Nucleus Basalis of Meynert Degeneration Predicts Cognitive Decline in Corticobasal Syndrome

BACKGROUND: Cognitive changes are common in corticobasal syndrome (CBS) and signi ﬁ cantly impact quality of life and caregiver burden. However, relatively few studies have investigated the neural substrates of cognitive changes in CBS, and reliable predictors of cognitive impairment are currently lacking. The nucleus basalis of Meynert (NbM), which serves as the primary source of cortical cholinergic innervation, has been functionally associated with cognition. This study aimed to explore whether patients with CBS exhibit reduced NbM volumes compared with healthy control participants and whether NbM degeneration can serve as a predictor of cognitive impairment in patients with CBS. METHODS: In this study, we investigated in vivo volumetric changes of the NbM in 38 patients with CBS and 84 healthy control participants. Next, we assessed whether gray matter degeneration of the NbM evaluated at baseline could predict cognitive impairment during a 12-month follow-up period in patients with CBS. All volumetric analyses were performed using 3T T1-weighted images obtained from the 4-Repeat Tauopathy Neuroimaging Initiative. RESULTS: Patients with CBS displayed signi ﬁ cantly lower NbM volumes than control participants ( p , .001). Structural damage of the NbM also predicted the development of cognitive impairment in patients with CBS as assessed by longitudinal measurements of the Clinical Dementia Rating Sum of Boxes ( p , .001) and Mini-Mental State Examination ( p = .035). CONCLUSIONS: Our ﬁ ndings suggest that NbM atrophy may represent a promising noninvasive in vivo marker of cognitive decline in CBS and provide new insights into the neural mechanisms that underlie cognitive impairment in CBS. https://doi.org/10.1016/j.biopsych.2024.01.018

https://doi.org/10.1016/j.biopsych.2024.01.018 Corticobasal syndrome (CBS) is a rare atypical Parkinsonian syndrome characterized by a range of motor and nonmotor symptoms (1).CBS is the prototypical clinical phenotype of corticobasal degeneration (CBD), a 4 microtubule binding repeat tauopathy-related disorder that is associated with severe neuronal depletion, swollen achromatic neurons, and diffusely staining tau-positive astrocytic plaques (2,3).However, the underlying etiology of CBS is very heterogeneous, and up to 40% of CBS cases are actually caused by underlying Alzheimer's disease (AD).Patients with CBS are characterized by an asymmetric pattern of brain atrophy that involves the bilateral premotor cortex, superior parietal lobules, and striatum (4).Although early cognitive impairment was once thought to be rare (5), it is now clear that most patients have cognitive impairment from the early stages of the disease, which has a significant impact on their quality of life and caregiver burden (6).Armstrong et al. found that 52% of cases had cognitive impairment at disease onset and 70% during the disease course (1).The presence of such deficits has been incorporated into diagnostic criteria (7).The pattern and severity of neuropsychological impairments can be highly variable across patients (6).However, the most characteristic impairments are executive dysfunction, visuospatial difficulties, limb apraxia, language problems, and deficits in episodic memory (8).Cognitive impairment can be prominent in CBS and may mimic Alzheimer's dementia early in its disease course (9).Furthermore, a subset of patients with CBD and cognitive predominant syndrome has been described recently (10).Despite the prevalence and impact of cognitive impairment in patients with CBS, relatively few studies have investigated the neural substrates of cognitive changes (11).It has been suggested that loss of cholinergic innervation of the cerebral cortex is one of the mechanisms that contributes to cognitive impairment in other neurodegenerative disorders, such as AD and Parkinson's disease (12)(13)(14).However, it is still unclear whether cognitive decline is associated with cholinergic neural loss in CBS.Postmortem studies have shown that a substantial number of neurons of the nucleus basalis of Meynert (NbM), the primary source of cholinergic innervation of the cerebral cortex, are lost in patients with CBD, suggesting that cholinergic degeneration can be a hallmark of this condition (15).Interestingly, an in vivo positron emission tomography study showed that cerebral cortical acetylcholinesterase activity, which is a measure of cholinergic function, was moderately reduced in patients with CBS and was correlated with cognitive deficits (16).
In this study, we used T1-weighted magnetic resonance imaging (MRI) to explore whether 1) patients with CBS have reduced NbM volumes compared with healthy control (HC) participants, and 2) NbM degeneration can predict cognitive impairment in patients with CBS.To assess whether the relationships between volumetric measures of NbM and cognition were specific to this region, we also evaluated the relationships between the neocortical volumes and cognitive outcomes.Our underlying hypothesis was that NbM volume loss occurs in patients with CBS in association with cognitive impairment and therefore holds potential as a noninvasive in vivo marker of clinical progression in this disease.

