Article Text

Download PDFPDF

Original research
Visuospatial dysfunction predicts dementia-first phenoconversion in isolated REM sleep behaviour disorder
  1. Jing Wang1,2,3,
  2. Bei Huang2,4,
  3. Li Zhou2,
  4. Shi Tang2,
  5. Hongliang Feng1,2,3,
  6. Joey W Y Chan2,
  7. Steven W H Chau2,
  8. Jihui Zhang1,2,3,
  9. Shirley X Li5,
  10. Vincent Mok4,6,
  11. Yun Kwok Wing2,4,
  12. Yaping Liu1,2,3
  1. 1Center for Sleep and Circadian Medicine, Affiliated Brain Hospital of Guangzhou Medical University, Guangzhou, Guangdong, China
  2. 2Li Chiu Kong Family Sleep Assessment Unit, Department of Psychiatry, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, Guangdong, China
  3. 3Key Laboratory of Neurogenetics and Channelopathies of Guangdong Province and the Ministry of Education, Guangzhou Medical University, Guangzhou, Guangdong, China
  4. 4Li Ka Shing Institute of Health Sciences, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, Guangdong, China
  5. 5Department of Psychology and the State Key Laboratory of Brain and Cognitive Sciences, University of Hong Kong, Hong Kong SAR, Guangdong, China
  6. 6Margaret K.L. Cheung Research Centre for Management of Parkinsonism, Division of Neurology, Department of Medicine and Therapeutics, Faculty of Medicine, The Chinese University of Hong Kong, Hong Kong SAR, Guangdong, China
  1. Correspondence to Dr Yaping Liu, Professor, Director of Center for Sleep and Circadian Medicine, The Affiliated Brain Hospital of Guangzhou Medical University, Guangzhou, Guangdong, China; Adjunct Assistant Professor, Department of Psychiatry, Faculty of Medicine, The Chinese University of Hong Kong, Shatin, Hong Kong, China; yaping.liu{at}cuhk.edu.hk; Professor Yun Kwok Wing, Choh-Ming Li Professor of Psychiatry, Director of the Li Chiu Kong Family Sleep Assessment Unit, Department of Psychiatry, Shatin Hospital, The Chinese University of Hong Kong, Shatin, Hong Kong, People's Republic of China; ykwing{at}cuhk.edu.hk

Abstract

Objective While isolated rapid eye movement sleep behaviour disorder (iRBD) is known as a prodrome of α-synucleinopathies, the prediction for its future phenoconversion to parkinsonism-first or dementia-first subtype remains a challenge. This study aimed to investigate whether visuospatial dysfunction predicts dementia-first phenoconversion in iRBD.

Methods Patients with iRBD and control subjects were enrolled in this prospective cohort study. Baseline neuropsychological assessment included the Unified Parkinson’s Disease Rating Scale part III, Montreal Cognitive Assessment (MoCA), Rey-Osterrieth complex figure (ROCF), Colour Trails test (CTT), Farnsworth-Munsell 100-hue test and Digit Span test. The anterior and posterior subscores of MoCA as well as their modified versions were explored. A composite score derived from ROCF and CTT was also explored. Regular follow-up was conducted to determine the phenoconversion status of iRBD patients.

Results The study included 175 iRBD patients and 98 controls. During a mean follow-up of 5.1 years, 25.7% of patients experienced phenoconversion. Most of the neuropsychological tests could differentiate dementia-first but not parkinsonism-first convertors from non-convertors. The modified posterior subscore of MoCA, by integrating the Alternating Trail Making and Clock Drawing components into original the posterior subscore, which mainly reflects visuospatial function, was the strongest predictor for dementia-first phenoconversion (adjusted HR 5.48, 95% CI 1.67 to 17.98).

Conclusion Visuospatial dysfunction, as reflected mainly by the modified posterior subscore of MoCA, is a predictive factor for dementia-first phenoconversion in iRBD, suggesting its potential for being a biomarker for clinical prognostic prediction and potential neuroprotective trials aiming to delay or prevent dementia.

