This article was downloaded by: [Institute of Neurology] On: 9 September 2010 Access details: Access Details: [subscription number 917397660] Publisher Psychology Press Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37- 41 Mortimer Street, London W1T 3JH, UK

Neurocase Publication details, including instructions for authors and subscription information: http://www.informaworld.com/smpp/title~content=t713658146

Abnormal visual phenomena in posterior cortical atrophy Sebastian J. Crutcha; Manja Lehmanna; Nikos Gorgoraptisb; Diego Kaskic; Natalie Ryana; Masud Husainb; Elizabeth K. Warringtona a Dementia Research Centre, Department of Neurodegeneration, UCL Institute of Neurology, University College London, UK b University College London, Institute of Cognitive Neuroscience, UK c Department of Neuro-otology, Charing Cross Hospital, London, UK

First published on: 02 September 2010

To cite this Article Crutch, Sebastian J. , Lehmann, Manja , Gorgoraptis, Nikos , Kaski, Diego , Ryan, Natalie , Husain, Masud and Warrington, Elizabeth K.(2010) 'Abnormal visual phenomena in posterior cortical atrophy', Neurocase,, First published on: 02 September 2010 (iFirst) To link to this Article: DOI: 10.1080/13554794.2010.504729 URL: http://dx.doi.org/10.1080/13554794.2010.504729

PLEASE SCROLL DOWN FOR ARTICLE

Full terms and conditions of use: http://www.informaworld.com/terms-and-conditions-of-access.pdf

This article may be used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material. NEUROCASE 2010, iFirst, 1–18

NNCS Abnormal visual phenomena in posterior cortical atrophy

Abnormal Visual Phenomena in PCA Sebastian J. Crutch,1 Manja Lehmann,1 Nikos Gorgoraptis,2 Diego Kaski,3 Natalie Ryan,1 Masud Husain,2 and Elizabeth K. Warrington1 1Dementia Research Centre, Department of Neurodegeneration, UCL Institute of Neurology, University College London, UK 2Institute of Cognitive Neuroscience, University College London, UK 3Department of Neuro-otology, Charing Cross Hospital, London, UK

Individuals with posterior cortical atrophy (PCA) report a host of unusual and poorly explained visual distur- bances. This preliminary report describes a single patient (CRO), and documents and investigates abnormally prolonged colour afterimages (concurrent and prolonged of colours complimentary to the colour of an observed stimulus), perceived motion of static stimuli, and better reading of small than large letters. We also evaluate CRO’s visual and vestibular functions in an effort to understand the origin of her experience of room tilt illusion, a disturbing phenomenon not previously observed in individuals with cortical degenerative disease. These visual symptoms are set in the context of a 4-year longitudinal neuropsychological and neuroimaging investigation of CRO’s visual and other cognitive skills. We hypothesise that prolonged colour after-images are attributable to relative sparing of V1 inhibitory interneurons; perceived motion of static stimuli reflects weak magnocellular function; better reading of small than large letters indicates a reduced effective field of vision; and room tilt illusion effects are caused by disordered integration of visual and vestibular information. This study contributes to the growing charac- terisation of PCA whose atypical early visual symptoms are often heterogeneous and frequently under-recognised.

Keywords: Posterior cortical atrophy (PCA); Alzheimer’s disease (AD); Reversal of vision metamorphopsia; instability; Colour perception.

INTRODUCTION Rogelet, Delafosse, & Destee, 1996; Ross et al., 1996) and indeed the condition is often referred to Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 Posterior cortical atrophy (PCA) is a progressive as the ‘biparietal’ or ‘visual’ variant of AD, or neurodegenerative condition involving prominent alternatively as Benson’s syndrome after the ori- tissue loss in the occipital, parietal and posterior ginal clinical description (Benson et al., 1988). The temporal cortex. PCA is characterised by the pro- distinction between PCA and the typical amnestic gressive deterioration of higher presentation of AD reflects the distribution of the and other posterior cortical functions, with first AD pathology. Individuals with PCA tend to have symptoms often occurring before the age of 65 a much greater density of senile plaques and neu- (e.g., Benson, Davis, & Snyder, 1988; Cogan, 1985; rofibrillary tangles in occipital cortex and regions De Renzi, 1986; Galton, Patterson, Xuereb, & of posterior parietal cortex and temporo-occipital Hodges, 2000). Alzheimer’s disease (AD) is the junction than those with typical AD, whilst showing most common underlying pathology (Cogan, 1985; fewer pathological changes in more anterior areas

Address correspondence to Dr Sebastian J. Crutch, Dementia Research Centre, Box 16, National Hospital for Neurology and Neurosur- gery, Queen Square, London WC1N 3BG, UK. (E-mail: [email protected]).

© 2010 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business http://www.psypress.com/neurocase DOI: 10.1080/13554794.2010.504729 2 CRUTCH ET AL.

(e.g., prefrontal cortex; Hof, Vogt, Bouras, & which are poorly recognised but nonetheless may Morrison, 1997; Tang-Wai et al., 2004). These and prove to be characteristic of PCA. This preliminary other differences between PCA and typical AD report replicates evidence of prolonged colour may be underpinned partially by genetic risk factors afterimages and better reading of small than large in the ApoE ε4 allele, which is strongly associated letters, and describes two further abnormal visual with typical AD but which does not seem to be a phenomena, namely perceived motion of static risk factor in PCA (Schott et al., 2006; Snowden stimuli and room tilt illusion effects. We argue that et al., 2007). these cognitive phenomena should be targets for In accordance with the distribution of pathology systematic evaluation in future cohort studies of and atrophy, patients with PCA tend to show rela- PCA, and have significant implications for tive sparing of many aspects of cognition such as patients’ functional abilities. memory and language and often have full insight into their condition. Thus in the early stages of the disease, they do not meet the standard diagnostic CLINICAL DESCRIPTION criteria for typical AD. Instead, the most promi- nent symptoms tend to include early visual pro- CRO is a right-handed retired health professional cessing deficits (e.g., achromatopsia, figure-ground who was referred to the National Hospital in 2004 discrimination problems), apperceptive agnosia, (aged 57 years old) with a 4-year history of progres- prosopagnosia, spatial deficits (e.g., visual disori- sive visual impairment. Early visual symptoms entation, spatial agnosia) and other symptoms included difficulty in making depth and distance such as optic ataxia, gaze apraxia and limb apraxia judgements, and the appearance of coloured patches (Benson et al., 1988; Freedman et al., 1991; Levine, at the edge of her visual field. Investigations con- Lee, & Fisher, 1993; McMonagle, Deering, Berliner, ducted prior to referral included a brain MRI which & Kertesz, 2006; Mendez, Ghajarania, & Perryman, was reported as normal, and an EEG showing poorly 2002; Rogelet et al., 1996; Ross et al., 1996). sustained alpha rhythm with temporal lobe slowing. Patients with PCA also experience other related Normal electroretinography (ERG) and visual and unrelated behavioural symptoms including evoked potentials (VEPs) investigations suggested dyslexia, acalculia, dysgraphia and visual inatten- normal retinal and functions. Initial neu- tion (see Caine, 2004 for a review). ropsychological assessment also corroborated the However, in addition to these well-recognised clinical impression of predominant posterior cortical neuropsychological deficits, patients with PCA damage, with particular weaknesses on tests of visuo- also report a host of other symptoms, the anatomical spatial and visuoperceptual functions especially for and cognitive basis of which are much more poorly complex stimuli. Visual inspection of a brain MRI understood. For example, the tendency for indicated subtle biparietal atrophy but no hippocam- patients with progressive visual disturbance (who pal volume loss or vascular disease. There was no

Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 might now be considered to have PCA) to have evidence of brainstem or cerebellar infarction on this greater difficulty recognising large as compared or subsequent MRI scans, which included diffusion with small pictures and other stimuli has been weighted imaging sequences. described (e.g., Pick, 1908 cited in Bender and Consequently, CRO was given a diagnosis of Feldman, 1972; Coslett, Stark, Rajaram, & Saffran, posterior cortical atrophy (PCA) owing to proba- 1995; Kartsounis and Warrington, 1991; Saffran, ble Alzheimer’s disease. This diagnosis was made Fitzpatrick-DeSalme, & Coslett, 1990; Stark, on the grounds of the clinical information summa- Grafman, & Fertig, 1997). Another abnormal rised above but also upon the fulfilment of behav- effect described previously in a patient with PCA is ioural criteria employed routinely at the Dementia the experience of abnormal colour afterimages and Research Centre. These criteria require an individ- washes of colour (e.g., Chan, Crutch, & Warrington, ual to demonstrate memory function above the 5th 2001). However, despite occasional reports, such percentile and at least 2 out 4 scores below the 5th unusual or atypical symptoms are often poorly rec- percentile on the following 4 tests: Number loca- ognised, and sometimes contribute to incorrect tion and Object Decision subtests from the Visual diagnoses of malingering, anxiety or ophthalmolog- Object and Space Perception battery (VOSP; ical disease in the earliest stages of the condition. Warrington & James, 1991) and graded difficulty In this paper, we report a single patient with tests of arithmetic and spelling (Baxter & PCA who exhibits a number of unusual symptoms Warrington, 1994; Jackson & Warrington, 1986). ABNORMAL VISUAL PHENOMENA IN PCA 3

On further clinical examination in December Visual Object and Space Perception Battery 2009 (aged 62 years old), 8 years into her illness, (VOSP; Warrington & James, 1991), and a shape she scored 21/30 on the Mini-Mental State Examina- discrimination test (based on Efron, 1968). Visuo- tion. Neurological examination revealed marked perceptual processing was examined using the visual disorientation and difficulties with visual VOSP Fragmented Letters and Object Decision fixation and praxis. A full neuro-otological assess- subtests, and the canonical/non-canonical view ment including detailed examina- perception test of Warrington and James (1988). tion was conducted at the time and is reported in Finally, visuospatial processing was investigated section 7 (Room tilt illusion). She continued to using the VOSP Number Location and Dot Count- have no limb myoclonus or extrapyramidal signs ing subtests, and the Motor Screening test from the and her deep tendon reflexes were all depressed but Cambridge Neuropsychological Test Automated symmetrical with flexor plantar responses. Sensory Battery (CANTAB®, Cambridge Cognition). examination, including joint position sense, was The scores from the neuropsychological assess- normal throughout. There was no history of ments are shown in Table 1. During the course of formed visual hallucinations, fluctuating cognition the 4-year interval between the first and last or symptoms of a REM sleep disorder to suggest assessments, CRO continued to score in the nor- Lewy body disease. mal range on tests of verbal recognition memory, Written informed consent was obtained using word retrieval, word comprehension, spelling and procedures approved by the National Hospital for visual acuity. It was noted though that her epi- Neurology and Neurosurgery and Institute of sodic memory and spelling skills in particular did Neurology Joint Research Ethics Committee. show signs of deterioration during this period from their previously superior levels. CRO’s ver- bal short-term memory, upper limb praxis and LONGITUDINAL NEUROPSYCHOLOGICAL cognitive estimation skills were also intact ini- AND NEURORADIOLOGICAL EVIDENCE tially but were observed to be impaired at visits 3 and 4, whilst calculation was impaired at all CRO completed a neuropsychological assess- assessments. Considering tasks with a visual com- ment in August 2004, and then participated in ponent, CRO was observed to be significantly several research visits before entering a clinical impaired on all tests of visuoperceptual and visu- trial examining the effectiveness of the acetylcho- ospatial processing from the outset of our investi- linesterase inhibitor donepezil in posterior corti- gations, with particularly poor performance of cal atrophy between October 2004 and April tasks involving the apperception of non-canonical, 2005. Following completion of the trial, CRO degraded and spatially dispersed stimuli. It is of attended for a series of 4 yearly research assess- note that at the baseline assessment, impover- ments between January 2006 and November 2008, ished space and object apperception were con-

Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 each involving a neuropsychological assessment trasted with intact performance on a range of and a 1.5-T MRI scan. more basic, early visual processing tasks (shape The neuropsychological evaluation included detection, shape discrimination, hue discrimina- measures of general cognitive function (Folstein et al., tion). However, her early visual processing skills 1975), verbal recognition memory (Warrington, soon followed the deterioration in her higher 1996), verbal synonym comprehension (Warrington order perceptual functioning with increasing diffi- et al., 1998), naming to verbal description, cogni- culties on discrimination and visual crowding tive estimation (Shallice & Evans, 1978), and psy- tasks over subsequent visits. chomotor speed (Willison & Warrington, 1992). A MRI scans were acquired on a 1.5-T GE Signa series of literacy (reading: James et al., 2001; spelling: scanner (General Electric, Milwaukee, WI) using Baxter & Warrington, 1986), numeracy (Jackson & an inversion-recovery prepared fast SPGR Warrington, 1986), praxis (Crutch, 2004) and sequence (TE = 5.4 ms, TR = 12 ms, TI = 650 ms). short-term memory tasks (digit span forwards) T1-weighted volumetric images were obtained with were also administered. Early visual processing a 24-cm field of view and 256 × 256 matrix to pro- was assessed using the Visual Acuity, Hue Discrim- vide 124 contiguous 1.5-mm thick slices in the ination and Visual Crowding subtests from the coronal plane. All scans were corrected for intensity Cortical Visual Screening Test (CORVIST; James inhomogeneities using the N3 algorithm (Sled, et al., 1991), the Shape Detection subtest from the Zijdenbos, & Evans, 1998). Each repeat was %ile %ile th th Baxter and Baxter 3–4 27–30 %ile 7 ∼ ∼ nd 14 %ile %ile %ile %ile %ile %ile %ile %ile %ile %ile %ile %ile acy, and early visual, th st th st st th st st st st st st controls (N = 15): 909ms 909ms = 15): (N controls 296ms) (sd full GDST, and WAIS-R full GDST, *Pro-rated. UT, Untestable. Jackson and Warrington (1986). (1986). Warrington and Jackson e skills, literacy, numer 6 and Warrington, 1983), Warrington, and Warrington and James (1988). (1988). James and Warrington 12 Efron (1968). (1968). Efron 11 Willison and Warrington (1992). (1992). Warrington and Willison 5 uospatial processing trieval and comprehension, frontal-executiv ans (1978). TABLE 1 Warrington and James (1991). (1991). James and Warrington 10 Shallice and Ev 4 Baseline Baseline visuoperceptual and vis Aug 04–Jan 05 04–Jan Aug 06 1 Jan Visit 072 Jan Visit 3 Jan 08 Visit 08 4 Nov Visit Norms/comment 20/20 18/20 18/20 16/20 15/20 errors no controls: Healthy –– 28/30 – 28/30– 25/30 24/25 – 23/30 20/25 range: Normal 23/25 15/15 >25 12/15 11/15 4: <5 3 and Visits 45/50* 24/25 22/25 24/25 20/25 >5 – 14/26 9/26 9/26 9/26 < 1 Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 sts, including verbal memory, word re Letters identifiedLetters – 14/19 14/19 10/19Usual 7/19 10/20 <1 – – 6/20 – <1 Completion time – 46s 82s 60sUnusual 2/20 105sMissed touches <1 – – 2 2/20 – 5 <1 11 Warrington et al. (1998). (1998). al. et Warrington 3 ion Memory Test (Warrington, 1984), Graded Naming Test (visual presentation; McKenna presentation; (visual Test Naming Graded 1984), (Warrington, Test Memory ion ) 19/20 15/20 16/20 16/20 16/20 <2 1–4: Visits 10 James, Plant, and Warrington (2001). (2001). Warrington and Plant, James, 9 (words; 2 – addition – addition 6 20 items) 20 25/30* 16/20 15/20 10/20 9/20 >5 st – Snellen) 6/9 6/9 6/9 6/9 6/9 normal Warrington (1996). (1996). Warrington 9 12 2 3 8 (Oblong edge ratio ) 16/16 – 16/16 15/16 14/16 – 11 Crutch (2004). (2004). Crutch 9 (error score) (error – 2 2 11 10 4: <5 3 and Visits 8 4 – Set B, 1 – Set 7 5 1 1:1.20) auditory/visual) items only) items Performance on background neuropsychological te Folstein et al. (1975). Visual crowding processing Visuoperceptual Object decision (VOSP) 10/20 – 8/20 – 4/20 4/10 4/20* 7/10 UT 3/10 <1 Impaired Short Recognition Memory Test Memory Recognition Short Figure-ground discrimination (VOSP discrimination Figure-ground Calculation (Extended GDA Digit Span. Digit TEST MMSE Warrington (1994). (1994). Warrington Reading (CORVIST Reading (VOSP) letters Fragmented views usual and Unusual processing Visuospatial (VOSP) location Number Dot Counting (VOSP) (CANTAB)Test Screening MOT (ms) RT Mean – 11/20 – 5/10 – 3/10 3/10 – 1631 7/10 1/10 1375 0/20 2/10 0/10 2025 UT 1/10 UT female Age-matched <1 1/10 <1 <1 Shape discrimination Shape 1 Naming (verbal description) (verbal Naming Cognitive estimates 22/30* 19/20 18/20 19/20 19/20 range Normal Digit span (maximum forwards) 7*6755Normal range: 5–9 range: 7*6755Normal forwards) (maximum span A Cancellation Digit Early visual processing (CORVIST acuity Visual Concrete synonyms test synonyms Concrete test production Gesture Hue discrimination (CORVIST) discrimination Hue 4/4 3/4 3/4 4/4 1/4 range: Normal Spelling (GDST *Different testversions employed: Recognit 4 ABNORMAL VISUAL PHENOMENA IN PCA 5