Participants
Data used in the preparation of this manuscript were obtained from the 4-Repeat Tauopathy Neuroimaging Initiative (4RTNI) database (http://4rtni-ftldni.ini.usc.edu/).4RTNI was launched in early 2011 and is funded through the National Institute on Aging and The Tau Research Consortium.The primary goal of 4RTNI is to identify neuroimaging and biomarker indicators for disease progression in the 4-repeat tauopathy neurodegenerative diseases.The principal investigator of 4RTNI is Adam Boxer at the University of California, San Francisco.The data are the result of collaborative efforts at 4 sites in North America (University of California, San Francisco; University of California, San Diego; Massachusetts General Hospital, Boston; and University of Toronto).For more information on 4RTNI, please visit http://memory.ucsf.edu/research/studies/4rtni.Each participant met the criteria for possible or probable CBS (1).
We included 38 patients with CBS with baseline T1-weighted MR images.Participants were evaluated using the Mini-Mental State Examination (MMSE), the Unified Parkinson's Disease Rating Scale, the Progressive Supranuclear Palsy Rating Scale (PSPRS), the Clinical Dementia Rating Sum of Boxes (CDR-SB), the Functional Activities Questionnaire (FAQ), and the Schwab and England Activities of Daily Living scale.All participants with CBS who were included in the study underwent at least 2 cognitive evaluations at follow-up (after 6 months and 1 year).Neuropathological data were available for 3 participants who underwent autopsy.For baseline comparisons with the CBS group (38 individuals), 84 age-and sex-matched HC participants were included (Table 1) from the Frontotemporal Lobar Degeneration Neuroimaging Initiative dataset, which is a companion study to 4RTNI.All HC participants had an MMSE score $ 26, a CDR score of 0, and no subjective cognitive complaints and underwent T1-weighted MRI scanning using similar acquisition parameters.The study was approved by the institutional review board of each participating site, and all subjects or their legal guardians gave informed written consent.

MRI Acquisition and Processing
MR images were acquired from all individuals on a 3T Siemens Trio Tim system using a T1-weighted magnetization-prepared rapid acquisition gradient-echo sequence (repetition time/echo time = 2300/2.9ms, flip angle = 9.0 , matrix = 240 3 256 3 160, voxel size: 1 mm 3 ).T1-weighted images were preprocessed using the CAT12 toolbox (Structural Brain Mapping Group) implemented in SPM12.In particular, all images were segmented into gray matter (GM), white matter, and cerebrospinal fluid and then mapped to the standard Montreal Neurological Institute space using the highly accurate highdimensional DARTEL registration method.Normalized images were modulated to guarantee that volumes were preserved after the spatial normalization procedure.Next, preprocessed GM data were smoothed with an 8 mm full width at half maximum isotropic Gaussian kernel.In addition, the intracranial volume was estimated for each participant by summing the GM, white matter, and cerebrospinal fluid volumes (17,18).

Region-of-Interest Analysis
The NbM region of interest was measured using probabilistic maps provided with the SPM Anatomy Toolbox r2.2b and derived from the 3-dimensional reconstruction of histological sections from 10 postmortem human brains (http://www.fzjuelich.de/inm/inm-1/DE/Forschung/).In particular, the NbM corresponds to the cholinergic cell group Ch4 (19) and was identified using an existing available probabilistic anatomical map (20).The region of interest was thresholded at 30% probability and then resampled and warped to the DARTEL template.Volumetric measures of the NbM were extracted from modulated images and by summing the total intensity of voxels within the NbM mask adjusted for the size and number of voxels (21).A similar approach was used to evaluate volumetric measures of the Ch123, which includes the medial septum and the diagonal band of Broca, the hippocampus, the entorhinal cortex (EC), and the neocortex in both patients with CBS and HC participants.Neuromorphometrics atlas (http:// Neuromorphometrics.com) was used to define the neocortical mask in the DARTEL template.

