Spatiotemporal Characteristics of Regional Brain Perfusion Associated with Neuropsychiatric Symptoms in Patients with Alzheimer’s Disease

Abstract

Background:

Current technology for exploring neuroimaging markers and neural circuits of neuropsychiatric symptoms (NPS) in patients with Alzheimer’s disease (AD) is expensive and usually invasive, limiting its use in clinical practice.

Objective:

To investigate the cerebral morphology and perfusion characteristics of NPS and identify the spatiotemporal perfusion circuits of NPS sub-symptoms.

Methods:

This nested case-control study included 102 AD patients with NPS and 51 age- and sex-matched AD patients without NPS. Gray matter volume, cerebral blood flow (CBF), and arterial transit time (ATT) were measured and generated using time-encoded 7-delay pseudo-continuous arterial spin labeling (pCASL). Multiple conditional logistic regression analysis was used to identify neuroimaging markers of NPS. The associations between the CBF or ATT of affected brain areas and NPS sub-symptoms were evaluated after adjusting for confounding factors. The neural circuits of sub-symptoms were identified based on spatiotemporal perfusion sequencing.

Results:

Lower Mini-Mental State Examination scores (p < 0.001), higher Caregiver Burden Inventory scores (p < 0.001), and higher CBF (p = 0.001) and ATT values (p < 0.003) of the right anteroventral thalamic nucleus (ATN) were risk factors for NPS in patients with AD. Six spatiotemporal perfusion circuits were found from 12 sub-symptoms, including the anterior cingulate gyri-temporal pole/subcortical thalamus-cerebellum circuit, insula-limbic-cortex circuit, subcortical thalamus-temporal pole-cortex circuit, subcortical thalamus-cerebellum circuit, frontal cortex-cerebellum-occipital cortex circuit, and subcortical thalamus-hippocampus-dorsal raphe nucleus circuit.

Conclusions:

Prolonged ATT and increased CBF of the right ATN may be neuroimaging markers for detecting NPS in patients with AD. Time-encoded pCASL could be a reliable technique to explore the neural perfusional circuits of NPS.

Keywords

Alzheimer’s disease neuroimaging neuropsychiatric symptoms perfusion

INTRODUCTION

Neuropsychiatric symptoms (NPS), including agitation, aggression, anxiety, depression, sleep disorders, and appetite disturbances, are an important group of highly heterogeneous symptoms observed in patients with Alzheimer’s disease (AD) [1]. More than 80% patients with AD develop at least one NPS sub-symptom, accelerating disease progression and increasing the burden on caregivers [2, 3]. Despite their high prevalence, NPS are not part of any diagnostic paradigm for AD, and advances in the pharmacological treatment of NPS remain extremely inadequate [4, 5]. The causes of limited NPS research progression can mainly be attributed to its complex and unknown pathogenesis, heterogeneous sub-symptoms, and delayed intervention [6]. Therefore, identifying novel markers that can help track the onset and progression of NPS and exploring the potential neural pathogenesis of general or specific NPS for targeted treatment in these patients is crucial.

Previous neuroimaging studies have attempted to reveal brain areas or the underlying neural circuits associated with NPS in patients with AD using various modalities; however, most focused on the isolated sub-symptoms of NPS [7, 8]. Some studies did not find associations between NPS and changes in the regional brain structure or neural activity because of methodological issues [9]. Moreover, most patients with AD have coexisting NPS sub-symptoms or develop many sub-symptoms throughout the disease course, suggesting that NPS sub-symptoms may share common pathophysiological processes and similar trends over time [6]. A recent review summarized the NPS-related general and specific patterns of brain lesions, which may help track the treatment response; however, when exploring the mechanism of a particular NPS sub-symptom, other sub-symptoms were not adjusted for [10].

Pseudo-continuous arterial spin labeling (pCASL) is a non-invasive and reliable magnetic resonance (MR) perfusion technique with decreased susceptibility artifacts. The consensus of the International Society for Magnetic Resonance in Medicine (ISMRM) Perfusion Study Group and the European Consortium for ASL in dementia has recommended pCASL for clinical applications in evaluating brain perfusion [11]. An ASL follow-up study demonstrated that the baseline regional cerebral blood flow (CBF) in the posterior cingulate cortex (PCC) and bilateral superior medial frontal regions predicted the conversion and time to conversion from mild cognitive impairment (MCI) to AD [12]. However, traditional ASL techniques have low signal-to-noise ratios (SNR), leading to reductions in image quality that limit their use in clinical practice [13]. Recent developments in ASL involve the multi-time point acquisition of dynamic ASL images using time-encoded pCASL to improve the SNR across different time points and objectively maintain higher CBF accuracy across a wider range of arterial transit times (ATTs) [14, 15]. A recent study found that combining CBF and ATT with time-encoded pCASL resulted in higher performance for identifying MCI than CBF of 1-delay ASL when added to sex, age, apolipoprotein E (APOE) epsilon 4 allele (ɛ4) carrier status, and education years [16]. The interest in clinical research utilizing pCASL in AD-related cognitive impairment is growing; however, previous work on regional CBF related to NPS is limited and focused on investigating specific NPS sub-symptoms [17, 18]. Moreover, the selection of examined brain areas was arbitrary across different studies, and the methods used to evaluate cerebral perfusions, such as MR imaging, ASL, single-photon emission computed tomography, and positron emission tomography, have high heterogeneity [10 , 17–19]. Research on the spatiotemporal characteristics of cerebral perfusion in general and specific NPS sub-symptoms is limited. Thus, this study aimed to 1) explore the cerebral morphology and perfusion characteristics of NPS in patients with AD for early detection and 2) examine the spatiotemporal characteristics of cerebral perfusion in different NPS sub-symptoms to explore potential neuroregulatory mechanisms using the non-invasive and convenient time-encoded pCASL technique.

