This scientific commentary refers to ‘Abnormalities of fixation, saccade and pursuit in posterior cortical atrophy’, by Shakespeare et al. (doi:10.1093/brain/awv103).
Posterior cortical atrophy (PCA) is a progressive degenerative condition characterized by a gradual loss of visual skills and other posterior cortical functions due to atrophy of parietal, occipital and occipito-temporal brain regions (Lehmann et al., 2011). It is associated with a relative preservation of memory, language skills and judgement until late in the clinical course. Core features of PCA include partial or complete Bálint’s syndrome (simultanagnosia, optic ataxia and ocular apraxia) and Gerstmann’s syndrome (acalculia, agraphia, finger agnosia), along with various combinations of visuo-spatial and visuo-perceptual impairments (Crutch et al., 2012), apparent despite an otherwise normal ophthalmological examination. In this issue of Brain, Shakespeare et al. (2015) compare and contrast eye movement abnormalities revealed by basic oculomotor tests in patients with PCA, typical Alzheimer’s disease and healthy controls, in order to determine the extent to which abnormalities in basic (lower-order) oculomotor function contribute to visuo-perceptual disturbances in PCA (Shakespeare et al., 2015).
Patients with PCA typically present when they are between 50 and 65 years of age (Crutch et al., 2012). One of the most important manifestations of PCA is simultanagnosia, an inability to synthesize the overall meaning of a visual scene despite being able to identify its individual elements; this should be suspected when a patient presents with an inability to read pseudo-isochromatic plates, despite intact colour vision (Beh et al., 2015). The relative rarity of PCA makes it difficult to recruit large cohorts, but recently the number of studies of this disorder, often referred to as the cardinal visual dementia and the most common atypical Alzheimer’s disease phenotype, has been increasing.
The study of eye movements in neurodegenerative diseases has also become increasingly common in recent years. Such studies aim to provide biomarkers for early disease, in some cases preclinical, to assist in the differential diagnosis of phenotypically similar syndromes and to monitor the effects of therapy. Many studies have concentrated on saccades, which are relatively simple to record and analyse, and can be tested with increasingly complex paradigms used to evaluate cognitive abilities. A number of different abnormalities have been found in Alzheimer’s disease, including hypometric saccades, prolonged saccade latencies, reduced peak velocities and disorganized visual scanning (Mosimann et al., 2005). However, the results of studies looking at the saccadic velocity and gain (saccade amplitude divided by target amplitude) in Alzheimer’s disease are mixed: some have detected impairments (Shafiq-Antonacci et al., 2003) whereas others have not (Garbutt et al., 2008). Despite these discrepancies, two consistent impairments of saccades have emerged from Alzheimer’s disease oculomotor research; first, a high frequency of saccadic intrusions during attempted fixation usually in the form of square wave jerks, confirmed in the study by Shakespeare et al., and second, visual capture by the target in the anti-saccade paradigm (Garbutt et al., 2008), in which the subject has to suppress a reflexive saccade to a peripheral target and execute an endogenously driven saccade to an equal and opposite location.
Smooth pursuit eye movements are also usually abnormal in Alzheimer’s disease, with an increased frequency of saccades during pursuit resulting from a reduced gain (eye velocity divided by target velocity) in the pursuit system (Fletcher and Sharpe, 1988). However, in addition, large-amplitude saccadic intrusions in the direction of target motion are also observed, probably reflecting increased saccadic distractibility.
Shakespeare and colleagues have used a battery of simple oculomotor paradigms to examine fixation stability, saccade generation and smooth pursuit eye movements. The ability of patients to disengage attention and generate targets for subsequent eye movements was tested using a saccadic gap/overlap paradigm. The authors also looked at the relationship between oculomotor metrics and performance on tests of basic visual function and higher order object and space perception.
In the fixation stability tasks, Shakespeare et al. report a significantly higher frequency of square wave jerks in the typical Alzheimer’s disease group than in the PCA group, with the latter no different from healthy controls. However, patients with PCA and those with typical Alzheimer’s disease both showed a greater frequency of large intrusive saccades and shorter fixation times than the control group.