Statistical Analysis
Continuous variables are expressed as mean 6 standard deviation.Between-group comparisons were performed using 1way analysis of variance or Mann-Whitney U tests for normally or non-normally distributed variables, respectively.The normality of variables was tested with Shapiro-Wilk tests.Categorical variables were compared using Pearson's c 2 test.Differences in GM density between HC participants and patients with CBS were assessed using threshold-free cluster enhancement from FSL's randomise (22), a nonparametric permutation-based approach performed with 5000 permutations.Only clusters with p , .05 after correcting for familywise error were considered statistically significant.Differences between groups were evaluated while controlling for age, sex, and total intracranial volume.In the group of participants with CBS, we also conducted a voxelwise regression analysis using the volume of NbM as a regressor.To assess whether basal forebrain (NbM, Ch123) and neocortical volumes differed between groups, we carried out an analysis of covariance, including NbM and neocortical volumes as dependent variables; group as a factor (HC, CBS); and age, sex, disease duration, education, and intracranial volume as covariates.Partial Pearson correlations tests controlling for age, sex, education, and intracranial volume were used (17,18).Baseline statistical analyses were performed using SPSS, version 25.0 (IBM Corp.).To evaluate whether baseline basal volumes were associated with longitudinal cognitive decline within the CBS group, we used linear mixed-effects models implemented in R (R Foundation for Statistical Computing) using the lme4 package.In these models, we included cognitive scores as dependent variables and time, baseline basal forebrain and neocortical volumes, age, sex, disease duration, education, and intracranial volume as fixed effects.We also included an interaction term between time, basal forebrain, and sex to investigate the potential sex effect in the relationship between basal forebrain volume and cognitive decline.Adjustment for multiple comparisons was carried out in all analyses for each regional volume using false discovery rate correction.In addition, we used Spearman correlations to test the association between basal forebrain volumes at baseline and individuals' changes in CDR-SB scores over time (visits 1 or 2 compared with baseline).This allowed us to further understand the relationship between baseline basal forebrain volumes and cognitive decline over time.

Demographic and Clinical Data
Demographic and clinical data at baseline and follow-up are summarized in Table 1.Among the 38 patients with CBS, 3 underwent autopsy and showed CBD pathology.As expected, no significant differences were found in age, sex, or years of education between groups.MMSE scores were lower in patients with CBS than in HC participants (p , .001).

Voxel-Based Morphometry
Compared with HC participants, patients with CBS showed an extended pattern of GM decrease in cortical and subcortical brain regions, including the NbM (Figure 1A).Specifically, patients with CBS exhibited decreased GM in the frontotemporo-parietal regions (familywise error-corrected p , .05).Additionally, patients with CBS demonstrated GM atrophy in the basal ganglia, thalamus, hippocampus, and amygdala compared with the HC group (familywise error-corrected p , .05).Voxelwise regression analysis revealed a significant correlation of the NbM volume with a substantial portion of the neocortex (Figure S1).

Region-of-Interest Analysis
Patients with CBS displayed significantly lower NbM volumes than HCs (p , .001) (Figure 1B, C).Neocortical, hippocampal, and Ch123 volumes were also lower in patients with CBS than in HC participants (p , .001).Neocortical, hippocampal, and NbM volumes were correlated in both the HC and CBS groups (p , .001).Age at onset was correlated with NbM (p = .012)and hippocampal volumes (p = .030)but not with neocortical volume (p = .080).Disease duration was not correlated with NbM or with neocortical or hippocampal volume.There was a significant association between the CDR-SB and neocortex volumes (r = 20.469;p = .01)and between the CDR-SB and hippocampus volume (r = 20.479;p = .007)at baseline in patients with CBS.A trend toward a significant association was found between CDR-SB and NbM volumes (r = 20.338;p = .063)at baseline in patients with CBS.