  • DEMENTIA
  • SLEEP DISORDERS

Data availability statement

Data are available on reasonable request. The data collected in this study were not planned to be made to the public currently but might be available on reasonable request to the corresponding author.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

WHAT IS ALREADY KNOWN ON THIS TOPIC

  • Visuospatial dysfunction is a distinct symptom of dementia driven by α-synuclein pathology, however, little is known on whether it exists in the prodromal stage of α-synucleinopathies, that is, isolated rapid eye movement sleep behaviour disorder (iRBD), and if so whether it predicts dementia-first phenoconversion in iRBD.

WHAT THIS STUDY ADDS

  • Visuospatial dysfunction already occurred in iRBD and can predict the dementia-first but not parkinsonism-first phenoconversion of iRBD, with the modified posterior subscore of Montreal Cognitive Assessment (MoCA) showing the maximum discriminative power.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

  • The findings suggested the involvement of neural circuitry underlying the visuospatial function in the aetiology of dementia with Lewy bodies, which may serve as a potential neuroprotective target. In addition, MoCA should be appreciated as a simple and effective tool to identify the patients of iRBD at risk of dementia, which would help stratify the patients for future trials in dementia prevention.

Introduction

Isolated rapid eye movement (REM) sleep behaviour disorder (iRBD) is a parasomnia characterised by a loss of REM sleep muscle atonia during REM sleep, resulting in patients physically acting out their dreams.1 It has been observed that approximately 90% of patients with iRBD would eventually develop α-synucleinopathies such as Parkinson’s disease (PD) and dementia with Lewy bodies (DLB) within 15 years.2 Consequently, iRBD is recognised as a distinct precursor of α-synucleinopathies, and identifying early and accurate markers in iRBD that can predict the trajectory towards a specific type of α-synucleinopathies is of significant importance and timely relevance. Such markers may potentially pave the way for more focused and effective neuroprotective trials.3 4

Visuospatial impairment is a notable and distinct cognitive dysfunction in DLB, distinguishing it from both PD and Alzheimer’s disease.5 6 Studies have demonstrated that visuospatial dysfunction, as assessed by tasks like the intersecting pentagons figure copy, could predict the onset of dementia in PD, independent of impaired executive functions.7 8 In the prodromal stage of α-synucleinopathies, particularly iRBD, emerging evidence has suggested an association between visuospatial dysfunction and a higher risk of future dementia. A postmortem study found more pronounced impairment in visuospatial function in iRBD subjects who converted to DLB within a 7-year follow-up period compared with those who did not.9 Similarly, a prospective study reported that iRBD patients who developed dementia displayed greater baseline visuospatial deficit, alongside other cognitive impairments, compared with those who developed parkinsonism.10 Very recently, the International RBD Study Group highlighted the predictive capacity of visuospatial dysfunction, as measured by figure copy tests, for prospective dementia-first phenoconversion in iRBD patients.11 However, these findings should be further corroborated considering the small sample sizes and the heterogeneity of the test scores merged from multicentres.10 11

Visuospatial function could be assessed not only by the individual neuropsychological tests as mentioned above but also by subscores of a validated test. For example, the posterior subscore of the Montreal Cognitive Assessment (MoCA) was reported to be highly associated with visuospatial function of the posterior cortex.12 In addition, a previous study has reported that a composite score derived from multiple motor tests outperformed the individual tests in predicting clinical outcomes in a cohort of RBD and PD.13 A composite score for visual function might generate more information than individual tests as well. Therefore, the current study aimed to investigate whether visuospatial dysfunction in iRBD patients, as determined by individual neuropsychological tests, the posterior subscore of MoCA, or a composite score derived from individual neuropsychological tests, could effectively predict dementia-first phenoconversion in iRBD.

Methods

Study design and subject recruitment

This was a longitudinal cohort study with a case–control comparison at baseline. The study subjects consisted of patients diagnosed with iRBD based on the criteria outlined in the International Classification of Sleep Disorders (third edition). Age-comparable and sex-comparable individuals without RBD were recruited from the community as control subjects. The recruitment of subjects was based on clinical interviews and video-polysomnography assessments to confirm or exclude RBD diagnosis for patients with RBD and control subjects, respectively. The exclusion criteria for both iRBD and control subjects included the presence of neurodegenerative disorders at baseline, colour blindness and any eye disease that could influence visual function tests, such as cataracts, macular degeneration and glaucoma.

