter tracts in posterior brain regions. [44][45][46] However, considerable regional overlap in atrophy has also been reported, with regions including the posterior cingulate gyri, precuneus, and inferior parietal lobe being aff ected in both PCA and typical Alzheimer's disease. 23 Such fi ndings suggest that PCA, when associated with Alzheimer's disease, exists on a spectrum of variation with other phenotypes of Alzheimer's disease. 23,32 Fluid registration of longitudinally acquired structural magnetic resonance (MR) images shows the evolution of PCA (fi gure 3), with fi ndings of group studies indicating that, by 5 years of symptom duration, atrophy is widespread throughout the cortex, including medial temporal lobe structures. 31 Data from functional imaging studies using single photon emission computed tomography (SPECT) and fl uorodeoxyglucose (FDG) PET accord largely with structural changes in parieto-occipital areas (fi gure 2B). 20,[47][48][49][50][51] In addition to posterior regions, FDG-PET has indicated specifi c areas of hypometabolism in the frontal eye fi elds bilaterally, which can occur secondary to loss of input from occipitoparietal regions and be the cause of oculomotor apraxia in PCA (fi gure 1). 20,52 In a few studies, researchers have also assessed patterns of amyloid deposition with Pittsburgh compound B (PiB)-PET in patients with PCA. Case studies and small series have shown increased accumulation of amyloid-β, predominantly in the occipital and parietal lobes, relative to individuals with typical Alzheimer's disease. [53][54][55][56] However, in two studies in which PiB uptake was compared in large groups of patients with PCA and typical Alzheimer's disease, no signifi cant diff erence was reported in amyloid deposition between these groups, with both showing diff use PiB uptake through out frontal, temporoparietal, and occipital cortex (fi gure 2B). 22,57

Genetics
To date, no reports have been published of a PCA phenotype associated with autosomal dominant familial Alzheimer's disease; however, in a case series, 58 the PCA syndrome was described in familial prion disease associated with a 5-octapeptide insertion into the prion protein. Findings from available PCA case series suggest that there is no signifi cant diff erence in the number of patients with a positive family history of dementia in PCA and with typical Alzheimer's disease. 17,21 In some studies, signifi cant diff erences between apolipoprotein E (APOE) genotypes have been noted in patients with PCA versus amnestic Alzheimer's disease. 14,18,59 However, in other studies, no diff erence in APOE was recorded between PCA and typical Alzheimer's disease (table). 17,[21][22][23][60][61][62][63][64] Discrepancies between these studies could be attributable to diff erences in the inclusion criteria used to defi ne PCA and typical Alzheimer's disease and age at onset. In particular, scant pathological confi rmation is a major limitation of these studies.
Does the APOE ε4 allele drive the degenerative pattern in patients with Alzheimer's disease (ie, to or away from medial temporal structures)? Is the low frequency of APOE ε4 in PCA that is reported in some studies attributable to patients with non-Alzheimer's disease Genome-wide association studies to assess frequencies of other genetic risk factors for sporadic Alzheimer's disease will be useful.

Pathology
Findings of pathological studies all show that Alzheimer's disease is the most common underlying cause of PCA. 13 21 Although the distribution patterns of pathological features diff er between PCA and typical Alzheimer's disease, the exact pattern of pathological changes is inconsistent and based on very few cases. Some studies have shown diff erences in both plaques and neurofi brillary tangles between PCA and typical Alzheimer's disease, 10,12,72 whereas others have recorded no diff erences in plaque distribution. 15,21 For example, Levine and colleagues 10 reported pathological fi ndings of one patient with PCA who showed the greatest density of senile plaques and neurofi brillary tangles in occipitoparietal regions and the lowest density in frontal lobe regions. Hof and co-workers 72,73 reported similar fi ndings, with plaques and tangles found predominantly in primary visual and visual association areas around the occipitoparietotemporal junction, whereas frontal regions-such as the prefrontal cortex-had very low densities of pathological changes. By contrast, Tang-Wai and colleagues 21 looked at pathological changes in nine patients with PCA versus 30 with typical Alzheimer's disease. The PCA group showed signifi cantly higher densities of neurofi brillary tangles in visual and visualassociation cortices and fewer tangles and senile plaques in the hippocampus and subiculum. However, density of senile plaques in other cortical areas was comparable in both groups. 21 Reasons for the discrepant fi ndings in these autopsy studies could include diff erences in inclusion criteria and demographic characteristics (such as age and disease severity) and diff erences in the methods used to quantify pathological changes (such as diff erent staining techniques and discrimination between diff use and neuritic plaques). Studies in which CSF biomarkers (Aβ 1-42 , T-tau, and P-tau 181 ) were assessed have recorded similar fi ndings in patients with PCA compared with Alzheimer's disease, 27,56,60,74 lending support to previous reports that PCA is associated typically with underlying Alzheimer's disease pathology.