Association Between Baseline Basal Forebrain Volumes and Longitudinal Cognition in Patients With CBS
We observed a significant interaction between time and baseline NbM volumes in the prediction of longitudinal CDR-SB scores (p , .001) (Table 2; Figure 2) as well as MMSE scores (p = .035).This interaction indicated that patients with reduced NbM volumes at baseline had more rapid cognitive decline over time than patients with higher baseline volumes.
We also observed a significant interaction between time and baseline NbM volumes in the prediction of instrumental activities of daily living as assessed by longitudinal FAQ scores (p = .027).There was no interaction between the global burden NbM Degeneration in Corticobasal Syndrome Using an interaction term between NbM, time, and sex in our model predicting changes in CDR-SB scores, the interaction was not significant (p = .96),suggesting that there was no effect modification by sex.Baseline Ch123 volumes were associated with cognitive decline as measured by longitudinal CDR-SB scores (p = .005)but not MMSE (p = .077).No interaction was found between the global burden of disease as assessed by the PSPRS (p = .173)or instrumental activities of daily living as assessed by longitudinal FAQ scores (p = .052)and Ch123.
Baseline neocortical volumes did not predict cognitive decline assessed by longitudinal measurements of CDR-SB (p = .317)and MMSE (p = .534).However, there was an interaction between the global burden of disease as measured by PSPRS and neocortex volume (p = .003).Baseline hippocampal volumes predicted cognitive decline assessed by longitudinal measurements of CDR-SB (p = .014)and MMSE (p = .012)but not FAQ (p = .129).Baseline EC volumes only predicted cognitive decline assessed by longitudinal measurements of CDR-SB (p = .010)(see Table S1).
The significant interaction between time and baseline NbM volumes in the prediction of longitudinal CDR-SB, MMSE, and FAQ scores remained significant even when neocortex and hippocampal volumes were included as covariates in the linear mixed model (CDR-SB, p , .001;MMSE, p = .03;FAQ, p = .02).Similarly, the association of NbM volumes at baseline with individuals' change in CDR-SB scores between baseline and visit 1 showed a significant negative correlation (r = 20.442,p = .024)(Figure S2).Furthermore, the correlation of NbM volumes at baseline with individuals' change in CDR-SB scores between baseline and visit 2 was even stronger and was also significant (r = 20.62,p = .002)(Figure S3).However, the correlation between individuals' Ch123 volumes at baseline and their change in CDR-SB scores over time (1 or 2 years) was not  significant (p ..1).These results confirm that lower NbM, but not Ch123, volumes at baseline were associated with greater cognitive decline over time.