Analysis of individual saccades in a reflexive (pro-saccade) task showed that patients with PCA had longer latencies than individuals in the Alzheimer’s disease group, particularly for the first major (primary) saccade towards the target at both 5° and 10°, although not at 15°. As far as the saccadic amplitude is concerned, Shakespeare and co-workers report that in PCA, the main (primary) saccade towards the target is hypometric when compared to those in the typical Alzheimer’s disease and control groups, with a greater difference the greater the target distance. Although there was no difference in peak saccadic velocity between the PCA and control groups, the typical Alzheimer’s disease group did show an increased saccadic velocity. In the smooth pursuit task, the PCA group had a significantly lower mean pursuit gain and an increase in catch-up saccades compared to the control group.
Shakespeare et al. have used two specific methods of analysis, the rate of oculomotor impairment and receiver operating characteristic (ROC) analysis, to investigate to what extent the groups can be separated at an individual patient level, as opposed to simply looking at the group means and variances: an approach that is often misleading and that has been the subject of considerable debate. Using a combination of saccadic parameters provides greater discriminative power than the use of only one parameter. The authors report that patients with PCA are impaired on more than one saccadic metric when compared to patients with typical Alzheimer’s disease and healthy controls. Moreover, ROC analysis showed that the saccadic amplitude error has the greatest sensitivity/specificity to discriminate between the PCA and typical Alzheimer’s disease groups. This is of particular interest, as in the past, studies have not used such rigorous statistical methodology and this might account for some of the discrepancies present in the literature.
In PCA, some of the abnormalities seen in the oculomotor study might be expected to relate to patients’ perceptual abnormalities. However, after Bonferroni corrections, the only association that survived was a negative correlation between greater major saccade latency and poorer basic visual processing. Parietal and occipital cortical thinning was significantly associated with increased saccadic latency in agreement with the role of parietal cortex in the generation of reflexive (pro-)saccades. However, it was not associated with the increased frequency of large intrusive saccades, which may have arisen either as a result of impaired attentional control associated with parietal dysfunction or impaired inhibition associated with frontal lobe dysfunction.
Although significant differences in several aspects of saccadic performance were observed in the PCA group compared to controls, Shakespeare et al. acknowledge that the saccadic tasks used in the study were limited and a future study could usefully include more complex tasks to try and increase our understanding of the pathophysiology of these saccadic disturbances in PCA. For example, the authors do mention the possibility of using an anti-saccade task to help delineate the presence of spontaneous motor activity and impaired inhibitory control. A number of studies have found that inhibition errors in the anti-saccadic paradigm can be predicted in Alzheimer’s disease by measures of dementia severity (Shafiq-Antonacci et al., 2003; Crawford et al., 2005). Anti-saccades may provide a functional index of the dorsolateral prefrontal cortex, which is damaged in the later stages of Alzheimer’s disease. Heuer et al. (2013) used this paradigm as a way of testing inhibitory control in Alzheimer’s disease. The results showed that patients were impaired relative to those with minimal cognitive impairment (MCI) and healthy controls, but no such data exist for a PCA group. The paradigm has also been used as a tool for monitoring the progression of Alzheimer’s disease (Kaufman et al., 2010). The anti-saccadic task, therefore, might be a useful and relatively easy way of measuring and comparing executive function in PCA and Alzheimer’s disease.
Another metric to explore in the oculomotor battery relates to microsaccades. A study by Kapoula and colleagues (2014) measured fixational eye movements and in particular microsaccades in Alzheimer’s disease, MCI and healthy individuals. Microsaccade direction differed significantly in patients versus controls, being more oblique in Alzheimer’s disease and MCI, but no abnormalities were observed in microsaccade dynamics (such as duration, peak velocity and the peak duration–magnitude relationship), which are more directly related to the function of the brainstem saccadic generator. Such studies lend support to the idea that microsaccadic metrics may be a useful tool for characterizing neurodegenerative disorders, but to date no such studies exist for PCA.
If oculomotor dysfunction can be used as a reliable biomarker, then when disease-modifying therapies become available, such markers could help with both tracking disease progression and assessing drug efficacy.