Diagnostic and research criteria
Two sets of diagnostic criteria for PCA have been proposed. 17,21 Suggested core features for a diagnosis of PCA include insidious onset and gradual progression; presentation of visual defi cits in the absence of ocular disease; relatively preserved episodic memory, verbal fl uency, and personal insight; presence of symptoms including visual agnosia, simultanagnosia, optic ataxia, ocular apraxia, dyspraxia and environmental disorientation; and absence of stroke or tumour. Supportive features include alexia, ideomotor apraxia, agraphia, acalculia, onset before the age of 65 years, and neuroimaging evidence of PCA or hypoperfusion.
Although these criteria have proved useful in several clinical and research contexts, they are based on clinical experience at single centres and have not been validated more widely. Without objective evidence linking clinical phenotype to underlying pathology, there continues to be inconsistency, with the term PCA being used as a descriptive syndromic term and as a diagnostic label. Such inconsistencies present several diffi culties in Review attributing and assessing the validity of a diagnosis of PCA and, particularly, in the design and interpretation of research studies and clinical trials. First, although a syndromic classifi cation could be adequate for some types of research study (eg, brain-behaviour, behavioural intervention), other investigations will need direct consideration of probable underlying pathological features (eg, clinical trials of disease-specifi c drugs). Second, at present, we have no evidence base on which to judge the eff ectiveness of pharmacological treatments for Alzheimer's disease in individuals with PCA attributable to probable Alzheimer's disease or to decide whether individuals with PCA should be included or excluded from conventional clinical trials of Alzheimer's disease-eg, because of the potential unsuitability of study outcome measures (eg, visual memory tasks) selected for patients with more typical amnestic or global clinical presentations. Third, current criteria provide no guidance about the degree of specifi city needed for a diagnosis of PCA. For example, in the relatively large series reported by Renner and colleagues, 15 nine of 27 patients presented with PCA as a fairly isolated disorder, whereas in the remaining 18 people it was the prominent feature of a more generalised dementia. Several groups have suggested that PCA, when attributable to probable Alzheimer's disease, lies on a phenotypic continuum with other typical and atypical Alzheimer's phenotypes (eg, amnestic Alzheimer's disease, global cognitive impairment, logopenic or phonological aphasia), 18,23,32 but the boundaries between such phenotypes are defi ned imprecisely. Fourth, the presentation of visual complaints is a core feature of existing criteria but some patients with neurodegenerative disorders present with predominant impairment of other posterior cortical functions, such as calculation, spelling, and praxis; 6,18,27,48,75 such individuals could be deemed to fall within the PCA spectrum. Finally, the value of biomarkers might diff er in PCA compared with typical Alzheimer's disease or dementia with Lewy bodies (eg, relative absence of hippocampal atrophy). This issue is especially important in view of the increasing incorporation of such biomarkers in disease-specifi c diagnostic criteria. 76,77 Future resolution of these issues and development of clinical and research criteria for the defi nition of PCA are likely to be based on a consensus of opinion from many specialist centres, supported by objective evidence of the relation between clinical presentation, neuroimaging and CSF biomarkers, and histopathological data. Establishment of the relative likelihood of diff erent pathologies in large, multicentre datasets would improve the discrim ination of potential disease subtypes necessary for trials of disease-modifying agents. One possible approach would be to apply a range of criteria to a multicentre dataset to establish sets of inclusion and exclusion criteria that identify specifi c disease subgroups (eg, PCA with Alzheimer's disease). By consensus, experts could also investigate frameworks for making criteria useable, in terms of a quantifi able set of diagnostic markers, to help with enrolment into research studies and to improve the comparability of data between institutions.