subsequently co-registered to baseline using an aff- the whole brain, however, the anterior-posterior ine (12 dof) (3 translations, 3 rotations, 3 scalings gradient was still clearly visible. and 3 shears) brain–brain registration, to correct for any voxel-size changes due to slight variation in the scanner magnetic field. This registration also ABNORMAL COLOUR AFTER-IMAGES used a differential bias correction to correct for dif- ferences in intensity inhomogeneity between serial As noted in the clinical description, one of CRO’s scans (Lewis & Fox, 2004). presenting ;features was of perceiving coloured Fluid registration was used to visualize areas of patches at the edge of her visual field. During sub- contraction and expansion across the whole brain sequent visits, she also described ‘washes of colour’ for each time step (Christensen, Rabbitt, & Miller, across the visual environment and particular diffi- 1996). The fluid algorithm used here was largely as culty when moving from areas of bright light to described by Freeborough and Fox (1998). A areas of lower light (e.g., going from the garden medial sagittal view of CRO’s right hemisphere into the house on a bright day), with her vision from the baseline scan at visit 1 as well as the fluid- taking a long time to adapt to the new levels of registered scans at visits 2–4 are shown in Figure 1. ambient light. She also reported seeing cream-col- As indicated by the colour-coded voxel-compression oured stripes when clearing her white bed linen. map, CRO showed progressive atrophic changes This observation was reminiscent of a previously with a marked anterior-posterior gradient over this reported individual with PCA (Patient VL; Chan 3-year interval. The pattern of atrophy was charac- et al., 2001) who had reported perceiving her terised by ventricular enlargement, as well as hands to be for 30–60 s after laying out her marked contraction in regions of the parietal and red bed linen. Thus we decided to assess CRO’s occipital lobes, with noticeable sparing of the cere- colour perception and then to administer a subset bellum across all time steps. By the time of CRO’s of the tasks used to investigate the abnormal dura- last visit there was widespread atrophy across tion of VL’s colour after-images. Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010

Figure 1. Registered serial MR images showing a medial sagittal view of CRO’s right hemisphere at Visits 1–4 (ages 58–61 years old). Repeat scans were fluid-registered to the baseline image and colour-coded voxel-compression maps were produced. The scale shows the percentage volume change per voxel (–20 to 20%) with green and blue representing contraction and yellow and red representing expansion. 6 CRUTCH ET AL.

Experiment 1a: Hue discrimination Results The perceived duration of each abnormal col- Methods our is shown in Figure 2, together with the previ- A background test of hue discrimination was ously collected data from VL and 3 age-matched administered to assess colour perception skills. The healthy control participants reported in Chan et al. stimuli were pairs of matte colour chips presented (2001). CRO’s after-images were in the majority of adjacent to one another mounted on white A6 cases notably longer than in controls, and her cards. The colours were selected from the red, colour descriptions tended to be the complimen- green and purple hue ranges of the Munsell colour tary colour of the previously presented colour system, and were of fixed value and chroma (6/6). stimulus. She reported seeing ‘pink’ following all Task difficulty was varied by altering the distance green stimuli, ‘purple’ or ‘lilac’ after all yellow between hue pairs on the Munsell hue scale. There stimuli, ‘turquoise’ after the 15-s red stimulus, and were three levels of difficulty: Easy (8 degrees of ‘yellow’ or ‘sunrise’ after the 15- and 10-s blue separation), Intermediate (4 degrees of separation) stimuli. The only exceptions were red 5 s (‘noth- and Difficult (2 degrees of separation). At each ing’), red 10 s (‘grey’) and blue 5 s (‘pink at the level of difficulty, 8 pairs of colour chips were bottom’). shown in the red, green and purple ranges of the spectrum. In each set of 8 pairs, half the pairs were Comment the same hue and half the pairs were different hues. CRO’s descriptions of abnormal colour per- For all stimuli, CRO was requested to state ceptions were very similar to those of patient whether the colour chips were the same hue or dif- VL, and when tested using the same colour patch ferent hues. stimuli, CRO also showed comparable increases in colour afterimage duration compared with Results healthy control subjects. As with VL, these after- CRO made no errors in the easy condition (8/8 images were observed in the context of mildly in each colour range), occasional errors in the impaired colour perception. In the Discussion, intermediate condition (Red = 8/8; Green = 6/8; we evaluate whether these prolonged afterimages Purple = 6/8) and frequent errors in the difficult reflect a relative sparing of inhibitory inter- condition (Red = 6/8; Green = 5/8; Purple = 5/8). neurons in V1. This corresponds to weak but not grossly impaired colour perception, as healthy controls tend to make occasional errors only on the hard condition. 50 45 VL CRO Experiment 1b: Colour and duration of 40 Controls Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 after-images 35 30 Methods 25 Under white room lighting conditions, the 20 patient was asked to fixate binocularly on red, 15 blue, green or yellow coloured A5 sized cards of 10

equal luminance placed on a neutral light grey Duration of after-images (s) 5 background. The colour cards were presented for 0 durations of 5, 10 or 15 s. After presentation, the test card was removed and the patient was asked to

fixate on the grey background and to describe the Red (5s) Blue (5s) Red (10s) Red (15s) Blue (10s) Blue (15s) Green (5s) Yellow (5s)

duration and colour of any after-images. Each col- Green (10s) Green (15s) Yellow (10s) Yellow (15s) our was presented only once at each presentation duration. CRO reported the colour and duration Figure 2. Duration of CRO perceived colour after-images. Also shown are the mean after-image durations reported by patient VL of any after-images perceived immediately after the and 3 healthy age-matched control participants based on three presentation of the colour stimuli. presentations of each colour stimulus (from Chan et al., 2001). ABNORMAL VISUAL PHENOMENA IN PCA 7