MATERIALS AND METHODS

Patient and data collection

The data used in this study were collected between August 2021 and October 2022 from the Chinese Imaging, Biomarkers, and Lifestyle (CIBL) study, an ongoing, large-scale, prospective cohort study of the risk factors, genetics, biomarkers, neuroimaging, and lifestyles of Chinese individuals with AD. The study was registered at chictr.org.cn (identifier: ChiCTR2100049131). The inclusion criteria were as follows: 1) aged 55–85 years; 2) met the criteria for probable AD according to the 2011 or 2018 National Institute on Aging-Alzheimer’s Association workgroup diagnostic criteria [20, 21]; 3) objective cognitive impairment based on Mini-Mental State Examination (MMSE) scores after adjusting for educational levels; and 4) significant interference with daily living abilities. The exclusion criteria were as follows: 1) severe AD; 2) core features of other central neurodegenerative diseases, such as dementia with Lewy bodies, frontotemporal dementia, and Parkinson’s disease; 3) concomitant central nervous system diseases that may cause cognitive impairment, such as cerebrovascular disease, tumor, encephalitis, and epilepsy; 4) a history of mental disorders according to the Diagnostic and Statistical Manual of Mental Disorders-5 [22]; 5) cognitive impairment due to traumatic brain injury; 6) history of drug abuse or exposure to toxic environments; 7) new cerebral infarction, hemorrhage, or tumors while undergoing pCASL in this study; and 8) incomplete MR sequence or missing data.

The Institutional Review Board of Capital Medical University, Beijing Tiantan Hospital (Beijing, China; KY-2021-028-01) approved the CIBL study. All participants and their caregivers provided written informed consent.

Nested case-control study

We conducted a nested case-control study of 102 patients with mild to moderate AD and NPS. Controls were randomly selected from the baseline AD participants who did not have NPS at baseline and were matched for age (±1 year) and sex (male or female) with the cases in a 1 : 2 ratio. After excluding those with missing data, 102 incident cases in the NPS group and 51 matched controls were included in the final analysis (Fig. 1).

Fig. 1

Flow diagram explaining the inclusion and exclusion criteria. AD, Alzheimer’s disease; CIBL, the Chinese Imaging, Biomarkers, and Lifestyle of Alzheimer’s Disease Cohort; NPS, neuropsychiatric symptoms; MRI, magnetic resonance imaging; pCASL, pseudo-continuous arterial spin labeling.

Clinical assessment

In this study, we collected demographic data, including age at initial enrollment, sex, years of education, body mass index (BMI), and the waist-to-hip ratio (WHR). Information on the APOE ɛ4 allele carrier status was also collected. Patients or their caregivers completed a comprehensive neuropsychological battery, including the MMSE, Neuropsychiatric Inventory (NPI), Activities of Daily Living (ADL), Pittsburgh Sleep Quality Index (PSQI), Mini-Nutritional Assessment (MNA), and Caregiver Burden Inventory (CBI). Two trained neurologists and neuropsychologists administered these neuropsychological tests.

NPS were evaluated using the NPI, a fully structured interview used to investigate 12 different behavioral and neuropsychiatric domains, including delusion, hallucination, depression, anxiety, agitation/aggression, euphoria, disinhibition, irritability/lability, apathy, aberrant motor activity, sleep disturbances, and appetite disturbances [23]. The NPI scale provides a severity score (on a 3-point scale) and a frequency score (on a 4-point scale) for each reported symptom, and the NPI total severity score was calculated by summing the 12 severity sub-scores (severity score*frequency score).

Objective cognition was assessed using the MMSE, and a total score of≤24 for>6 years of education,≤20 for 1–6 years of education, and≤17 for 0 years of education was considered to indicate cognitive impairment [24].

Sleep quality was assessed using the Chinese version of the PSQI which consists of 19 items with seven subscales, subjective sleep quality, sleep latency, sleep duration, habitual sleep efficiency, sleep disturbances, use of sleep medications, and daytime dysfunction [25]. The global PSQI score is equal to the sum of each subscale score (on a 3-point scale).

The MNA scale comprises simple measurements and brief questions, including anthropometric measurements, global assessment, dietary questionnaires, and subjective assessment [26].

The ADL questionnaire includes 10 items that assess basic ADL tasks and 10 items that assess instrumental ADL tasks. The total score (ranging from 20–80) is the sum of each item’s score (on a 4-point scale).

The CBI is a 24-item multidimensional questionnaire that quantifies the impact of caregiver burden in five domains: time dependence and developmental, physical, social, and emotional burden [27]. The total CBI score ranges from 0 to 96, with higher scores indicating greater levels of perceived burden.

MR imaging acquisition and processing

A 3.0 T MR scanner (SIGNA Premier; GE Healthcare, Milwaukee, WI, USA) with a 48-channel head coil was used for imaging acquisition. High-resolution three-dimensional T1 scans were acquired using the inversion recovery gradient recalled echo sequence with the following parameters: repetition time (TR) = 7.3 ms, echo time (TE) = 3.0 ms, inversion time = 450 ms, flip angle = 12°, field of view (FOV) = 256 mm×256 mm, acquisition matrix = 256×256, slice thickness = 1.0 mm, slice number = 176, and scan time = 4 min 56 s. ASL perfusion imaging was performed using time-encoded 7-delay pCASL with the following parameters: axial acquisition, repetition time = 9315.0 ms, echo time = 11.2 ms, field of view = 220 mm×220 mm, acquisition matrix = 512×512, slice thickness = 3.0 mm, and slice number = 48; the label durations of the seven labeling blocks were 0.361, 0.378, 0.402, 0.436, 0.491, 0.591, and 0.842 s, the post-labeling delays were 1.000, 1.361, 1.739, 2.141, 2.577, 3.067, and 3.658 s, and the scan time was 15 min 55 s. Figure 2 demonstrates the diagram of cerebral perfusion by 7-delay pCASL.