Management
As far as we know, no reports have been published that assess the eff ectiveness of acetylcholinesterase inhibitors (eg, donepezil, rivastigmine, and galantamine) in PCA. However, these drugs are frequently and, in our view (since Alzheimer's disease is the most likely underlying pathology), appropriately administered. Clinical experience and a few case reports suggest some clinical benefi t, 78 most probably in patients with underlying pathological features of Alzheimer's disease or dementia with Lewy bodies. Antidepressant drugs might also be appropriate in patients with persistent low mood, and trials of levodopa or carbidopa could be useful in individuals with parkinsonism.

Review
Because of poor awareness of PCA, patients with this disorder typically receive scant or inappropriate care and advice, whereby diffi culties that are less important to the individual (eg, memory problems) are targeted while vital perceptual diffi culties (eg, many activities in day centres and nursing homes are visually mediated) are, in general, not considered. Preservation of skills such as memory, language, and insight in PCA, especially in mild-tomoderate stages of the disease, enables patients to take advantage of peer-support meetings and group, couple, and individual psychological treatments when the need exists. Support group meetings are especially useful in tackling social isolation, allowing patients to share the experience of what is often a long and diffi cult route to diagnosis, and to exchange practical tips and coping strategies and advice. Patients with PCA often benefi t from resources designed primarily for blind and partially sighted people, such as mobile phones with simplifi ed displays, voice recognition software, talking books and watches, culinary aids, and lamps to increase ambient light levels in the home. Referral to an occupational therapist or sensory team could be appropriate to help a patient participate more fully in activities of daily living. An individual might also need to be referred to an ophthalmologist to register as partially sighted under statutory invalidity schemes, which might then provide access to fi nancial and social benefi ts and services. Driving a car is not appropriate for many people with PCA, particularly those with prominent visual disturbance. Physical therapy can also be helpful for individuals with parkinsonism and gait disturbance. Scant empirical evidence shows the eff ect of management strategies in PCA, but a rehabilitation programme that included psychoeducation, compensatory strategies, and cognitive exercises was tested in an individual with PCA, 79 resulting in small improvements in visuo perceptual functioning.

Conclusions
Posterior cortical atrophy is a debilitating and underrecognised focal degenerative syndrome that is associated with a range of diff erent disease pathologies. The core features of the syndrome are suffi ciently homogeneous to justify regarding PCA as an independent nosology, with Alzheimer's disease as the most common underlying cause. However a lack of consistency in the classifi cation of PCA is likely to continue unless diagnostic criteria and terminology are standardised. The criteria proposed in this Review attempt to take both the clinical and histo pathological features of PCA into account and to introduce quantifi able behavioural inclusion criteria for research studies of PCA. Better understanding and aware ness of the syndrome by medical and lay workers is necessary to improve diagnosis and treatment and to enhance support services for individuals with PCA and their families. Identifi cation of the distinctive patterns of struc tural, functional, cognitive, and genetic changes in PCA could provide new insights into the pathogenesis and clinical features of typical Alzheimer's disease and into general mechanisms of visual network function and degeneration. Dedicated trials are needed to assess the eff ectiveness of pharmacological and non-pharmacological interventions in PCA, and to identify factors that drive phenotypic variability in this small but important population of patients who typically have early-onset dementia.

Contributors
All authors contributed to the writing and reviewing of this Review.

Confl icts of interest
NCF has served on scientifi c advisory boards of the Alzheimer's Research Forum, the Alzheimer's Society, and the Alzheimer's Research Trust, and holds a patent for QA Box that might accrue revenue. In the past 5 years his research group has received payment for consultancy or for undertaking studies from Abbott Laboratories, Elan Pharmaceuticals, Eisai, Eli Lilly, GE Healthcare, IXICO, Lundbeck, Pfi zer, Sanofi -Aventis, and Wyeth Pharmaceuticals. JMS has received payment for undertaking studies from AVID Radiopharmaceuticals. SJC, ML, GDR, and MNR declare that they have no confl icts of interest.

Search strategy and selection criteria
We identifi ed references for this Review by searching PubMed with the terms: "posterior cortical atrophy", "biparietal Alzheimer's disease", "visual dementia", and "Balint's syndrome dementia". We looked for articles published between 1970 and November, 2011. We also identifi ed articles by searching our own fi les. Only reports published in English were reviewed, with the exception of historical manuscripts (before 1988). The fi nal reference list was generated on the basis of originality and relevance to the broad scope of this Review.