General data: A total of 80 cases of Alzheimer’s disease patients who visited the memory clinic from April 2021 to March 2022 were included in this study. Patient basic information was collected by reviewing case data, including age, gender, education level, occupation, handedness, etc. Additionally, past medical history as well as smoking and drinking history were obtained. The inclusion criteria were as follows: (1) AD diagnosised by the National Institute on Aging- Alzheimer’s Association (NIA-AA) in 2011, with an age of ≥ 50 years old. (2) Having a Magnetic Resonance Imaging hippocampal coronal section MTA score2 points. (3) Being diagnosed as probable Alzheimer’s disease by at least two deputy chief-level psychiatrists or neurologists. (4) Undergoing complete evaluations using MMSE, ADL, NPI scales in the memory clinic and Siemens 3TMR imaging within one week. (5) Reliable caregivers being present during assessments. (6) Approval from our hospital’s ethics committee for conducting this study protocol. (7) Ensuring that patients or their legal guardians were informed about this study and signed informed consent forms.

The exclusion criteria were defined as follows:

(1) Presence of other types of dementia such as vascular dementia(VaD: with a clear obvious dementia syndrome occurring within six months after cerebrovascular events), frontotemporal dementia(FTD), Lewy body dementia(DLB), etc., or presence of diseases that may cause other dementias or cognitive impairments such as head trauma, brain tumors, epilepsy, cerebral inflammatory diseases, demyelinating diseases of central nervous system, normo-pressure hydrocephalus, severe liver or kidney dysfunction, severe anemia, hypothyroidism, vitamin B12 deficiency, and so on. (2) Presence of white matter hyperintensities or other types of white matter lesions resulting from radiation injury, carbon monoxide poisoning, multiple sclerosis (MS), vasculitis, or cerebral white matter malnutrition.(3) Clear history of cerebrovascular disease with a Hachinski Ischemic Score (HIS) > 4 and Fazekas score ≥ 2.(4) Clear history of psychiatric disorders such as schizophrenia or severe depression (HAMD scale score < 8), or clear history of substance abuse including drugs and alcohol.(5) Accompanied by severe disorders, severe aphasia, or inability to complete neuropsychological tests.

Neuropsychological testing: Professional evaluators trained in consistency at the Memory Clinic conduct cognitive assessments of patients. The Mini-Mental State Examination (MMSE), with a maximum score of 30, is utilized. The MMSE scale comprises orientation (10 points), memory (3 points), attention and calculation abilities (5 points), recall ability (3 points), language skills (8 points), and visuospatial skills (1 point). Daily living activities assessment: The Activities of Daily Living Scale (ADL) is utilized, with a total score of 80. This includes the Basic Activities of Daily Living Scale (BADL) consisting of 11 items, and the Instrumental Activities of Daily Living Scale (IADL) comprising 9 items. A higher score indicates poorer daily living activity capabilities. Neuropsychiatric behavior evaluation: Reliable caregivers are interviewed using the Neuropsychiatric Inventory questionnaire(NPI), which covers symptoms such as delusions, hallucinations, agitation/aggression, depression/mood disturbances, anxiety, euphoria/elation, apathy/indifference, disinhibition, irritability/emotional lability abnormal motor behavior, sleep/nighttime behavior, and appetite/eating disorders. Scores are assigned based on frequency and severity of episodes as well as distress caused to caregivers. The total score is 144; higher scores indicate more severe neuropsychiatric behaviors.

According to the NPI scale, a total of 80 patients were categorized into two groups: delusional (n = 36) and non-delusional (n = 44). Among the 36 delusional patients, 12 exhibited mild severity scores (score = 1), while 9 had moderate severity scores (score = 2), and the remaining 15 experienced severe symptoms.

Equipment: 3.0T Magnetom Syngo Skyra, Siemens, Erlangen, Germany, standard 20-channel magnetic head receiving coil for imaging.

Quantitative WHM analysis of white matter: magnetic resonance imaging scan sequence and parameters Structural magnetic resonance imaging: three-dimensional magnetization was used to prepare rapid gradient echo MP-RAGE sequence. Imaging parameters were as follows: The in-plane resolution was 1mmx1mm, the layer thickness was 1 mm, the total number of layers was 180, FA = 8, TR = 7.1ms, TE = 3.0ms, TI = 900ms, TD = 2400ms, FOV = 24cmx18cm, matrix = 256 × 256. Structural magnetic resonance data preprocessing and analysis: SPM software package was used to register the original structural image to the MNI standard template by affine transformation, and then the registered image was segmented into gray matter, white matter, and cerebrospinal fluid. FLAIR sequences were segmented by automatic WMH segmentation. T1, white matter hyperintensity processing: WMH volume (in MM3) was calculated for the whole brain, paraventricular and deep subcortical regions, and each brain region. The head size was normalized and log-transformed to achieve normal distribution.

Flowchart of WMH extraction. (A) Input the raw T1-weighted and FLAIR images into the UBO detector. (B) UBO Detector Processing Pipeline. UBO Detector first performs a rigid-body registration with T1 as the reference and FLAIR as source image and conducts T1 segmentation using SPM12 function. And then the individual T1, coregistered FLAIR, as well as the GM, WM and CSF images were registered to a DARTEL template provided by UBO. Finally, FMRIB’s FAST Segmentation was conducted on skull-stripped and bias field-corrected FLAIR images to generate candidate WMH clusters. (C) K-NN algorithm was performed to determine whether the candidate clusters are WMH or non-WMH

①Statistical analyses of all cognitive and questionnaire data were analyzed using SPSS version 21 ②The Shapiro-Wilk test was used to assess if the continuous variables showed signifificant deviations from normality. For variables that did not signifificantly deviate from normality, a two-sample t test was used to evaluate the difference between delusional patients and non-delusional control groups. For signifificantly non-normal data, the Mann-Whitney U test was used. The x2 test of independence was used to statistically assess categorical variables. Signifificance level was set at p < 0.05. As demographic and cognitive data were primarily used to assess if the two cohorts were comparable, no corrections were used for multiple comparisons to provide a more conservative test for between-group differences.