Neuropsychological profile of mild cognitive impairment with Lewy body disease

Abstract

Objective: Efforts are being made to identify dementia with Lewy bodies (DLB), the second commonest cause of neurodegenerative dementia after Alzheimer's disease (AD), in the Mild Cognitive Impairment (MCI) phase, during which intervention on the disease processes would likely be most successful. Few studies have targeted this group and the cognitive profile of MCI with Lewy bodies (MCI-LB) is therefore unclear. The present study aims to elucidate the neuropsychology of MCI-LB relative to MCI due to AD (MCI-AD) and healthy controls. Methods: In addition to age-matched controls (n = 31), participants with MCI and symptoms suggestive of LB disease were recruited from local clinics. Baseline assessment of all subjects included clinical examination, imaging (123iodinemetaiodobenzylguanidine [MIBG], dopamine transporter imaging [DaTscan]), and comprehensive neuropsychological assessments. Simple and Choice Reaction Time (SRT and CRT) and a Continuous Performance Test-AX (CPT-AX) were also administered to measure intraindividual variability (IIV) in attention using ex-Gaussian modelling of reaction times. MCI patients were diagnosed firstly following National Institute on Aging-Alzheimer’s Association (NIA-AA) criteria for MCI. Participants with demonstrable cognitive impairment but no clinical symptoms or biomarkers for DLB were considered MCI-AD (n = 18). Within MCI-LB (n = 44), individuals with two or more consensus criteria for the diagnostic features or biomarkers of DLB (McKeith et al., 2017) were considered "Probable" MCI-LB (n = 30). White matter integrity was quantified using diffusion tensor imaging (DTI) and tract-based spatial statistics. Results: While both groups are impaired relative to controls, MCI-LB Probable performed worse than MCI-AD on processing speed (Digit Symbol Substitution Test [DSST, p = .011]), executive function (Verbal Fluency [FAS], p = .027) and visuospatial function (pareidolia task, p = .010; Visual Patterns Test, p = .019) tests. In contrast, MCI-AD scored significantly lower than MCI-LB Probable on tests of verbal learning and memory (Rey Auditory Verbal Learning Test short, p = .047, and long delay, p = .025, and retroactive interference, p = .029). DSST was the best predictor of group allocation using a stepwise discriminant analysis, F(2,76) = 36.89, p < .001, and 92.6% of MCI-LB Probable scored at or below the 16th percentile of control DSST scores. Using hierarchical linear regression, a control-informed processing speed composite fully explained group-associated variance in the visuospatial composite, RAVLT learning and RAVLT short delay recall. In contrast, FAS explained only 25.0% of group variance in the visuospatial composite and is not significantly correlated with RAVLT short delay (p = .132). MCI-LB Probable showed increased IIV using ex-Gaussian tau in CRT (p = .021, d = 1.12) and CPT-AX (p = .007, d =0.80) relative to controls, while MCI-AD differed significantly from controls in SRT tau (p = .002, d = 0.93). No difference between groups was found in white mater integrity, although the DSST showed substantial correlation with fractional anisotropy in the sample as a whole. Conclusions: The present study succeeded in demonstrating that the cognitive dysfunction typical of advanced DLB and AD is observable in the MCI phase of clinically-defined MCI-LB and MCI-AD, respectively. MCI-LB showed visuospatial, attentional and processing speed impairments. Processing speed emerged as particularly important to MCI-LB neuropsychology, suggesting a processing speed, rather than executive, mediated model of decline in MCI-LB. MCI-AD, in contrast, shows verbal learning and memory impairment. Future work should pursue this promising evidence of subtle, aetiologically-specific differences in cognition in MCI