Subject recruitment took place at a university-affiliated sleep centre in Hong Kong spanning from March 2008 to July 2021.14–16

Measurements

Baseline assessments

At baseline, general demographic information was collected using a structured questionnaire. The age at RBD onset, age at diagnosis, comorbidities and medication information were obtained from medical records. Moreover, the diagnosis of depression and the use of antidepressants were further confirmed by the assessment results of the Mini International Neuropsychiatric Interview administered by trained clinicians.14 17 Duration of RBD at baseline was calculated by subtracting the age at diagnosis from the age at onset.

Global cognitive function was evaluated using the locally validated Hong Kong version of MoCA.18 The final score of the MoCA was adjusted for education, with an additional point added for those with less than 6 years of education.18 In addition, the anterior and posterior subscores of the MoCA were calculated using a published method.12 Briefly, the anterior subscore was derived from component scores of MoCA including items of Serial 7 Subtraction subtest, Letter P Fluency subtest, Alternating Trail Making subtest, Letter A Tapping subtest, Backward Digit Span subtest, Abstraction subtest and Clock Drawing subtest. The posterior subscore, as reported to be relevant for visuospatial function, was calculated based on component scores of MoCA, which included the Delayed Recall subtest, Forward Digit Span subtest, Naming, Orientation and Copy of the cube subtest.

As previous research suggested that the Colour Trails test (CTT) and Clock Drawing Test also closely involve visuospatial skills,19 20 we further modified and recalculated the anterior and posterior subscores of MoCA, by integrating the component scores of the Alternating Trail Making subtest (analogous to CTT) and the Clock Drawing subtest (which were supposed to constitute part of the original anterior subscore) into the posterior subscore, to explore whether these modified anterior and posterior subscores of MoCA could improve the predictive performance for phenoconversion in iRBD.

In addition, an independent CTT was conducted, which is a culture fair version of the Trail Making test and is designed to assess selective attention and mental flexibility. Nonetheless, it also requires visuospatial skills to complete, as shown in a previous study.19 21 The CTT consists of two parts. In part 1, subjects were instructed to quickly connect circles numbered 1–25 in sequential order. In part 2, they were required to connect numbered circles with alternating colours (pink and yellow) in the correct sequence. The time taken to complete each trial was recorded as a measure of performance.

Visuospatial function was also assessed using the Rey-Osterrieth Complex Figure (ROCF).22 Subjects were asked to copy a complex figure (copy) and then immediately recall and redraw the figure (immediate recall). After a 30 min delay, subjects were asked to draw the figure again (delayed recall). The score was determined based on the accurately recalled components of the complex figure. We did not analyse the duration of the test completion since subjects with poor recall tended to finish the task earlier, which would inherently introduce bias into the results.

Furthermore, colour vision was evaluated with the Farnsworth-Munsell 100 Hue (FM 100 Hue) Test.23 An independent Digit Span Task was also administered to subjects to evaluate working memory.24 Lastly, the Unified Parkinson’s Disease Rating Scale (UPDRS) part III was conducted to evaluate the motor function.25

Clinical follow-up

All patients with iRBD underwent regular clinical follow-ups (every 3–6 months) for symptomatic management and evaluation of motor and cognitive functioning, with additional periodic (annual to biannual) reassessments including neuropsychological tests at the research site as per research protocol. Patients who exhibited prominent neurodegenerative signs at the clinic or the research site were referred to the neurologists, geriatricians and psychogeriatricians for further evaluation of the presence of neurodegenerative disorders, based on the clinical criteria including UK PD Society Brain Bank Diagnostic Criteria for PD26 and DLB Consortium Criteria for DLB.27 In general, the ‘1-year rule’ was adopted when differentiating DLB and PD dementia (PDD): if dementia begins within 1 year of Parkinsonism, it is diagnosed as DLB; if later, PDD is the diagnosis.27 In this study, both the diagnoses of PD and PDD were classified as parkinsonism-first while DLB was classified as dementia-first. As only a few patients were diagnosed with multiple system atrophy (n=2), they were not included in the analysis.

As for control subjects, the majority of the subjects (86.7%) were actively participating in the RBD-related projects at our research site, at which the neurodegenerative status could be evaluated.15 16 28 Besides, the medical records of all the subjects, including the diagnosis of neurodegenerative disorders or death (censor date of the study: 30 May 2023), could be assessed through the computerised medical system, which covers over 90% of the local citizens.29

Statistical analysis

Continuous variables were reported as mean±SD. Between-group differences were assessed using independent t-tests or nonparametric Mann-Whitney U tests. Categorical variables were presented as n (percentage) and compared using χ2 tests or Fisher’s exact tests, as appropriate.