DISCUSSION
In this study, we found that patients with CBS had significantly lower GM volume in the NbM than HC participants.Moreover, we demonstrated that structural damage of the NbM was predictive and specific for the development of cognitive impairment in patients with CBS as measured by MMSE and CDR-SB.Lower GM volume of the NbM was also associated with the development of loss of independence as measured by FAQ.These results remained significant even when we corrected for baseline neocortical and hippocampal volumes and suggest that NbM may be a promising biomarker for cognitive decline in CBS.
The NbM is well known as one of the main cholinergic nuclei in the brain (23), which projects mainly to the cerebral cortex and hippocampus, and plays a major role in memory and cognition (24,25).Neuropathological changes in the NbM have previously been demonstrated in CBS, providing initial evidence in support of cholinergic degeneration in this condition (15).Kasashima and Oda (15) performed a comparative pathological study of cholinergic neuronal changes in the NbM and the laterodorsal tegmental and pedunculopontine tegmental nuclei in patients with CBD and progressive supranuclear palsy (PSP), another 4 microtubule binding repeat tauopathy, by immunohistochemistry for choline acetyl-transferase staining.Pedunculopontine tegmental nuclei is positioned within the upper pons area of the brainstem, while laterodorsal tegmental nuclei extends from the midbrain tegmentum to the pontine tegmentum within the brainstem, both serve as major sources of cholinergic innervation to the thalamus (26).The authors found that the number of neurons in the NbM and the choline acetyl-transferase-positivity rate of remaining neurons showed a more pronounced decrease in patients with CBD than in patients with PSP.On the other hand, the pedunculopontine tegmental and laterodorsal tegmental neurons, which are known to control the states of wakefulness and the cycle of rapid eye movement sleep, (27), were more prominently reduced in PSP than CBD.However, it is worth noting that while some studies have reported minimal neuronal loss in the NbM in CBS, the extent of neuronal loss is significantly lower than that seen in AD cases (2,28).Despite this, it has been consistently reported that all CBS cases display phospho-tau inclusions that affect a significant proportion of Ch4-NbM neurons, ranging from 30% to 70% (28).The hypothesis of a cholinergic deficit in CBS has been further tested in a positron emission tomography study by Hirano et (16).Consistent with these studies, our study provides new in vivo evidence of structural changes in NbM in patients with CBS.
Regarding cognition, the relationship between cognitive impairment and degeneration in the NbM has been extensively elucidated in other neurodegenerative disorders such as AD and Parkinson's disease (12,21,29,30).Intriguingly, recent models of AD have proposed the NbM as the initial site of structural degeneration, subsequently affecting the EC (14,21).This challenges the dominant model that posits that the onset of Alzheimer's degeneration in the ECs occurs before it extends to the temporoparietal cortex.For example, the loss of GM was more pronounced in the NbM than in the EC among cognitively healthy individuals with abnormal cerebrospinal fluid biomarkers of amyloid-b and hyperphosphorylated tau (14).Furthermore, the baseline volume of the NbM was predictive of the longitudinal structural degeneration in the EC, implying a trans-synaptic spread of neuropathology originating from the NbM (21).However, to the best of our knowledge, this is the first time that in vivo structural imaging of the cholinergic NbM has been used to predict longitudinal changes in cognition in patients with CBS.In this study, we used the MMSE and CDR-SB to assess cognitive impairment.MMSE is the bestknown and most commonly used short screening tool for overall cognitive impairment in clinical, research, and community settings (31).CDR-SB is a widely used scale that has demonstrated validity and reliability in longitudinal assessment of both cognition and function (32,33).Our findings suggest that the NbM is associated with cognitive decline in CBS.Our findings are also consistent with the results of a previous study that showed that cortical acetylcholinesterase activity was correlated with general cognitive function as measured by MMSE in patients with CBS (16).
To substantiate our findings, we also explored the relationship between neocortex and hippocampal volumes and clinical progression.The inclusion of the hippocampus in the analysis was guided by 2 factors: first, it is considered to be a structure the atrophy of which is linked to AD (34); second, it receives its cholinergic innervation from the septal diagonal band neurons rather than from the NbM (26).Although we found that neocortex volumes at baseline did not predict cognitive impairment, neocortex volumes were correlated with the progression of the PSPRS, which is a well-validated, multidomain clinical rating scale used to measure disease severity and progression in both PSP and CBS and is mainly sensitive to motor impairment (35).Hippocampal volumes predicted cognitive impairment as measured by MMSE and CDR-SB but did not predict the development of loss of independence as measured by FAQ.Furthermore, the significant interaction between cognitive progression and baseline NbM volumes remained significant despite the inclusion of neocortex and hippocampal volumes as covariates in the linear mixed model.Therefore, we found that neocortex volume atrophy predicted disease progression and especially motor features, while the progression of cognitive impairment was specifically predicted by NbM atrophy, as well as the decline in activities of daily living.
A limitation of the current study is the lack of pathological data, particularly data pertaining to the biological definition of AD.This includes the lack of neuropathological data for all participants, amyloid positron emission tomography scans, cerebrospinal fluid amyloid, and APOE allele status in the patients with clinically defined CBS.It is important to note that up to 40% of CBS cases are caused by underlying AD (36,37), and the critical and early pathophysiological role of the NbM in AD has been well established (13,14).Therefore, while our study suggests that the NbM may be a promising biomarker for cognitive decline in CBS, we cannot rule out the possibility that many, if not all, patients with CBS included in the study have a neuropathological diagnosis of AD, which may have influenced our findings.However, in the 3 CBS cases for which autopsy data were available, the pathology was CBD and not AD, underscoring the careful selection of participants in this study.Biomarker research is usually most reliable when it is based on neuropathologically confirmed cases, and the lack of such data in our study limits the interpretability of our results.While our study was designed to test NbM as biomarker of cognitive impairment for the clinical syndrome of CBS, which is diagnosed based on clinical criteria, additional studies are required to explore this hypothesis in the pathologically defined entity of CBD.Diagnostic certainty of underlying pathology is essential for the development and application of effective mechanismbased therapies, and biomarker research based on neuropathologically confirmed cases would greatly strengthen our understanding of the biological mechanisms that underlie CBS.Secondly, a detailed neuropsychological test battery, data on drug intake (particularly drugs that have an action on the cholinergic system such as acetylcholinesterase inhibitors), and a longer follow up would be important to fully understand the pattern of cognitive deficits in CBS and its relationship to NbM volume loss.Thirdly, we used only structural data derived from T1-weighted sequences to measure NbM.Future studies with larger sample sizes and more comprehensive neuropsychological evaluations are needed to better understand the relationship between NbM volume loss and cognitive impairment in CBS, as well as to further characterize potential subtypes of the disease.