PERCEIVED MOTION OF STATIC STIMULI i. Dot Counting: The stimuli were arrays (N = 28) of 3–9 dots in a randomised spatial arrange- During the standard assessment of CRO’s visuo- ment. Trial 2 arrays had the same spatial spatial processing skills, she was administered arrangement as Trial 1 but were reproduced at several subtests from the VOSP (Warrington & 50% scale. James, 1991). Of particular note was her com- ii. Number Location: Taken from the VOSP ment following completion of the Dot Counting Number Location task (Warrington & James, subtest (which as the name suggests involves 1991), the stimuli consisted of two identically counting 5–9 black dots printed on a white back- sized squares printed one above the other; the ground), that the task was rather unfair because top square contained randomly positioned the dots kept moving around on the page. This Arabic numerals and the bottom square a sin- tallied with her conscious experience of other vis- gle black dot. CRO was requested to identify ual arrays containing small items (e.g., words on which number in the top square occupied the a page) in which the location of items appeared same spatial position within its square as the to alter. Her description particularly of letters dot in the bottom square. For both Trial 1 moving within a word was reminiscent of some (N = 12) and Trial 2 (N = 22), the presence of a forms of developmental dyslexia, in which read- colour overlay was varied in an ABBA design. ing difficulties are associated with unsteady eye fixations and attributed to impairments of the Results visual magnocellular system (Ray, Fowler, & Stein, 2005; Stein, 2003; Wilkins, 2003). In the i. Dot Counting: On Trial 1, CRO was signifi- current section, we report attempts to modify this cantly more accurate counting dots with the perceived motion of static stimuli using colour green tint than without [Tint = 22/28, No tint = filters (Experiment 2a), and an analysis of the 13/28; χ2(1) = 6.17, p = .013]. However, this relationship between eye movement and per- significant effect did not replicate on Trial 2 ceived motion (Experiment 2b). where overall response accuracy was much lower and there was only a small advantage for Experiment 2a: Dot Counting and Number the green tint condition [Tint = 9/28, No tint = χ2 Location with and without a colour tint 5/28; (1) = 1.52, p = .22]. ii. Number Location: On Trial 1, CRO responded One technique used to counter unsteady fixations more accurately with the tint than without, an and poor control in developmental dys- advantage which approached formal levels of lexics is to read through colour filters. Although statistical significance [Tint = 10/12, No tint = χ2 magnocellular cells are not involved in colour per- 6/12; (1) = 3.00, p = .083]. However, this ception, M-ganglion cells receive inhibitory input effect did not replicate as on Trial 2 CRO Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 from long- (L-; red), medium- (M-; green) and showed no difference between the two condi- short- (S-; blue) wavelength sensitive cones tions (Tint = 12/22, No tint = 12/22). (Roorda & Williams, 1999). Different colour filters may act to reduce components of this inhibitory Experiment 2b: Eye movements associated input (e.g., yellow filters reducing S-cone inhibi- with perceived motion of static stimuli tory input), thus increasing magnocellular function and improving dependent functions such as con- Further investigations of CRO’s perceived motion trast and motion sensitivity. Here we piloted the of static stimuli effect were conducted in December use of a similar colour tint as a technique for 2009. Eye movements are of importance in under- reducing CRO’s sense of perceived motion of static standing CRO’s perception of motion of static stimuli. stimuli as eye position signals play a critical role in normal space constancy. The healthy brain Methods achieves a steady perception of the world despite CRO was administered two spatial tasks (Dot changes in the retinal position of visual stimuli by Counting and Number Location) with and without remapping space into co-ordinates based on a colour overlay (transparent green cellophane) on upcoming rather than current fixation point. In two occasions (Trials 1 and 2). Experiment 2b, an eye-tracking system was used to 8 CRUTCH ET AL.

monitor the stability of CRO’s fixation when was acquired only from trial 5 (a 4-item array; see observing single and multiple item arrays. Figure 4). The heat map of gaze position density indicates that on this trial, the correct response Methods may have been provided somewhat by a chance summation, with two dots fixated twice and two CRO was presented with a single 0.8° diameter dots never fixated. black dot on a light grey background under ambi- ent lighting and was asked to fixate and to report Comment whether the dot appeared to be static or moving. CRO was subsequently presented with 8 arrays of There was some limited evidence that, in some 1–4 dots, again with diameter 0.8° and with a min- instances, an intervention which may have imum spacing of 1.6° and a maximum eccentricity increased aspects of magnocellular function of 6°, and was requested to count how many dots improved the accuracy of CRO’s performance on were on the screen. All stimuli were presented at two tests of visuospatial processing, both of which the centre of the display, with onset triggered by involve multiple target stimuli spatially dispersed CRO fixating upon a central fixation cross which within presentation arrays. This poor performance preceded each stimulus. Gaze position was moni- may in part by influenced by impaired numerosity tored online at 1000 Hz, using a frame-mounted judgements. However, measurement of eye move- infrared eye tracker (EyeLink 1000, SR Research ments indicated some form of relationship between Ltd, Canada). The patient’s fixation was unstable CRO’s perception of motion among static items when we attempted to record eye position, making and the occurrence of eye movements. Indeed, fix- it extremely difficult to obtain good calibration of ation instability had a particularly pronounced the eye tracking device. Consequently only a rela- cost when CRO was attempting tasks requiring tively small amount of data was collected. The more than a single point of fixation. However, the patient completed two long trials, of durations: 30 causal relationship between her perception of and 20 s, respectively, in the stable fixation task. motion and eye movements remains to be estab- Calibration was acceptable in Trial 1, but less opti- lished. The perception of motion among static mal in Trial 2. items could be caused directly by involuntary eye movements. Alternatively, rather than being invol- Results untary, the eye movements detected in this experi- ment could reflect the attempt to fixate an CRO’s horizontal and vertical eye positions over incorrectly remapped stimulus, therefore produc- the entire duration of Trial 1 of the single dot fixa- ing a vicious circle of voluntary . This sac- tion task are shown in Figure 3a. A complemen- cadic cycle could be initiated by a spurious, small tary heat map showing the density of gaze amplitude eye movement that might be voluntary positions at each point on the display screen is or involuntary. Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 shown in Figure 3b. There was minimal eye move- ment with no saccades during the period when CRO reported stable perception of the dot. How- INVERSE SIZE EFFECTS ever, numerous shifts in eye position were noted during the period when the patient reported dot From the time of her referral, CRO reported movement. In order to calculate the proportion of increasing difficulties with reading, which initially time when fixation was steady or unstable, the fol- were characterised by losing her place on the page lowing criteria were used: (i) Gaze within 5º from and having difficulty moving from the end of one the centre (i.e., looking at the target); (ii) Gaze line of text to the beginning of the next. Such spa- position did not change more than 0.1º in the last tial dyslexic deficits are reported commonly by 200 ms. Using these criteria to define ‘steady fixa- individuals with PCA. However, in 2005, CRO tion’, gaze was stable on only 20.1% (Trial 1) or made two comments which suggested that her 11.8% (Trial 2) of the time. On the arrays of 1–4 problems with reading sometimes related to the dots, CRO was able to count the items correctly on size of the text. First, she mentioned that she had only 5/8 occasions. The single item display was the particular difficulty reading the headlines in news- only ‘array’ which CRO enumerated correctly on papers. Second, she also described how, in an both the occasions it was presented. Unfortu- effort to counteract her reading problems, she had nately, due to inadequate calibration, reliable data gone to her local public library to borrow large ABNORMAL VISUAL PHENOMENA IN PCA 9

A

B

Figure 3. (A) Horizontal and vertical eye positions are shown for the entire duration of Trial 1 of the single dot fixation task. The green Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 segment indicates the period during which the patient reported stable perception, and corresponds with fixation. The red segment corresponds to the reported perception of movement. (B) A heat map showing the density of gaze positions for each point on the screen.

print books, designed for the ‘visually impaired’. were presented at 3 different font sizes: large (1000; Ironically, she found these texts infinitely more dif- letter height 20.4°), medium (250; 5.0°) and small ficult to read than normal size print. These obser- (80; 1.6°). The stimuli were viewed on a Dell vations led us to evaluate the effect of size upon her Inspiron 510m laptop computer from a viewing processing of letters, numbers, words, and a selec- distance of 50 cm for 150 ms with no fixation or tion of stimuli designed for testing aspects of early mask using Superlab Pro software. The large, visual processing. medium and small stimuli were administered in blocks of 8 items using a Latin square design. CRO was asked to name each letter. Experiment 3a: Letter size

Methods Results The stimuli were 24 uppercase letters (all alpha- The percentage correct responses are shown in betic items except I and O) in Arial font. All stimuli Figure 5. A Cochran Q-test of differences between 10 CRUTCH ET AL.

Figure 4. A heat map showing the density of gaze positions for each point on the screen. Xs denote dot positions, and the green dots are refixations on previously visited locations (within <1° of a previously fixated location, when the immediately previous fixation was >2.5 away).