Fig. 2

Diagram of cerebral perfusion by time-encoded 7-delay pseudo-continuous arterial spin labeling (pCASL). The raw maps of pCASL imaging, including different post label delays ranging from 1.000 to 3.658 seconds (A). The arrival-time-corrected cerebral blood flow colormaps (B). The arterial arrival time colormaps (C).

A junior radiologist with 5 years of experience in neuroradiology performed data processing using CereFlow software 1.0 (Anying Technology Beijing Co., Ltd., Beijing, China), which was checked by a senior radiologist with 20 years of experience in neuroradiology. The following steps were performed: 1) importing automatically generated CBF/ATT images from the default vendor’s postprocessing pipeline; 2) co-registration of the M0 image (GE ASL’s PD image) with the anatomical T1-weighted image; the CBF/ATT images were also co-registered to the T1 image with the same transformation parameters; 3) normalization of the T1 images to the Montreal Neurological Institute template; 4) warping the CBF/ATT images into the Montreal Neurological Institute space using the forward transformation matrix derived from T1; and 5) extraction of the regional CBF/ATT using the third edition of the automated anatomical atlas (AAL3) [28].

Statistical analysis

All statistical analyses were performed using SAS Version 9.4 software (SAS Institute, Inc., Cary, NC, USA) and R (version 4.2.1; R Foundation for Statistical Computing, Vienna, Austria). The chi-square or Fisher’s exact test was used to assess statistical differences in categorical variables, whereas independent samples t-tests for normally distributed data or the Mann–Whitney U test for skewed data were used to evaluate continuous variables. Multivariate conditional logistic stepwise regression was used to determine independent predictors of all clinical and radiological influencing factors with a p < 0.05 in the univariate analysis (α_in = 0.10, α_out = 0.05). Univariate and multivariate logistic regression were performed to investigate the association between NPS sub-symptoms and the CBF or ATT of 169 brain regions based on the AAL3. The adjusted confounding factors include age, sex, and the CBF or ATT of the whole brain in the multivariate logistic regression. The significance level of all tests was set at p < 0.05 (two-sided).

RESULTS

Baseline clinical characteristics

Table 1 shows the clinical characteristics of 102 AD patients with NPS and 51 AD patients without NPS. The patients in the NPS group had lower MMSE scores than those without NPS (median [interquartile range, IQR]: 18.00 [11.00, 23.00] versus 24.00 [20.00, 25.00], Z = –4.619, p < 0.001). The MNA score correlated positively with nutritional status and was lower in the NPS group (median [IQR]: 20.25 [18.50, 23.00] versus 23.00 [22.00, 24.00], Z = –4.257, p < 0.001) than in the control group. However, the NPS group had higher ADL and CBI scores than the control group (median [IQR]: 25.50 [21.00, 38.00] versus 21.00 [20.00, 24.00], Z = –4.195, p < 0.001; 21.50 [8.75, 43.50] versus 2.00 [0.00, 10.00], Z = –6.310, p < 0.001). No significant differences were observed in age, sex, education, BMI, WHR, and APOE ɛ4 carrier status between the two groups (p>0.05).

Table 1

Clinical and neuropsychological characteristics of the two groups

Variable	NPS group (N = 102)	Control group (N = 51)	t/χ²/Z	p
Demographics
Age [y, x±s]	68.60±8.10	68.55±8.12	0.035	0.972
Sex [male (%)]	38.00 (37.25)	19.00 (37.25)	0.000	1.000
BMI [kg/m², median (IQR)]	22.71 (20.93, 25.60)	23.55 (21.98, 25.46)	–1.277	0.202
WHR [%, median (IQR)]	0.88 (0.82, 0.92)	0.87 (0.83, 0.92)	–0.453	0.651
Education [y, median (IQR)]	11.00 (9.00, 13.00)	10.00 (7.00, 15.00)	–0.451	0.652
APOE ɛ4 carrier (%)	42.00 (41.18)	16.00 (31.37)	1.388	0.239
Neuropsychological battery
MMSE [score, median (IQR)]	18.00 (11.00, 23.00)	24.00 (20.00, 25.00)	–4.619	<0.001
ADL [score, median (IQR)]	25.50 (21.00, 38.00)	21.00 (20.00, 24.00)	–4.195	<0.001
PSQI [score, median (IQR)]	4.00 (2.00, 7.00)	4.00 (2.00, 7.00)	–0.180	0.857
MNA [score, median (IQR)]	20.25 (18.50, 23.00)	23.00 (22.00, 24.00)	–4.257	<0.001
CBI [score, median (IQR)]	21.50 (8.75, 43.50)	2.00 (0.00, 10.00)	–6.310	<0.001

NPS, neuropsychiatric symptoms; BMI, body mass index; WHR, waist-to-hip ratio; APOE ɛ4, apolipoprotein E epsilon4 allele; MMSE, Mini-Mental State Examination; ADL, Activities of Daily Living scale; PSQI, Pittsburgh Sleep Quality Index, MNA, Mini-Nutritional Assessment; CBI, Caregiver Burden Inventory; IQR, interquartile range.