To integrate the visuospatial components from the CTT (time) and the ROCF (score), a composite score was derived using the principal component analysis (PCA). The first component of the PCA, which captures the maximum variance in the data, represented the composite score. In our dataset, the first component accounted for 58.5% of the total variance. Since the numerical values of the CTT time and the ROCF score were in opposite directions, the reciprocal of the time values was taken to ensure that the numerical values of both tests aligned in the same direction. Subsequently, the composite score was rescaled to a range of 0–100 by subtracting the minimum value, dividing by the range, and multiplying by 100.13

To evaluate the discriminatory performance of baseline neuropsychological tests on phenoconversion, receiver operating characteristic (ROC) analysis was conducted, and the area under the curve (AUC) with its 95% CI was calculated. The distribution of baseline neuropsychological test scores among iRBD patients who did not convert (non-convertors), those who converted to PD (parkinsonism-first convertors), and those who converted to DLB (dementia-first convertors) were visualised using the R programme (V.4.3.1) with the ‘ggplot2’ package. Differences in mean scores of neuropsychological tests among the three groups were assessed using analysis of variance. Since the sample sizes were unequal between three groups and equal variances were not assumed, the post hoc test of Games-Howell was employed.

Furthermore, patients were categorised into two groups based on the median scores (ie, upper half vs lower half). Cox regression models were used to ascertain the risk of phenoconversion, with HR and their corresponding 95% CI calculated. Age, sex, depression and duration from RBD onset to assessment were further adjusted in the models. The follow-up duration was calculated from the date of the baseline neuropsychological assessment to the date of the last clinic visit, date of death (censored data), or date of diagnosis of a neurodegenerative disease. Patients with missing data (5 for MoCA test, 4 for ROCF delay recall test, 3 for FM100 Hue test and 11 for UPDRS part III data) were excluded from the analyses. Statistical significance was defined as a two-tailed p<0.05 for all the tests.

Results

General information and baseline performance

A total of 175 patients diagnosed with iRBD (age at diagnosis, 67.7±6.6 years, 74.3% male) and 98 control subjects (66.6±8.3 years, 66.3% male) were recruited for this study. Over a mean follow-up period of 5.1 years, 25.7% of the patients converted (22 parkinsonism-first convertors, 21 dementia-first convertors and 2 patients converted to multiple system atrophy). Three (3.1%) of the control subjects were diagnosed with dementia (one vascular dementia, one mixed dementia and one dementia type not specified) during a mean follow-up duration of 5.8 years. 14 patients and 4 control subjects deceased before developing any neurodegenerative diseases.

At baseline, patients with iRBD performed worse than control subjects on UPDRS part III, CTT (part 1), and ROCF (copy and immediate recall), and Digit Span test (backward). There were no significant differences in age, sex, education level, follow-up duration and other neuropsychological test results between the groups. However, patients with iRBD showed a higher rate of comorbid depression and use of antidepressants (table 1).

Table 1

Demographic features and measures of cognitive function among participants

Compared with non-convertors, dementia-first convertors had higher error scores on the FM100 Hue test and poorer performance on the CTT (time on part 1 and part 2) and had lower baseline score on the MoCA (total score, posterior and modified posterior subscores), ROCF (immediate and delayed recall) and the composite score of CTT and ROCF. Furthermore, dementia-first convertors preformed significantly worse than parkinsonism-first convertors on CTT part 2 (time), ROCF (immediate and delayed recall, score) and had a lower Composite Score (figure 1).

Figure 1

Distribution of scores on neuropsychological tests at baseline. *p<0.05; **p<0.01. CTT, Colour Trails Test; MoCA, Montreal Cognitive Assessment; ROCF, Rey-Osterrieth Complex Figure; UPDRS, Unified Parkinson’s Disease Rating Scale.