100 90 80 70 60 50 40 30 20 Percentage correct (%) 10 0 Small (font 80) Small (font 60) Small (font 80) Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 Large (font 150) Large (font 500) Large (font 1000) Medium (font 250) Medium (font 250)

Letters Arabic numerals Words

Figure 5. Percentage correct responses for the identification of large, medium and small uppercase letters, Arabic numerals, and words.

conditions demonstrated a significant trend Arabic numerals (1–9) in Arial font. All stimuli toward an inverse size effect, with response accu- were presented twice at 3 different font sizes: large racy significantly higher with smaller stimuli (Q = (500; 10.3°), medium (250; 5.0°) and small (80; 43.5, p < .001, df = 2). 1.6°). The stimuli were all viewed for 150 ms using the same apparatus and design as in the letter experiment. Experiment 3b: Number size Results Methods The percentage numbers named correctly are A comparable test of the effect of font size upon shown in Figure 5. Consistent with her perform- number identification was conducted, using 9 single ance in Experiment 3a, CRO named smaller items ABNORMAL VISUAL PHENOMENA IN PCA 11

more accurately than larger items, an inverse size 2009, using an eye-tracking system to assess the effect which was shown to be statistically signific- fixation pattern associated with observing two ant (Q = 25.1, p < .001, df = 2). example words from Experiment 3c presented at different font sizes. Experiment 3c: Word size Methods Methods The word ‘NATURAL’ was presented twice, at different font sizes: 12 (letter height 1°) and 60 The effect of font size upon word reading accu- (2.12°). The word ‘KING’ was also presented racy was assessed using 16 four-letter words and 16 twice, at font sizes 12 (1°) and 150 (5.29°). Both eight-letter words shown in uppercase Arial font. words were presented at the centre of the screen, The 4- and 8-letter words were matched for triggered by CRO’s fixation upon a preceding cent- CELEX frequency but the 8-letter words naturally ral fixation cross. CRO was requested to name had fewer orthographic neighbours than the 4-letter each item. Owing to calibration difficulties, only a words (4-letter words: mean CELEX = 168.4, SD limited number of stimuli were presented; the word CELEX = 195.3; Mean N = 8.94); 8-letter words: stimuli ‘NATURAL’ and ‘KING’ were administered mean CELEX = 153.8, SD = 96.3; Mean N = interspersed with 9 other letter, number and word 0.38). All words were presented at 2 different font stimuli, the data for which are not presented here. sizes: large (150; letter height 3.0°) and small (60; 1.2°). The words were presented for 150 ms in an Results overall ABBA design using the same apparatus as in the letter and number experiments. CRO read both target words correctly in the font 12 condition, but was unable to identify the Results word or even any constituent letters in the larger font conditions (e.g., NATURAL: ‘a lot of letters, The percentage correct responses for large and I can’t make them out’). The density of eye posi- small font size words are shown in Figure 5. A tions at each location on the screen when attempt- more detailed breakdown of the number of correct ing to read the word NATURAL at both small and responses is shown in Table 2. CRO read small large font sizes are shown in Figure 6. These font words significantly more accurately than large example figures demonstrate a relatively stable fix- font words (Sign test: N = 15, x = 0, p < .01). She ation point near the initial letter for the success- was also significantly more accurate reading 8- fully read small font item, but a much more than 4-letter words, χ2(1) = 4.02, p = .045. The spatially distributed array of shorter duration errors made by CRO were either visual errors (N = fixations for the larger font presentation. 13; e.g., date → skate, gray → gravy), omissions (N = 11) or miscellaneous (N = 3; e.g., trouble → village). Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 Comment CRO’s clinical descriptions of having greater Experiment 3d: Eye positions associated with difficulty reading large sized print were confirmed inverse size effects empirically. She was able to identify single letters, Arabic numerals and single words significantly Further investigations of CRO’s better reading of more accurately when printed in smaller fonts. small than large letters were conducted in December CRO’s inability to perceive larger items was associ- ated with multiple fixations in and around the

TABLE 2 stimulus area. These observations suggest a reduc- Pilot data from CRO showing the number of large and small tion in receptive field size (and consequently a font 4- and 8-letter words read correctly (Experiment 3c) reduction in the field of effective vision), with eye movements partly reflecting an unsuccessful search Large Small for a fixation point where sufficient perceptual (font size 150) (font size 60) TOTAL information can be obtained to permit accurate 4-letter words 2/16 11/16 13/32 identification of target features. 8-letter words 9/16 15/16 21/32 It was also noted that she read long words more TOTAL 11/32 26/32 accurately than short words, a fact which is likely 12 CRUTCH ET AL.

A 15 900 800 10 700

5 600 500 0 400 –5 300

200 –10 100 –15 0 –25 –20 –15 –10 –5 0 5 10 15 20 25

B 15 60 10 50

5 40

0 30

–5 20

–10 10

–15 –25 –20 –15 –10 –5 0 5 10 15 20 25

Figure 6. Example heat maps showing density of gaze positions at each location on the screen for the same word presented at (A) small font size (and consequently correctly read) and (B) large font size (and consequently incorrectly read).

to reflect that multiple factors other than the physical It was a couple of, two or three or four weeks ago, Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 size (length) of word stimuli affect the likelihood of and it was very odd, and I was in the kitchen or accurate identification. For example, the 4-letter something and I think probably I’d come downstairs words had far higher orthographic neighbourhood or something. Anyway, when I got downstairs the ratings (the number of alternative words which can whole of the room was upside down, which was actually very scary but I got over that when I real- be formed by changing only one letter in the target), ised it was OK if you know what I mean. But it meaning there was a much larger neighbourhood was completely, the bottom was on the top and of competitor words with similar orthography the top was on the bottom. Do you see what I which might be mis-selected to produce visual mean? It was most peculiar. [Had that ever hap- reading errors. pened to you before?] No. It was quite scary actu- ally, and it was really quite difficult to orientate myself because I wasn’t quite sure whether I really ROOM TILT ILLUSION was upside down, you know, silly things can hap- pen can’t they, or whether it was just the image in One recent experience which CRO reported at the my brain or whatever else had turned me inside beginning of 2009 was of a transient episode of out basically. [And what did you do when that room tilt illusion. Transcribed from an audio happened to you?] Erm, sat down, and well I did a recording [with investigator questions in brackets], bit of blinking and things like that as you would, CRO described this unusual event as follows: and nothing really happened and then it just sort ABNORMAL VISUAL PHENOMENA IN PCA 13