Regional brain volumes and perfusions based on AAL3

We found significant differences in gray matter volumes, CBF, and ATT between the two groups (Table 2). Patients in the NPS group had larger volumes of lobule I and II of the vermis (p = 0.038) and the left reuniens nucleus (p = 0.005) than those in the control group. The CBF values of five brain regions, including the left gyrus rectus (p = 0.013), left medial orbital gyrus (p = 0.016), right medial orbital gyrus (p = 0.037), left lateral orbital gyrus (p = 0.025), and left temporal pole (p = 0.049) were significantly lower in patients with NPS. However, the CBF values of the right anteroventral thalamic nucleus (ATN) were significantly higher in patients with NPS (median [IQR]: 27.98 [16.87, 41.42] versus 21.64 [15.14, 30.71], Z = –2.05, p = 0.040) than in those without. Compared with those in the control group, the patients with NPS had higher ATT values in 38 brain regions, including the bilateral inferior frontal gyrus (opercular part), bilateral inferior frontal gyrus (triangular part), bilateral inferior frontal gyrus (pars orbitalis), bilateral Rolandic operculum, bilateral superior frontal gyrus (medial orbital), left gyrus rectus, bilateral medial orbital gyrus, bilateral posterior orbital gyrus, insula, left middle cingulate and paracingulate gyri, bilateral caudate nucleus, left putamen, bilateral pallidum, left Heschl’s gyrus, left superior temporal gyrus, right ATN, lateral posterior nucleus of the right thalamus, ventral lateral nucleus of the right thalamus, ventral posterolateral nucleus of the right thalamus, right reuniens, right mediodorsal medial magnocellular part, right anterior pulvinar, right anterior cingulate cortex (subgenual), left anterior cingulate cortex (pregenual), left anterior cingulate cortex (supracallosal), bilateral ventral tegmental area, and right red nucleus (p < 0.05).

Table 2

Regional brain volumes and perfusions based on the third edition of the Automated Anatomical Atlas between the two groups

Variable	NPS group (N = 102)	Control group (N = 51)	t/χ²/Z	FDR-p
Volumes (mm³)
Lobule I, II of the vermis (mean±SD)	0.22±0.05	0.20±0.04	2.09	0.038
The left reuniens nucleus [median (IQR)]	0.01 (0.01, 0.02)	0.01 (0.01, 0.01)	–2.81	0.005
Cerebral blood flow (ml/100 g/min)
The left gyrus rectus [median (IQR)]	40.34 (28.75, 50.57)	45.53 (38.4, 52.23)	2.48	0.013
The left medial orbital gyrus [median (IQR)]	35.23 (23.4, 44.21)	40.27 (34.15, 44.7)	2.41	0.016
The right medial orbital gyrus (mean±SD)	37.07±12.62	41.66±12.98	2.10	0.037
The left lateral orbital gyrus (mean±SD)	35.70 ±16.78	41.86±13.90	2.26	0.025
The left temporal pole [median (IQR)]	32.56 (25.26, 40.2)	35.85 (30.01, 43.8)	1.97	0.049
The anteroventral nucleus of the right thalamus [median (IQR)]	27.98 (16.87, 41.42)	21.64 (15.14, 30.71)	–2.05	0.040
Arterial transit time (s)
The left inferior frontal gyrus, opercular part (mean±SD)	1.45±0.16	1.39±0.14	2.43	0.016
The right inferior frontal gyrus, opercular part (mean±SD)	1.46±0.17	1.38±0.13	2.67	0.008
The left inferior frontal gyrus, triangular part (mean±SD)	1.52±0.17	1.45±0.16	2.25	0.026
The right inferior frontal gyrus, triangular part [median (IQR)]	1.49 (1.37–1.57)	1.42 (1.32–1.52)	–2.19	0.028
The left inferior frontal gyrus, pars orbitalis [median (IQR)]	1.47 (1.37–1.6)	1.42 (1.29–1.54)	–1.96	0.049
The right inferior frontal gyrus, pars orbitalis (mean±SD)	1.47±0.18	1.40±0.18	2.34	0.021
The right rolandic operculum (mean±SD)	1.39±0.17	1.32±0.13	2.59	0.011
The left rolandic operculum (mean±SD)	1.41±0.19	1.34±0.15	2.15	0.033
The left superior frontal gyrus, medial (mean±SD)	1.48±0.16	1.42±0.15	2.23	0.027
The right superior frontal gyrus, medial orbital (mean±SD)	1.45±0.18	1.38±0.16	2.33	0.021
The left gyrus rectus (mean±SD)	1.41±0.15	1.35±0.15	2.23	0.027
The left medial orbital gyrus (mean±SD)	1.50±0.17	1.44±0.16	2.21	0.029
The right medial orbital gyrus (mean±SD)	1.49±0.17	1.41±0.15	2.84	0.005
The left posterior orbital gyrus (mean±SD)	1.45±0.19	1.38±0.18	2.03	0.044
The right posterior orbital gyrus [median (IQR)]	1.41 (1.31–1.5)	1.31 (1.25–1.42)	–2.69	0.007
The left insula (mean±SD)	1.34±0.16	1.28±0.13	2.45	0.016
The left middle cingulate & paracingulate gyri (mean±SD)	1.50±0.17	1.44±0.18	2.14	0.034
The left caudate nucleus (mean±SD)	1.41±0.17	1.34±0.16	2.33	0.021
The right caudate nucleus [median (IQR)]	1.40 (1.26–1.48)	1.34 (1.25–1.42)	–2.20	0.028
The left lenticular nucleus, putamen (mean±SD)	1.31±0.17	1.24±0.14	2.60	0.010
The left lenticular nucleus, pallidum (mean±SD)	1.28±0.17	1.22±0.15	2.18	0.031
The right lenticular nucleus, pallidum [median (IQR)]	1.24 (1.15–1.33)	1.18 (1.1–1.27)	–2.20	0.028
The left Heschl’s gyrus (mean±SD)	1.35±0.18	1.28±0.13	2.32	0.022
The left superior temporal gyrus (mean±SD)	1.45±0.17	1.38±0.15	2.42	0.017
The anteroventral nucleus of the right thalamus (mean±SD)	1.49±0.25	1.40±0.22	2.03	0.045
The lateral posterior nucleus of the right thalamus (mean±SD)	1.64±0.29	1.54±0.29	2.06	0.041
The ventral lateral nucleus of the right thalamus [median (IQR)]	1.42 (1.25–1.6)	1.32 (1.19–1.48)	–1.97	0.048
The ventral posterolateral nucleus of the right thalamus [median (IQR)]	1.43 (1.3–1.57)	1.30 (1.23–1.55)	–2.03	0.043
The right reuniens (mean±SD)	1.53±0.23	1.45±0.22	2.07	0.041
The right mediodorsal medial magnocellular part (mean±SD)	1.50±0.20	1.43±0.19	2.05	0.042
The right mediodorsal medial parvocellular part [median (IQR)]	1.57 (1.29–1.76)	1.43 (1.29–1.58)	–2.30	0.021
The right pulvinar anterior (mean±SD)	1.60±0.26	1.50±0.28	2.17	0.031
The right anterior cingulate cortex, subgenual [median (IQR)]	1.32 (1.21–1.43)	1.26 (1.2–1.36)	–2.08	0.037
The left anterior cingulate cortex, pregenual (mean±SD)	1.36±0.16	1.29±0.14	2.47	0.015
The left anterior cingulate cortex, supracallosal [median (IQR)]	1.31 (1.24–1.41)	1.26 (1.21–1.34)	–2.53	0.012
The left ventral tegmental area [median (IQR)]	1.33 (1.16–1.55)	1.26 (1.08–1.43)	–1.99	0.047
The right ventral tegmental area [median (IQR)]	1.38 (1.18–1.58)	1.29 (1.13–1.47)	–2.04	0.041
The right red nucleus (mean±SD)	1.43±0.23	1.35±0.19	2.12	0.036