ROC analysis revealed that most of the baseline neuropsychologic tests were capable of differentiating dementia-first convertors from those non-convertors. The CTT (part 1 and part 2, time), ROCF score, FM100 Hue test and Digit Span test (backward) scores, MoCA total score as well as the anterior and posterior subscores of MoCA (both the original and the modified ones), exhibited AUC values ranging from 0.65 to 0.81, with the CTT part 2 (time) had the highest AUC of 0.81 (95% CI 0.71 to 0.92). The composite score of CTT and ROCF had an AUC of 0.80 (95% CI 0.67 to 0.92). However, only the UPDRS part III score was able to differentiate parkinsonism-first convertors from non-convertors with an AUC of 0.66 (95% CI 0.53 to 0.80) (table 2).

Table 2

Baseline measures of cognitive function in differentiating phenoconversion

Predicting dementia-first versus parkinsonism-first using baseline measurements

In the Cox regression model, lower MoCA total score and higher UPDRS part III score initially demonstrated predictive power for parkinsonism-first phenoconversion. However, after adjusting for age, sex, depression and RBD duration at baseline, none of the neuropsychological measures showed significant predictive power for parkinsonism-first conversion. On the other hand, for the prediction of dementia-first phenoconversion, higher error score on the FM100 Hue test, longer time on the CTT part 2 and lower score on the ROCF (copy, immediate recall, delayed recall), lower composite score derived from CTT and ROCF, and lower scores of MoCA (total score, anterior and posterior subscores) were associated with a higher risk of dementia-first phenoconversion (figure 2). After adjusting for the aforementioned covariates, only lower ROCF score (copy: HR 3.39, 95% CI 1.28 to 8.98), lower composite score of CTT and ROCF (HR 3.63, 95% CI 1.17 to 11.22) and lower scores of MoCA (total score: HR 3.94, 95% CI 1.31 to 11.86; posterior subscore: HR 4.82, 95% CI 1.46 to 15.90) remained as predictors of future risk of dementia-first phenoconversion (figures 2 and 3).

Figure 2

Predictive performance on risk of phenoconversion by using different neuropsychological tests. Adjusted model: Cox regression model was adjusted for age, sex, depression and disease duration from age of RBD onset to assessment. *p<0.05, **p<0.01, ***p<0.001. CTT, Colour Trails Test; MoCA, Montreal Cognitive Assessment; ROCF, Rey-Osterrieth Complex Figure; UPDRS, Unified Parkinson’s Disease Rating Scale.

Figure 3

Risk of phenoconversion stratified by the anterior and posterior subscores of MoCA and their modified versions. MoCA, Montreal Cognitive Assessment.

With additional scores from the Alternating Trail Making subtest and the Clock Drawing subtest, the performance of the modified posterior subscore of MoCA (lower half vs upper half) in the prediction of dementia-first phenoconversion, increased from an HR of 4.82 to 5.48 (figure 3). However, both the anterior and posterior subscores of MoCA, regardless whether they were modified or not, failed to predict parkinsonism-first phenoconversion.

Discussion

The present longitudinal study provides compelling evidence that neurocognitive measures assessing visuospatial function could exhibit specific predictive potential for dementia-first but not parkinsonism-first phenoconversion in patients with iRBD. Consequently, these results suggest that visuospatial dysfunction in iRBD patients could serve as a distinctive clinical marker to predict dementia-first phenoconversion, with the modified posterior subscore of MoCA demonstrating the maximum discriminative power.

Visuospatial dysfunction in DLB and its prodromal stage, iRBD

Prior investigations have demonstrated the potential of visuospatial dysfunction to differentiate DLB from Alzheimer’s disease5 6 and PD dementia (PDD).30 Consistently, we also observed that impairment in visuospatial function can manifest even in the prodromal stage of DLB (ie, at the diagnosis of iRBD) and discriminate the phenoconversion of DLB from PD.31 The International RBD Study Group recently reported poor performance on the Trail Making Test Part B as the most robust indicator of dementia-first phenoconversion among 503 iRBD patients.11 Although the Trail Making Test Part B is commonly considered as a measure of attention and executive function, it indeed encompasses some visuospatial elements such as visual scanning, visual sequencing and visuospatial memory.32 Similarly, as a culture fair version of Trail Making test, CTT has also been suggested to evaluate visuospatial abilities.19 In the current study, both the modified posterior subscore of MoCA (by integrating the Alternating Trail Making and Clock Drawing subtest) and the composite score derived from CTT and ROCF, greatly improved the predictive performance. Collectively, these findings underscore visuospatial dysfunction as a salient and distinctive characteristic of DLB, thereby highlighting the potential implication of early identification of this impairment during the prodromal stage.