of, I suppose I must have stood up or done some- saccadic eye movements, the target was first moved thing different or something, and it just was OK. 30º to the right, then 30º to the left, and the patient . . . [And how long do you think it lasted for?] was asked to follow this with her eyes. Pursuit Probably three or four minutes, maybe a bit longer, movements were tested by moving the target hori- but not much longer than that. [And when it zontally (sinusoidally) from left to right, and the stopped happening, did it just flip back suddenly?] patient was again asked to follow the movement Yes, it flipped back. I wasn’t sort of half and half. with her eyes. The chair was kept stationary during these conditions. To test peripheral vestibular Of relevance to the report of room tilt illusion was function, the patient was rotated at 90°/s in the CRO’s history of benign paroxysmal positional ver- dark, first to the right, then to the left and eye tigo (BPPV). However, CRO was very clear that this movements were again recorded., Vestibular episode was quite distinct from any previous experi- responses to chair rotation were symmetrical and ences of vertigo, which occur following rapid of normal duration suggesting intact vestibular changes in posture. ‘If I get up too suddenly or down function. A bithermal caloric showed no evidence too suddenly my head can spin, but it’s not that, this of a canal paresis and no asymmetry of vestibular was entirely different’. Nonetheless, to consider the function. The patient also underwent assessment of basis for CRO’s room tilt illusion further, a detailed the subjective visual vertical (SVV). Whilst seated neuro-otological assessment was conducted. 60 cm away from a laser target (rod-shaped) pro- jected onto the wall, in an otherwise dark room, Neuro-otological assessment she was asked to align the laser rod with the per- ceived gravitational vertical. While the patient’s Bedside eye movement examination was carried eyes were shut, the experimenter offset the rod out using a large 30-cm diameter black circle as a (between 30° and 60° to either side). Six trials were target fixation. There was a full range of eye move- recorded. There was a mean leftward offset of 2.5° ments in vertical and horizontal planes, with no which is within the normal range. spontaneous or gaze-evoked nystagmus. Pursuit movements were normal horizontally, with occa- Comment sional saccadic intrusions. Vertical pursuit was This episode of apparent room tilt illusion was broken but within the normal age range. Saccadic clearly and coherently described, qualitatively dif- eye movements appeared normal, although a pre- ferent to previous episodes of BPPV, and occurred cise assessment was limited due to the patient’s dif- in the context of severe visual dysfunction. To ficulty with target fixation. The vestibulo-ocular summarise the neuro-otological examination, a reflex (VOR) – assessed using the head impulse test positioning (Hallpike) manoeuvre revealed nystag- (ref: Halmagyi GM, Curthoys IS, Arch Neurol mus consistent with posterior canal BPPV but this 1988) – was intact bilaterally, implying normal ves- manoeuvre did not trigger a room tilt illusion, and Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 tibular nerve function. There was no skew devia- clinical laboratory investigations showed normal tion of the eyes and no head tilt. Given the history and symmetrical vestibular function. We are una- suggestive of benign positional vertigo, a position- ware of any previous reports of room tilt illusion ing (Hallpike) manoeuvre was performed. This linked to cortical degenerative disease. As the elicited nystagmus consistent with right sided pos- room tilt illusion has occurred only once to date terior canal BPPV. Importantly, the Hallpike and not in the presence of the investigator, direct manoeuvre did not trigger a room tilt illusion. examination and questioning of the participant whilst experiencing the phenomenon has of course Clinical investigations not been possible. In the Discussion, consideration is given to whether this occurrence of room tilt illu- Electronystagmography (ENG) was limited due sion reflects a discrepancy between visual, vestibu- to the patient’s difficulty with small target fixation lar and proprioceptive spatial maps of verticality. and roaming eye movements. Whilst seated on a motorised chair, the patient was asked to fixate on a computer-controlled coloured target 60 cm away. DISCUSSION The patient’s head was secured using a padded chin rest, and eye movements were recorded with Posterior cortical atrophy (PCA), the degenerative surface electrodes placed around the eyes. To test condition most commonly considered to be the 14 CRUTCH ET AL.

‘visual’ or ‘biparietal’ variant of Alzheimer’s dis- associated with activity in V8 just anterior to V4 ease, is attracting increased clinical recognition and (Hadjikhani, Liu, Dale, Cavanagh, & Tootell, research interest. In this study, we describe a single 1998). As Chan et al. (2001) argued one possible individual with PCA who reported a number of mechanism for the abnormal elongation of these abnormal visual phenomena, and our preliminary after images is excessive rebound inhibition of pre- attempts to quantify and examine each of these viously excited wavelength selective neurons in V1. symptoms. In our clinical experience, some of these This may occur as a consequence of the relative phenomena are experienced commonly by other sparing of inhibitory interneurons in V1 in the con- PCA patients (perceived motion of static stimuli, text of the degeneration of excitatory neurons that and colour after image effects) whilst others have occurs in Alzheimer’s disease. It is an open ques- not been reported previously in this clinical popu- tion whether only some or all brain regions within lation (room tilt illusion). Here we summarise the normal colour processing system remain acti- briefly the characteristics of each of these phenom- vated for an abnormal period following the presen- ena and consider potential aetiologies for these tation of colour stimuli in such PCA patients. symptoms. Perceived motion of static stimuli Prolonged colour after-images The perceived motion described by CRO when The first unusual visual phenomenon examined viewing an array of static stimuli in the absence of was that of prolonged colour after-images, which any moving stimuli can be distinguished from a are reported commonly in the context of PCA. number of normally occurring motion phenomena Sometimes these colour abnormalities take the and after effects such as ‘induced motion’, the form of non-specific patches of colour in the visual impression that a static stimulus moves relative to periphery (as with CRO in the earliest sympto- the head in the context of other stimuli which actu- matic phases of the disease). In other situations, ally are moving (e.g., Nefs & Harris, 2008). One and possibly later in the disease course, these possibility is that this sense that static stimuli were abnormal colours are experienced more centrally moving occurred as a result of disordered eye fixa- in the visual field, and can often be understood as tions and movements. being complementary in colour to other prominent A variety of visual, proprioceptive and vestibu- stimuli in the current or recent visual environment. lar cues act to determine the amplitude and direc- For example, another patient with PCA (Patient 2; tion of eye movements required to maintain Crutch & Warrington, 2007) reported having to stability of the image on the fovea (i.e., space con- ask her partner why the England football team stancy e.g., Hess & Angelaki, 2003; Wurtz, 2008). were playing in a pink strip; this perception of pink Indeed, eye movements play an important role in is likely to have reflected a concurrent and persist- normal space constancy, with information about Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 ent pink/red afterimage overlaying the neutral the intention to make a saccadic eye movement white England shirts owing to the background used to compensate for the movement by remap- green stimulus of the football pitch grass. ping spatial representations upon the co-ordinates After-images of abnormal duration have previ- of the intended rather than current eye position ously been quantified in a patient with PCA (VL; (e.g., Duhamel, Colby, & Goldberg, 1992; Lappe, Chan et al., 2001). Tested using the same stimuli as Awater, & Krekelberg, 2000). In this way, the per- VL, CRO also showed considerably longer duration ceived motion of static stimuli could reflect prob- of colour afterimages following exposure to various lems with one or more of these constituent colour stimuli. Although the durations of after processes (e.g., visual reafferance [the visual images was only measured on one occasion for each motion signal generated by eye movements], eye colour at each exposure duration (as opposed to muscle proprioception, corollary discharge of eye averaging across three presentations under each con- movement commands). For example, gaze stability dition as in Chan et al., 2001), CRO gave clear self- could be affected because inaccurate vestibular sig- reports and the colour of each afterimage tended to nals are stimulating eye movements away from the be a complementary colour of the stimulus. current fixation point, or because of a deficit in eye Functional neuroimaging studies have reported movement planning and control in the context of that colour afterimages in healthy individuals are accurate vestibular signals. On either account, it ABNORMAL VISUAL PHENOMENA IN PCA 15

would be predicted that the perceived motion of patient NJ whose clinical picture was compatible static stimuli should not be observed when fixation with PCA and who was more accurate in identify- is held in one location in the absence of head or eye ing small than large pictures, words and single let- movements. ters. Most of these reports have interpreted reverse Eye-tracking data in the current study confirm size effects or better reading of small than large let- this prediction: when looking at a single dot, CRO ters in terms of a deficit in selective attention or only reported the position of the dot as static when slightly more specifically a restricted ‘spotlight’ of there were no concurrent horizontal or vertical attention such that visual features cannot be inte- changes in eye position. The instability of fixation grated over as large an area as in healthy controls in CRO is also consistent with FDG-PET evidence (e.g., Stark et al., 1997; cf. Treisman & Gelade, showing reduced glucose metabolism in patients 1980). with PCA not only in posterior brain regions but A related formulation is that PCA patients show also in the bilaterally (Nestor, a reduction in their ‘effective visual field’, consist- Caine, Fryer, Clarke, & Hodges, 2003). It was sug- ent with data from healthy individuals and patients gested that this hypometabolism might be second- with right temporo-parietal damage that attention ary to a loss of input from occipital and parietal dynamically modulates the detection of stimuli in regions and might account for the ocular apraxia the periphery (Russell, Malhotra, & Husain, 2004). seen in some patients with PCA. In addition, better reading of small than large let- The attempts described in the current paper to ters may be exacerbated by problems with direct- ameliorate the perceived motion of static stimuli ing and detecting eye movements (see the using a coloured filter were only provisional in explanation of perceived motion of static stimuli nature, and were based on the notion of magnocel- phenomena above). Larger stimuli may require lular-upregulation (e.g., Stein, 2003). However, or stimulate shifts in fixation to appreciate key CRO’s improved spatial perception under colour features, with disordered eye movements and com- conditions in two tasks represents some evidence promised oculomotor feedback leading to the that should motivate further detailed studies of the impaired integration of featural information across effects of colour filters upon gaze stability and per- the visual field. ceived motion of static stimuli in PCA patients. Room tilt illusion Better reading of small than large letters We finish by considering the most striking and Turning to the reading of small and large letters, unusual visual experience described by CRO, CRO demonstrated a reliable and significant namely room tilt illusion. The perception of verti- advantage for identifying smaller letters, Arabic cality relies upon the integration of visual, vestibu- numerals and words. These experimental findings lar and proprioceptive afferent stimuli, with Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 concur with CRO’s reported difficulties with read- multimodal cells in posterior parietal cortex heav- ing headlines and large print ‘easy reading’ books. ily implicated in the integration process (e.g., The data are also compatible with the anecdotal Andersen, Essick, & Siegel, 1985; Brandt, 1997; case of an individual with PCA who described to Sakata & Kusunoki, 1992). A discrepancy between his clinician sitting in a train carriage and being these three modalities can create an erroneous illu- unable to read the headlines in his own newspaper sion of movement or altered spatial position (Malis but being able to perceive the headlines in a copy & Guyot, 2003). of the same newspaper being read by another The complete 180° inversion of the visual world passenger seated further down the carriage (Fox, was first reported by Bishop in 1805 and attributed personal communication). to hysteria (Bishop, 1805). Room tilt illusion (also The tendency for patients with progressive visual sometimes referred to as reversal of vision meta- disturbance to have greater difficulty recognising morphopsia) involves a transient 180° or 90° rota- large as compared with small pictures and other tion in the visual field in the coronal plane stimuli has been described on several occasions (Hernandez et al., 2006), although sagittal rotations (e.g., Pick, 1908 cited in Bender & Feldman, 1972; have been reported (Tanaka, Takeda, & Hamanaka, Coslett et al., 1995; Saffran et al., 1990; Stark et al., 1996). This phenomenon has been reported follow- 1997). For example, Stark et al. (1997) reported ing damage to the brain stem, cerebellum or other 16 CRUTCH ET AL.