NPS, neuropsychiatric symptom; SD, standard deviation; IQR, interquartile range ratio.

Conditional multivariate logistic regression analyses

The multivariate conditional logistic regression included four significant clinical factors (MMSE scores, ADL scores, MNA scores, and CBI scores) combined with the volumes, CBF values, or ATT values of statistically significant brain regions, respectively, from the univariate analysis (Table 3). For clinical factors with regional brain volumes, the findings of conditional multivariate logistic regression analyses revealed that only lower MMSE scores (OR = 0.836, 95% CI: 0.736–0.920, p = 0.002) and higher CBI scores (OR = 1.064, 95% CI: 1.022–1.119, p = 0.006) were independent risk factors for NPS in patients with AD. For clinical factors with CBF values, the findings demonstrated that lower MMSE scores (OR = 0.910, 95% CI: 0.867–0.949, p < 0.001), higher CBI scores (OR = 1.086, 95% CI: 1.054–1.126, p < 0.001), and higher CBF values of the right ATN (OR = 1.053, 95% CI: 1.023–1.090, p = 0.001) were independently associated with NPS in patients with AD. For clinical factors with ATT values, lower MMSE scores (OR = 0.874, 95% CI: 0.809–0.935, p < 0.001), higher CBI scores (OR = 1.074, 95% CI: 1.042–1.115, p < 0.001), and higher ATT values of the right ATN (OR = 5.407, 95% CI: 1.923–17.416, p = 0.003) were independent risk factors for NPS in patients with AD.

Table 3

Multivariate conditional logistic regression of risk factors of Alzheimer’s disease in patients with neuropsychiatric symptoms

Variable	OR (95% CI)	p
Clinical factors with regional brain volumes
MMSE scores	0.836 (0.736–0.920)	0.002
CBI scores	1.064 (1.022–1.119)	0.006
Clinical factors with CBF
MMSE scores	0.910 (0.867–0.949)	<0.001
CBI scores	1.086 (1.054–1.126)	<0.001
Right anteroventral thalamic nucleus	1.053 (1.023–1.090)	0.001
Clinical factors with ATT
MMSE scores	0.874 (0.809–0.935)	<0.001
CBI scores	1.074 (1.042–1.115)	<0.001
Right anteroventral thalamic nucleus	5.407 (1.923–17.416)	0.003

MMSE, Mini-Mental State Examination; CBI, Caregiver Burden Inventory; CBF, cerebral blood flow; ATT, arterial transit time; OR, odds ratio; CI, confidence interval.

Relationship between specific NPS sub-symptoms and the regional brain CBF values in patients with AD

The most common NPS sub-symptoms were depression (n = 84, 54.90%) and anxiety (n = 84, 54.90%), followed by apathy (n = 73, 47.71%), irritability (n = 52, 33.99%), sleep disturbances (n = 48, 31.37%), agitation (n = 44, 28.76%), appetite disturbance (n = 32, 20.92%), delusion (n = 31, 20.26%), disinhibition (n = 28, 18.30%), hallucination (n = 21, 13.73%), aberrant motor activity (n = 20, 13.07%), and euphoria (n = 14, 9.15%).

Supplementary Table 1 presents a detailed overview of the association between regional brain CBF values and 12 NPS sub-symptoms with or without adjustment for confounding factors. A significant negative association was observed between the CBF values of seven brain areas, including the rectus, orbital gyrus, temporal lobe, olfactory cortex, anterior cingulate and paracingulate gyri, amygdala, and insula, and specific sub-symptoms of NPS, which suggested that the higher CBF values of these brain areas had a protective effect against NPS (OR < 1, p < 0.05 after adjusting for confounding factors). However, a significant positive association was observed between the CBF values of seven brain areas including the cerebellum, locus coeruleus, calcarine, cuneus, lingual gyrus, occipital lobe, and parietal lobe and specific sub-symptoms of NPS in patients with AD, which suggested that the higher CBF values of these brain areas increased the risk of NPS (OR>1, p < 0.05 after adjusting for confounding factors). The perfusion characteristics based on pCASL in a patient with depression and anxiety has been shown in Fig. 3. Interestingly, the cerebellum was the brain region most widely associated with these sub-symptoms, showing a positive association with all of them except for disinhibition, irritability, and sleep disturbances.