Other studies also reported that iRBD patients exhibited lower scores on ROCF tests and Corsi Supraspan Learning (assessing visuospatial learning) compared with control subjects.33 Furthermore, a longitudinal case–control study further revealed that patients with iRBD not only displayed poorer visuoconstructional abilities compared with controls but also presented a significant decline in visuospatial learning over time.34 Notably, these visuospatial dysfunctions did not appear to predict future parkinsonism-first phenoconversion during a 2-year follow-up period.34 These findings align with the evidence from our current study that patients who would convert to dementia tend to exhibit poorer baseline visuospatial function.10

Potential pathophysiology underlying visuospatial dysfunction

DLB is characterised by a heightened susceptibility to visual function impairment in comparison to other types of neurodegenerative diseases. Studies have demonstrated reduced activation in the visual area V5/middle temporal and deficits in posterior cortical perfusion in DLB patients when compared with control group.35 Similarly, positron emission tomography investigations conducted by Minoshima et al revealed significant metabolic reductions in the occipital cortex, particularly in the primary visual cortex, among DLB patients, which was not observed in AD patients.36 Furthermore, baseline reductions in glucose metabolism in the primary visual cortex of non-demented individuals predicted a higher risk of conversion to DLB.37 Additionally, our recent meta-analysis examining resting-state functional connectivity in patients with α-synucleinopathy found hypoconnectivity between the visual network and default mode network in DLB patients.38 These neuroimaging findings underscore the involvement of specific cortical areas and networks in DLB, particularly the posterior cortex which is responsible for visual processing.

Cortical visual processing relies on two interconnected yet distinct networks known as the dorsal (occipitoparietal, ‘vision for action’) and ventral (occipitotemporal, ‘vision for perception’) streams. Although recent molecular studies have revealed similarities in the occurrence of pathological α-synuclein filaments (‘Lewy fold’) in the brains of individuals with Lewy pathology (PD, DLB and PDD),39 it has been proposed that Lewy pathology is more likely to deposit in the cerebral cortex in DLB while it primarily affects the substantia nigra in PD.40 41 The spread of Lewy pathology across cortical and subcortical regions, along with involvement of multiple biochemical systems, is particularly prominent in DLB patients, especially those experiencing visual hallucinations.42 Furthermore, studies have indicated reduced cholinergic innervation in the occipital cortex and impaired thalamic cholinergic function in patients with iRBD compared with controls or cases of secondary RBD.43 Given that aberrant cholinergic function is strongly associated with visual hallucinations in DLB,42 impairment of cholinergic function may partially contribute to visuospatial dysfunction in iRBD and potentially serve as a predictor of future phenoconversion to DLB. In addition, Jeong et al recently reported that baseline EEG-based machine learning models could predict the time and subtype of phenoconversion in iRBD with satisfying accuracy.44 Future studies employing machine learning models based on neuroimaging and EEG might help identify visuospatial dysfunction-related patterns in DLB, which would hold the promise to serve as a predictive biomarker for dementia-first phenoconversion in iRBD.

Strengths, limitations and implications

The inclusion of a relatively large sample size comprising iRBD patients with a medium-term follow-up duration enhances the statistical power and robustness of the current findings. Moreover, the utilisation of the modified posterior subscore of MoCA and a composite score derived from CTT and ROCF tests ensures a comprehensive evaluation of visuospatial dysfunction, surpassing the limitations of relying solely on individual single measurements. Of note, as MoCA is a simple and widely used cognitive screening tool, the predicting ability of its modified posterior subscore towards the dementia-first phenoconversion in iRBD might have potential implications for prognostic counselling, disease monitoring and prioritising patients at higher risk of dementia for clinical management and future disease-modifying trials. However, the usefulness of the current modified posterior subscore of MoCA in predicting dementia-first phenocoversion needs to be further corroborated in other independent cohorts and the underlying pathophysiological pathway deserves further clarification.