cortical lesions (River, Ben Hur, & Steiner, 1998; of visuo-vestibular processing. As this symptom Steiner, Shahin, & Melamed, 1987) but most com- has occurred only once in this patient, it is too monly after vertebro-basilar ischaemia. A review early to label room tilt illusion as a common fea- of 18,000 patients presenting to a neuro-otology ture of PCA. However, this finding, together with unit found 23 cases of room tilt illusion following clinical complaints of balance problems, giddiness peripheral vestibular disorders. In three cases, the and unsteadiness from other PCA patients seen in phenomenon was associated with BPPV, and our clinical service, suggest vestibular function should occurred during the diagnostic Hallpike manoeu- be examined in detail in this patient population. vre in all cases (Malis & Guyot, 2003). Hayashi, In conclusion, all of the deficits of visual cogni- Yamaguchi, Katsumata, and Mimura (2009) tion which individuals with PCA describe can be reported room tilt illusion of 90° in a patient with accurately characterised as ‘negative’ in the sense Parkinson’s disease associated with other visual that they contribute to a profound and progressive hallucinations. They hypothesised that this functional impairment which reduces or prevents resulted from hypoperfusion in the posterior performance and enjoyment of numerous pastimes region of the right intraparietal sulcus causing a and daily activities. However, some symptoms mismatch between visual and vestibular inputs. could be construed as ‘positive cognitive phenom- However, we are unaware of any previous ena’ in that the individual does not merely fail to reports of room tilt illusion in individuals with a perceive a stimulus property which is present, but cortical degenerative disease. We would suggest has a ‘positive’ perception of stimulus attributes that in the current case, this phenomenon is likely which are not present. One certainty is that the to reflect a discrepancy between visual, vestibular abnormal visual phenomena reported in the cur- and proprioceptive spatial maps of verticality. rent paper constitute but a few of the unusual and Defective visual processing is strongly implicated poorly understood symptoms experienced by indi- as this episode was observed in the context of pos- viduals with PCA. It is anticipated that further terior cortical atrophy, was never experienced prior detailed investigation of the basis of such phenom- to the PCA diagnosis, and was experienced subjec- ena, and systematic evaluation of their prevalence tively as a completely different phenomenon to in a prospective cohort of PCA patients, may con- previous episodes of BPPV. In addition, despite tribute to our elucidation of the disease process in demonstrating active posterior canal-BPPV, the PCA and the development of aids and strategies to clinical examination and vestibular investigations, reduce the real-life cost of visual dysfunction for showed normal peripheral vestibular function, and these individuals. excluded peripheral and central otolithic dysfunc- tion. More importantly, positioning manoeuvres ACKNOWLEDGEMENTS did not trigger the room tilt illusion in our patient. Indeed, if this illusion were caused by BPPV, one We would like to thank CRO for giving so gener- Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 would have expected the room tilt to have occurred ously of her time. We are grateful to Professor more than once, given how active the positional Martin Rossor for allowing us to work with a vertigo was when tested with provocation manoeu- patient under his care, and for helpful discussions vres (Hallpike). It is therefore more likely that the concerning reversal of vision metamorphopsia. We phenomenon occurred as a result of impaired are also indebted to Merle James for making neu- visuo-vestibular processing at the cortical level, ropsychological test scores available to us. This where visual and vestibular inputs were perhaps work was undertaken at UCLH/UCL who transiently aberrantly encoded. CRO also showed received a proportion of funding from the Depart- no evidence of vertebro-basilar infarction. The ment of Health’s NIHR Biomedical Research room tilt illusion may have occurred only tran- Centres funding scheme. The Dementia siently because subsequent additional visual ana- Research Centre is an Alzheimer’s Research Trust lysis of the scene improved the visual input signal Co-ordinating Centre. SC is supported by an Alzhe- which was then more successfully and adaptively imer’s Research Trust Research Fellowship. ML is integrated with online vestibular and propriocep- supported by an Alzheimer’s Society Studentship. tive information. The reason why the illusion has occurred only once remain unclear, but clinical Original manuscript received 29 May 2009 experience suggests an number of PCA patients Revised manuscript accepted 8 June 2010 experience more persistent and regular disruptions First published online ABNORMAL VISUAL PHENOMENA IN PCA 17