Fig. 3

The perfusion characteristics based on pseudocontinuous arterial spin labeling in a patient with depression and anxiety in this study. A 74-year-old woman presented with progressive cognitive impairment and significant depression and anxiety for six years. Brain pseudocontinuous arterial spin labeling revealed a marked reduction of the cerebral blood flow (CBF) in the left opercular part of inferior frontal gyrus (Ading172, 30.49 ml/100 g/min), left middle cingulate and paracingulate gyri (B-Cding173, 41.05 ml/100 g/min), left supramarginal gyrus (Bding174, 35.20 ml/100 g/min), but a relative increase of the CBF in the lobule IV and V of vermis (B-Cding176, 69.07 ml/100 g/min) and lobule VI of vermis (B-C ding176, 62.06 ml/100 g/min). The sequencing of arterial transit time (ATT) values were 1.29 s in left caudate (D-Eding177;), 1.315 s in the left amygdala (Fding178), 1.403 s in the right temporal pole (Dding179), 1.614 s in the left supramarginal gyrus (Fding180), and 1.685 s in the left angular gyrus (Fding181), suggesting a potential subcortical thalamus-temporal pole-frontal/parietal cortex circuit.

Spatiotemporal perfusion characteristics based on the ATT values of different brain regions in specific NPS sub-symptoms

Supplementary Table 2 presents the association between 12 NPS sub-symptoms and the mean ATT values of the involved brain regions. Six spatiotemporal perfusion circuits were identified for 12 NPS sub-symptoms in patients with AD based on the differences in ATT values among brain regions (Fig. 4). The anterior cingulate gyri-temporal pole/subcortical thalamus-cerebellum circuit was associated with delusion, euphoria, and apathy in patients with AD. The insula-limbic-cortex circuit was associated with hallucinations and agitation. The subcortical thalamus-temporal pole-cortex circuit was associated with depression, anxiety, disinhibition, and irritability. Aberrant motor activity was characterized by the subcortical thalamus-cerebellum circuit; sleep disturbances were characterized by the frontal cortex-cerebellum-occipital cortex circuit; and appetite disturbance was characterized by the subcortical thalamus-hippocampus-dorsal raphe nucleus circuit.

Fig. 4

Six spatiotemporal perfusion circuits for representative neuropsychiatric symptoms.

DISCUSSION

In this study, we found for the first time that a prolonged ATT, increased CBF, and unchanged volumes of the right ATN could serve as neuroimaging markers for detecting NPS in patients with AD. This finding unveiled the microvascular flow disturbance characteristics of NPS and highlighted the potential regulatory role of ATN in NPS, aiding in early detection and further exploration of the hemodynamic mechanism.

The time-encoded pCASL technique is a new and reliable ASL technique that is increasingly used by researchers and clinicians. It enables non-invasive measurement of brain perfusion and neurovascular function assessment at the tissue level because of its high labeling efficiency and SNR, low level of artifacts, and supplementation of other physiological parameters besides perfusion [29, 30]. Previous studies that aimed to quantify CBF in patients with AD typically used 1-delay pCASL and focused on the association between perfusion and cognition [31, 32]. However, in this study, we used the time-encoded pCASL technique to explore the spatiotemporal perfusion characteristics of NPS in patients with AD, which has rarely been studied. ATT is an important and interesting physiological parameter that represents the time taken for the labeled blood travelling from the feeding arteries to reach the microvasculature of the imaging volume [14]. Thus, prolonged ATT maps provide supplementary clinical information that assists in identifying collateral pathways and interpreting early perfusion images [33].

The ATN is an important subcortical structure that connects the hippocampus and fornix in the limbic circuit, contributing to higher cognitive functions, especially working memory processes [34]. Recent experiments have provided consistent causal evidence of the two-way functional significance of direct hippocampal-anterior thalamic interactions in spatial processing [35]. An observational study based on the Alzheimer’s Disease Neuroimaging Initiative database showed that the ATN volume was significantly lower in a population with late MCI and AD than in healthy participants [36]. A recent systematic review showed that neuronal loss in ATN could contribute to many neurological and neuropsychiatric conditions, including AD, frontotemporal dementia, schizophrenia, and Korsakoff’s syndrome [37]. Despite lacking evidence on a direct functional relationship between ATN and NPS, findings from an animal model demonstrated that many autism and schizophrenia risk genes were expressed in the ATN [38]. This finding and our results suggest a shared circuit mechanism between AD-related NPS and other psychiatric disorders, indicating a converging cellular-to-circuit mechanism underlying cognitive deficits and NPS.

In addition, we discovered that NPS exhibits high symptom heterogeneity, with depression, anxiety, and apathy being the most common sub-symptoms. However, we did not find, as expected, that a certain brain region exhibited opposite perfusion characteristics in different NPS sub-symptoms like anxiety or depression; in these different sub-symptoms, the role of a certain brain area was almost consistent, presenting positive or negative effects. Our study is the first to reveal that the cerebellum is a widely affected brain area in NPS, the CBF of which was positively associated with all the sub-symptoms in patients with AD except for disinhibition, irritability, and sleep disturbances. The cerebellum is increasingly recognized in AD for its role in modulation of cognition, but the evidence of the connection between the perfusion of the cerebellum and NPS sub-symptoms in AD is extremely limited [39]. A recent study found that lesions causing auditory hallucinations were associated with the dentate nucleus in the cerebellum [40]. Moreover, an observational study found that depression in patients with AD continuum was associated with a lower gray matter volume in the cerebellum [41]. Another observational study demonstrated that compared with patients with MCI but without depression, those with depression showed a significant correlation between the increased mean diffusivity of the cerebellum and the severity of depression [42]. Despite the limited evidence, these findings contribute to a better understanding of the regulatory characteristics of brain regions of different sub-symptoms and emphasize the important role of cerebellum perfusion in the NPS for further exploration.

