Among individuals with Parkinson’s Disease (PD), visual hallucinations (VH) and mild cognitive impairment (MCI) are highly prevalent and often co-occur. Atrophy in similar brain regions [e.g. cholinergic basal forebrain (BF) nuclei] as well as specific cognitive difficulties (e.g. posterior-cortical abilities such as semantic fluency and visuoperception) have been associated with the presentation of each symptom type. While there are separate lines of evidence implicating BF volume in MCI and VH, no study to date has examined BF integrity in patients with concurrent MCI and VH symptomology. Furthermore, no prior studies examining BF integrity in MCI and VH have accounted for the potential confounding effects of dopaminergic medications which are known to exacerbate both symptom types. The aims of this study were to harmonize or bridge the two bodies of literature to determine the common neural substrate of PD-VH and PD-MCI (with an emphasis on the BF), to examine the confounding effects of dopaminergic pharmacotherapy, and to examine whether nondopaminergic “posterior” cognitive abilities differ between PD-MCI with versus without VH.

This study used a clinical chart review and MRI data to examine the associations between BF volume in a large group (n=296) of advanced PD patients (∼10 years disease duration) with and without each VH and MCI, covarying the effect of dopaminergic therapy. A two-way ANCOVA was run on total and regional BF volumes (i.e., total BF volume, and four nuclei including Ch4, Ch4p, Ch1-2, Ch3) using VH and MCI as independent variables, while covarying for dopaminergic medication. Using Mann-Whitney U tests, we compared the performance of individuals with MCI-VH versus that of individuals with MCI-noVH on tasks of semantic verbal fluency and of visuoperceptual skills (e.g., judgement of line orientation, object decision, and silhouettes).

There were two major findings: (1) atrophy of the Ch4 region in the BF was associated with MCI with VH while Ch1-2 was associated with MCI regardless of VH status, and (2) patients with both MCI and VH had poorer performance than individuals with MCI without VH on tasks measuring object recognition but not on tasks of visuospatial perception or semantic verbal fluency. These results remained stable regardless of whether or not dopaminergic medication was included in the model.

PD is a heterogeneous disease with different subtypes reflecting both dopaminergic and cholinergic dysfunction. Our findings suggest further dissociations within the cholinergic system. First, atrophy in Ch4, which projects to the cortical mantle, was preferentially associated with VH symptoms and object-based visuoperception deficits. This is consistent with proposals that VH are real-world manifestations of visuoperceptual deficits. Second, Ch1-2 atrophy, which projects primarily to the hippocampus, was associated with MCI regardless of VH. Future research will extend this work to other cognitive abilities such as memory, to analyses of brain networks that implicate the BF, and to the investigation of the relationship between anti-cholinergic medications and symptom presentation in PD.