However, it is imperative to acknowledge certain limitations inherent in this study. First, as previously mentioned, the completion time in the ROCF test was not included as a variable in generating the composite score. This omission may not well capture the overall performance of visuospatial function. However, simultaneously assessment of visuospatial dysfunction and processing speed presents a challenge for existing evaluation tools. For instance, individuals who exhibit poor performance during the recall phase of the ROCF test may end the test earlier.45 Second, the classification of patients into two subgroups based on the median value of neuropsychological test scores might seem arbitrary, though it was necessary to ensure a statistically robust analysis with sufficient power given the limited number of convertors observed in this study. Third, it is crucial to recognise that visuospatial function encompasses various aspects, and the assessment tools used in this study may not cover all domains within this complex construct. Moreover, the modified posterior subscore of MoCA does not purely reflect visuospatial function as the Alternating Trail Making and Clock Drawing subtests may also measure multiple cognitive domains including attention and execution.19 20 Lastly, the current study did not capture the prospective evolution of the visuospatial dysfunction, which is warranted in future studies and better investigated with the combination of neuroimaging data. Besides, the implication of visuospatial dysfunction in the general population needs further investigation.

Conclusion

Visuospatial dysfunction predicts dementia-first phenoconversion in patients with iRBD. The posterior subscore of MoCA reflecting visuospatial function of the posterior cortical domain is a useful indicator for identifying patients of iRBD at risk of dementia. These findings suggest that the impairment of neural circuitry underlying the visuospatial function is specifically involved in the aetiology of prodromal DLB and may serve as a potential neuroprotective target.

Data availability statement

Data are available on reasonable request. The data collected in this study were not planned to be made to the public currently but might be available on reasonable request to the corresponding author.

Ethics statements

Patient consent for publication

Ethics approval

The study was conducted in accordance with the Declaration of Helsinki. The research protocol received approval from the Joint Chinese University of Hong Kong-New Territories East Cluster Clinical Research Ethics Committee (CREC Ref. No.: 2007.455 and 2018.641) and informed consent was obtained from all the subjects prior to their participation in the study.

Acknowledgments

The authors would like to thank all the participants, research staff and colleagues from the Li Chiu Kong Family Sleep Assessment Unit for the study.

References

Footnotes

  • JW and BH contributed equally.

  • Contributors JW and YL contributed to the statistical analyses, data interpretation, and JW had the primary responsibility for drafting the manuscript, with the supervision from YL. JW, BH, LZ, ST, HF, JWYC, SWHC and YL contributed to the subject recruitment and data acquisition. JW, BH and YL conducted the data cleansing and had accessed and verified the data. YL and YKW contributed to the conception and design of the study. JW, BH, YL and YKW had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. All listed authors critically reviewed and revised this manuscript. YL and YKW are responsible for the overall content as guarantor. All listed authors meet authorship criteria and that no others meeting the criteria have been omitted. All authors reviewed and approved the manuscript for submission.

  • Funding This study was supported by the Research Grants Council of Hong Kong under the General Research Fund (Ref No.: 476610) and the Collaborative Research Fund (Ref No.: C4044-21GF), the Health Bureau of Hong Kong under the Health and Medical Research Fund (Ref No.: 06170676), and the Guangdong Basic and Applied Basic Research Foundation (Ref No.: 2024A1515011349).

  • Disclaimer The study funding bodies had no role in the conception, design, conduct, interpretation or analysis of the study or the approval of the publication.

  • Competing interests JW supported by the Faculty Postdoctoral Fellowship Scheme of the Chinese University of Hong Kong and the International Postdoctoral Exchange Fellowship Program (YJ20220085). She is currently a principal investigator, supported by the China Postdoctoral Science Foundation under Grant Number 2022M720913 and 2023T160148. LZ and ST supported by the Faculty Postdoctoral Fellowship Scheme of the Chinese University of Hong Kong. JWYC reported grants from General Research Fund of University Grants Committee and Health and Medical Research Fund-Food and Health Bureau, Hong Kong SAR, which are outside of this study, and personal fees for joining an expert panel meeting from Eisai. YKW received consultation fee and personal fees from Eisai for lecture, and travel support from Lundbeck HK and Aculys Pharma, Japan, which are outside the submitted work. YL reported being as PI, funded by the Health and Medical Research Fund from Food and Health Bureau but not related to this manuscript and received presentation fee from the Chinese Sleep Research Society but not related to this study.

  • Provenance and peer review Not commissioned; externally peer reviewed.