Finally, this study identified six spatiotemporal perfusion characteristics using the ATT values of time-encoded pCASL to simulate the neurological circuit of a specific NPS sub-symptom. The anterior cingulate gyri-temporal pole/subcortical thalamus-cerebellum circuit was associated with delusion, euphoria, and apathy in patients with AD, consistent with previous evidence [10]. An observational study demonstrated that the severity of apathy was negatively correlated with functional connectivity of the anterior cingulate gyri within the default mode network in patients with amnestic MCI [43]. Another observational study also revealed that decreased hypothalamic precuneus/PCC functional connectivity was related to apathy in amyloid-verified AD [44]. We also found that the subcortical thalamus-temporal pole-frontal/parietal cortex circuit was associated with depression, anxiety, disinhibition, and irritability, which has been rarely described. Resting-state functional MR imaging data collected from 70 patients with major depressive disorder (MDD) showed that compared to healthy controls, patients with MDD demonstrated altered thalamo-cortical connectivity via a complex pattern of region-dependent hypo- or hyperconnectivity [45]. Another study demonstrated that a depressive episode and MDD shared common disrupted dynamic thalamo-cortical functional connectivity variability [46]. Direct evidence of other circuits of different sub-symptoms in patients with AD is lacking; however, network alterations across the cerebral cortex, subcortical regions, and cerebellum in these NPS sub-symptoms suggests a shared mechanism among different sub-symptoms, and between these NPS sub-symptoms in AD and other common psychiatric disorders, which was consistent with previous findings [47]. Our study found that the spatiotemporal perfusion characteristics based on time-encoded pCASL could be considered as a convenient and non-invasive technology for better understanding the regulatory mechanisms of NPS.

This study also has some limitations. First, we used a nested case-control study design based on the CIBL cohort; hence, some data were invalid because of mismatching, causing a relatively small sample size in the control group. Second, the ASL technique is sensitive to patient movement and requires at least 10 min for the labeling localization to acquire different vessels, making it unsuitable for uncooperative patients.

In conclusion, this study revealed that prolonged ATT and increased CBF of the right ATN can act as neuroimaging markers of NPS, thereby unveiling the microvascular flow disturbance characteristics of NPS in patients with AD. The perfusion of the cerebellum is the most widespread factor affecting NPS sub-symptoms. Furthermore, the spatiotemporal perfusion characteristics of different sub-symptoms indicate that time-encoded pCASL may be a reliable technique to explore the potential neural regulatory circuit of a specific NPS sub-symptom. Multicenter clinical validation is required to evaluate the external utility of these findings.

Footnotes

ACKNOWLEDGMENTS

The authors thank all participants of the present study as well as all members of the staff of the CIBL study for their role in data collection.

FUNDING

This study was supported by the National Natural Science Foundation of China (82071187, 81870821) and the National Key Research and Development Program of China (2021YFC2500100, 2021YFC2500103).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study are available on request from the corresponding author.

The supplementary material is available in the electronic version of this article: .

References

Bessey

, Walaszek

(2019) Management of behavioral and psychological symptoms of dementia. Curr Psychiatry Rep 21, 66.

Lyketsos

, Lopez

, Jones

, Fitzpatrick

, Breitner

, DeKosky

(2002) Prevalence of neuropsychiatric symptoms in dementia and mild cognitive impairment: Results from the cardiovascular health study. JAMA 288, 1475–1483.

Stella

, Laks

, Govone

, de Medeiros

, Forlenza

(2016) Association of neuropsychiatric syndromes with global clinical deterioration in Alzheimer’s disease patients. Int Psychogeriatr 28, 779–786.

Sanders

(2021) Cognition and neuropsychiatric symptoms in Alzheimer disease: A century-old problem. Neurology 97, 615–616.

Magierski

, Sobow

, Schwertner

, Religa

(2020) Pharmacotherapy of behavioral and psychological symptoms of dementia: State of the art and future progress. Front Pharmacol 11, 1168.

Creese

, Lunnon

(2022) Neuropsychiatric symptoms in AD: The search for mechanisms. Nat Rev Neurol 18, 639–640.

, Chiew

, Rosa-Neto

, Kandiah

, Ismail

, Gauthier

(2019) Brain metabolic dysfunction in early neuropsychiatric symptoms of dementia. Front Pharmacol 10, 1398.

Jenkins

, Wang

, Rosen

, Weintraub

(2022) A transdiagnostic review of neuroimaging studies of apathy and disinhibition in dementia. Brain 145, 1886–1905.

Boublay

, Schott

, Krolak-Salmon

(2016) Neuroimaging correlates of neuropsychiatric symptoms in Alzheimer’s disease: A review of 20 years of research. Eur J Neurol 23, 1500–1509.

10.

Chen

, Dang

, Zhang

(2021) Brain mechanisms underlying neuropsychiatric symptoms in Alzheimer’s disease: A systematic review of symptom-general and -specific lesion patterns. Mol Neurodegener 16, 38.

11.

Alsop

, Detre

, Golay

, Günther

, Hendrikse

, Hernandez-Garcia

, Lu

, MacIntosh

, Parkes

, Smits

, van Osch

, Wang

, Wong

, Zaharchuk

(2015) Recommended implementation of arterial spin-labeled perfusion MRI for clinical applications: A consensus of the ISMRM perfusion study group and the European consortium for ASL in dementia. Magn Reson Med 73, 102–116.

12.

Duan

, Zhou

, Balachandrasekaran

, Bhumkar

, Boraste

, Becker

, Kuller

, Lopez

, Gach

, Dai

(2021) Cerebral blood flow predicts conversion of mild cognitive impairment into Alzheimer’s disease and cognitive decline: An arterial spin labeling follow-up study. J Alzheimers Dis 82, 293–305.

13.

Chandra

, Dervenoulas

, Politis

; Alzheimer’s Disease Neuroimaging Initiative (2019) Magnetic resonance imaging in Alzheimer’s disease and mild cognitive impairment. J Neurol 266, 1293–1302.

14.

van Osch

, Teeuwisse

, Chen

, Suzuki

, Helle

, Schmid

(2018) Advances in arterial spin labelling MRI methods for measuring perfusion and collateral flow. J Cereb Blood Flow Metab 38, 1461–1480.

15.

Woods

, Chappell

, Okell

(2020) Designing and comparing optimized pseudo-continuous Arterial Spin Labeling protocols for measurement of cerebral blood flow. Neuroimage 223, 117246.