Conclusions
Taken together, our data provide new insights into the neural mechanisms underlying cognitive impairment in CBS and suggest that in vivo structural measures of the cholinergic NbM predict cognitive impairment in CBS.Additional studies are warranted to explore whether these findings can be replicated in pathologically confirmed CBD.Nonetheless, our study highlights the potential of NbM atrophy as a noninvasive imaging biomarker for monitoring cognitive progression in CBS.

Figure 1 .
Figure 1.Gray matter differences between patients with corticobasal syndrome (CBS) and healthy control (HC) participants.(A) Coronal views of voxel-based morphometry results showing an extended pattern of cortical and subcortical gray matter atrophy in patients with CBS when compared with HCs.Blue lines indicate the level of the coronal sections used to display voxel-based morphometry results.(B) Differences in nucleus basalis of Meynert volumes (mean voxel value) between patients with CBS and HC participants.(C) Mask of nucleus basalis of Meynert used to extract gray matter volume.FWE, familywise error.

Figure 2 .
Figure 2. Nucleus basalis of Meynert (NBM) as a predictor of cognitive decline.Predicted trajectories of Clinical Dementia Rating (CDR) Sum of Boxes (increased CDR-Sum of Boxes score means worsening cognitive and functional impairment) in corticobasal syndrome patients with high (.50th percentile) and low (,50th percentile) baseline NBM volumes (mean voxel value).Prediction models show a significant relationship between longitudinal cognitive decline and NBM volumes.

Table 1 .
Demographic and Clinical Features of Participants at Baseline Functional Activities Questionnaire; HC, healthy control; MMSE, Mini-Mental State Examination; PSPRS, Progressive Supranuclear Palsy Rating Scale; SEADL, Schwab and England Activities of Daily Living; UPDRS-III, Part-III (motor exams) of the Unified Parkinson's Disease Rating Scale.

Table 2 .
Main and Interaction Effects of the Linear Mixed-Effects Models Estimating the Longitudinal Changes of Clinical Features as a Function of Baseline Volume of NbM, Neocortex, and Hippocampus While Controlling for Possible Confounders CDR, Clinical Dementia Rating; Est, estimate; FAQ, Functional Activities Questionnaire; MMSE, Mini-Mental State Examination; NbM, nucleus basalis of Meynert; PSPRS, Progressive Supranuclear Palsy Rating Scale.a False discovery rate-corrected values.NbM Degeneration in Corticobasal Syndrome Biological Psychiatry June 1, 2024; 95:1048-1054 www.sobp.org/journal1051 Biological Psychiatry al. that investigated brain acetylcholinesterase activity by [ 11 C]N-methylpiperidin-4-yl acetate in patients with CBS or PSP.Volume-of-interest analyses showed that mean cortical acetylcholinesterase activity was reduced by 17.5% in CBS and by 9.4% in PSP compared with an HC group