REFERENCES Hayashi, R., Yamaguchi, S., Katsumata,Y., & Mimura, M. (2009). Room tilt illusion in Parkinson’s disease: Loss of spatial reference frames? Journal of the Neurologi- Andersen, R. A., Essick, G. K., & Siegel, R. M. (1985). cal Sciences, 287, 264–266. Encoding of spatial location by posterior parietal Hernandez, A. H., Pujadas, F., Purroy, F., Delgado, P., neurons. Science, 230, 456–458. Huertas, R., & Alvarez-Sabin, J. (2006). Upside down Baxter, D. M., & Warrington, E. K. (1994). Measuring reversal of vision due to an isolated acute cerebellar dysgraphia: A graded-difficulty spelling test. Behav- ischemic infarction. Journal of Neurology, 253, 953–954. ioural Neurology, 7, 107–116. Hess, B. J. M., & Angelaki, D. E. (2003). Vestibular con- Bender, M. B., & Feldman, M. (1972). So-called visual tributions to gaze stability during transient forward agnosias. Brain, 95, 173–186. and backward motion. Journal of Neurophysiology, Benson, F., Davis, J., & Snyder, B. D. (1988). Posterior 90, 1996–2004. cortical atrophy. Archives of Neurology, 45, 789–793. Hof, P. R., Vogt, B. A., Bouras, C., & Morrison, J. H. Brandt, T. (1997). Cortical matching of visual and ves- (1997). Atypical form of Alzheimer’s disease with tibular 3D coordinate maps. Annals of Neurology, 42, prominent posterior cortical atrophy: A review of 983–984. lesion distribution and circuit disconnection in corti- Burgess, N. (2008). Spatial cognition and the brain. Year cal visual pathways. Vision Research, 37, 3609–3625. in Cognitive Neuroscience 2008, 1124, 77–97. Jackson, M., & Warrington, E. K. (1986). Arithmetic Caine, D. (2004). Posterior cortical atrophy: A review of skills in patients with unilateral cerebral lesions. the literature. Neurocase, 10, 382–385. Cortex, 22, 611–620. Chan, D., Crutch, S. J., & Warrington, E. K. (2001). A Kartsounis, L. D., & Warrington, E. K. (1991). Failure disorder of colour perception associated with abnor- of object recognition due to a breakdown of figure mal colour after-images: A defect of the primary vis- ground discrimination in a patient with normal ual cortex. Journal of Neurology Neurosurgery and acuity. Neuropsychologia, 29, 969–980. Psychiatry, 71, 515–517. Lappe, M., Awater, H., & Krekelberg, B. (2000). Post- Christensen, G. E., Rabbitt, R. D., & Miller, M. I. saccadic visual references generate presaccadic com- (1996). Deformable templates using large deforma- pression of space. Nature, 403, 892–895. tion kinematics. IEEE Transactions on Image Levine, D. N., Lee, J. M., & Fisher, C. M. (1993). The Processing, 5, 1435–1447. visual variant of Alzheimer’s disease: A clinicopatho- Cogan, D. G. (1985). Visual disturbances with focal pro- logic case study. Neurology, 43, 305–313. gressive dementing disease. American Journal of Oph- Lewis, E. B., & Fox, N. C. (2004). Correction of differ- thalmology, 100, 68–72. ential intensity inhomogeneity in longitudinal MR Crutch, S. J., & Warrington, E. K. (2007). Foveal crowd- images. Neuroimage, 23, 75–83. ing in posterior cortical atrophy: A specific early- Malis, D. D., & Guyot, J. P. (2003). Room tilt illusion as visual-processing deficit affecting word reading. a manifestation of peripheral vestibular disorders. Cognitive Neuropsychology, 24, 843–866. Annals of Otology, Rhinology and Laryngology, 112, Coslett, H. B., Stark, M., Rajaram, S., & Saffran, E. M. 600–605. (1995). Narrowing the spotlight: A visual attentional McMonagle, P., Deering, F., Berliner, Y., & Kertesz, A. disorder in Alzheimer’s disease, Neurocase, l, 305–318. (2006). The cognitive profile of posterior cortical De Renzi, E. (1986). Slowly progressive or atrophy. Neurology, 66, 331–338. apraxia without dementia. Cortex, 22, 171–180. Mendez, M. F., Ghajarania, M., & Perryman, K. M. Duhamel, J.-R., Colby, C., & Goldberg, M. E. (1992). (2002). Posterior cortical atrophy: Clinical characteris- The updating of the representation of visual space in Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010 tics and differences compared to Alzheimer’s disease. parietal cortex by intended eye movements. Science, Dementia and Geriatric Cognitive Disorders, 14, 33–40. 255, 90–92. Nefs, H. T., & Harris, J. M. (2008). Induced motion in Freeborough, P. A., & Fox, N. C. (1998). Modeling depth and the effects of vergence eye movements. brain deformations in Alzheimer disease by fluid Journal of Vision, 8, 1–16. registration of serial 3D MR images. Journal of Com- Nestor, P. J., Caine, D., Fryer, T. D., Clarke, J., & puter Assisted Tomography, 22, 838–843. Hodges, J. R. (2003). The topography of metabolic def- Freedman, L., Selchen, D. H., Black, S. E., Kaplan, R., icits in posterior cortical atrophy (the visual variant of Garnett, E. S., & Nahmias, C. (1991). Posterior corti- Alzheimer’s disease) with FDG-PET. Journal of Neu- cal dementia with alexia: neurobehavioural, MRI, rology Neurosurgery and Psychiatry, 74, 1521–1529. and PET findings. Journal of Neurology Neurosurgery Ray, N. J., Fowler, S., & Stein, J. F. (2005). Yellow fil- and Psychiatry, 54, 443–448. ters can improve magnocellular function: Motion Galton, C. J., Patterson, K., Xuereb, J. H., & Hodges, sensitivity, convergence, , and read- J. R. (2000). Atypical and typical presentations of ing. Clinical and Basic Oculomotor Research: in Alzheimer’s disease: A clinical, neuropsychological, Honor of David S Zee, 1039, 283–293. neuroimaging and pathological study of 13 cases. River, Y., Ben Hur, T., & Steiner, I. (1998). Reversal of Brain, 123, 484–498. vision metamorphopsia: Clinical and anatomical Hadjikhani, N., Liu, A. K., Dale, A. M., Cavanagh, P., characteristics. Archives of Neurology, 55, 1362–1368. & Tootell, R. B. H. (1998). Retinotopy and sen- Rogelet, P., Delafosse, A., & Destee, A. (1996). Posterior sitivity in human visual cortical area V8. Nature Neu- cortical atrophy: Unusual feature of Alzheimer’s dis- roscience, 1, 235–241. ease. Neurocase, 2, 495–501. 18 CRUTCH ET AL.

Roorda, A., & Williams, D. R. (1999). The arrangement Lendon, C. L., Richardson, A. M., Varma, A., of the three cone classes in the living . Neary, D., & Mann, D. M. A. (2007). Cognitive Nature, 397, 520–522. phenotypes in Alzheimer’s disease and genetic risk. Ross, S. J., Graham, N., Stuart Green, L., Prins, M., Cortex, 43, 835–845. Xuereb, J., Patterson, K., & Hodges, J. R. (1996). Stark, M. E., Grafman, J., & Fertig, E. (1997). A Progressive biparietal atrophy: an atypical presenta- restricted ‘spotlight’ of attention in visual object rec- tion of Alzheimer’s disease. Journal of Neurology ognition. Neuropsychologia, 35, 1233–1249. Neurosurgery and Psychiatry, 61, 388–395. Stein, J. (2003). Visual motion sensitivity and reading. Russell, C., Malhotra, P., & Husain, M. (2004). Atten- Neuropsychologia, 41, 1785–1793. tion modulates the visual field in healthy observers Steiner, I., Shahin, R., & Melamed, E. (1987). and parietal patients. Neuroreport, 15, 2189–2193. Acute upside down reversal of vision in tran- Saffran, E., Fitzpatrick-DeSalme, E., & Coslett, H. B. (1990). sient vertebrobasilar ischemia. Neurology, 37, Visual disturbances in dementia. In M. F. Schwartz 1685–1686. (Ed.), Modular deficits in Alzheimer-type dementia (pp. Tanaka, H., Takeda, A., & Hamanaka, T. (1996). Visual 297–327). Cambridge, MA: MIT Press. illusion with 90-degree forward rotation on the sagit- Sakata, H., & Kusunoki, M. (1992). Organization of tal plane due to acute paramedian pontine infarction. space perception: neural representation of three- Rinsho Shinkeigaku, 36, 34–37. dimensional space in the posterior parietal cortex. Tang-Wai, D. F., Graff-Radford, N. R., Boeve, B. F., Current Opinion in Neurobiology, 2, 170–174. Dickson, D. W., Parisi, J. E., Crook, R., Caselli, Schott, J. M., Ridha, B. H., Crutch, S. J., Healy, D. G., R. J., Knopman, D. S., & Petersen, R. C. (2004). Uphill, J. B., Warrington, E. K., Rossor, M. N., & Clinical, genetic, and neuropathologic characteris- Fox, N. C. (2006). Apolipoprotein E genotype modi- tics of posterior cortical atrophy. Neurology, 63, fies the phenotype of Alzheimer disease. Archives of 1168–1174. Neurology, 63, 155–156. Warrington, E. K., & James, M. (1991). A new test of Sled, J. G., Zijdenbos, A. P., & Evans, A. C. (1998). A object decision – 2D silhouettes featuring a minimal nonparametric method for automatic correction of view. Cortex, 27, 377–383. intensity nonuniformity in MRI data. IEEE Transac- Wilkins A. J. (2003). Reading through colour. Chichester, tions on Medical Imaging 17, 87–97. UK: Wiley. Snowden, J. S., Stopford, C. L., Julien, C. L., Thompson, Wurtz, R. H. (2008). Neuronal mechanisms of visual J. C., Davidson, Y., Gibbons, L., Pritchard, A., stability. Vision Research, 48, 2070–2089. Downloaded By: [Institute of Neurology] At: 10:46 9 September 2010