16.

Sun

, Wang

, Li

, Jiang

, Zhang

, Li

, Zhang

, Jia

, Chappell

, Xu

(2022) Potential diagnostic applications of multi-delay arterial spin labeling in early Alzheimer’s disease: The Chinese Imaging, Biomarkers, and Lifestyle Study. Front Neurosci 16, 934471.

17.

Jeong

, Kang

, Park

, Na

, Kim

, Yoon

, Song

, Chung

(2022) Regional cerebral blood flow correlates of neuropsychiatric symptom domains in early Alzheimer’s disease. Diagnostics (Basel) 12, 1246.

18.

Batail

, Corouge

, Combès

, Conan

, Guillery-Sollier

, Vérin

, Sauleau

, Le Jeune

, Gauvrit

, Robert

, Barillot

, Ferre

, Drapier

(2023) Apathy in depression: An arterial spin labeling perfusion MRI study. J Psychiatr Res 157, 7–16.

19.

Valotassiou

, Sifakis

, Tzavara

, Lykou

, Tsinia

, Kamtsadeli

, Sali

, Angelidis

, Psimadas

, Tsougos

, Papageorgiou

, Georgoulias

, Papatriantafyllou

(2021) Correlation of neuropsychiatric symptoms in dementia with brain perfusion: A 99mTc-SPECT-HMPAO study with Brodmann areas analysis. Curr Alzheimer Res 18, 970–983.

20.

McKhann

, Knopman

, Chertkow

, Hyman

, Jack

Jr , Kawas

, Klunk

, Koroshetz

, Manly

, Mayeux

, Mohs

, Morris

, Rossor

, Scheltens

, Carrillo

, Thies

, Weintraub

, Phelps

(2011) The diagnosis of dementia due to Alzheimer’s disease: Recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement 7, 263–269.

21.

Jack

Jr , Bennett

, Blennow

, Carrillo

, Dunn

, Haeberlein

, Holtzman

, Jagust

, Jessen

, Karlawish

, Liu

, Molinuevo

, Montine

, Phelps

, Rankin

, Rowe

, Scheltens

, Siemers

, Snyder

, Sperling

; Contributors (2018) NIA-AA Research Framework: Toward a biological definition of Alzheimer’s disease. Alzheimers Dement 14, 535–562.

22.

American Psychiatric Association(2013) Diagnostic and statistical manual of mental disorders. 5th ed (DSM-5). American Psychiatric Publishing, Arlington, VA.

23.

Leung

, Lam

, Chiu

, Cummings

, Chen

(2001) Validation study of the Chinese version of the neuropsychiatric inventory (CNPI). Int J Geriatr Psychiatry 16, 789–793.

24.

Katzman

, Zhang

, Ouang-Ya-Qu , Wang

, Liu

, Yu

, Wong

, Salmon

, Grant

(1988) A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol 41, 971–978.

25.

Buysse

, Reynolds

, 3rd Monk

, Berman

, Kupfer

(1989) The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Res 28, 193–213.

26.

Vellas

, Guigoz

, Garry

, Nourhashemi

, Bennahum

, Lauque

, Albarede

(1999) The Mini Nutritional Assessment (MNA) and its use in grading the nutritional state of elderly patients. Nutrition 15, 116–122.

27.

Chou

, Jiann-Chyun

, Chu

(2002) The reliability and validity of the Chinese version of the caregiver burden inventory. Nurs Res 51, 324–331.

28.

Rolls

, Huang

, Lin

, Feng

, Joliot

(2020) Automated anatomical labelling atlas 3. Neuroimage 206, 116189.

29.

van der Plas

MCE

, Teeuwisse

, Schmid

, Chappell

, van Osch

MJP

(2019) High temporal resolution arterial spin labeling MRI with whole-brain coverage by combining time-encoding with Look-Locker and simultaneous multi-slice imaging. Magn Reson Med 81, 3734–3744.

30.

Woods

, Schauman

, Chiew

, Chappell

, Okell

(2023) Time-encoded pseudo-continuous arterial spin labeling: Increasing SNR in ASL dynamic angiography. Magn Reson Med 89, 1323–1341.

31.

Wang

, Wang

, Li

, Wang

, Gao

, Lou

, Pfeuffer

, Zhang

, Zhu

, Li

, Chen

(2022) A diagnostic index based on pseudo-continuous arterial spin labeling and T1-mapping improves efficacy in discriminating Alzheimer’s disease from normal cognition. Front Neurosci 16, 974651.

32.

Fazlollahi

, Calamante

, Liang

, Bourgeat

, Raniga

, Dore

, Fripp

, Ames

, Masters

, Rowe

, Connelly

, Villemagne

, Salvado

; Australian Imaging Biomarkers and Lifestyle (AIBL) Research Group (2020) Increased cerebral blood flow with increased amyloid burden in the preclinical phase of Alzheimer’s disease. J Magn Reson Imaging 51, 505–513.

33.

Hernandez-Garcia

, Lahiri

, Schollenberger

(2019) Recent progress in ASL. Neuroimage 187, 3–16.

34.

Roy

, Zhang

, Aida

, Shen

, Skaggs

, Hou

, Fleishman

, Mosto

, Weninger

, Feng

(2022) Anterior thalamic circuits crucial for working memory. Proc Natl Acad Sci U S A 119, e2118712119.

35.

Nelson

AJD

, Kinnavane

, Amin

, O’Mara

, Aggleton

(2020) Deconstructing the direct reciprocal hippocampal-anterior thalamic pathways for spatial learning. J Neurosci 40, 6978–6990.

36.

Bernstein

, Rapcsak

, Hornberger

, Saranathan

; Alzheimer’s Disease Neuroimaging Initiative (2021) Structural changes in thalamic nuclei across prodromal and clinical Alzheimer’s disease. J Alzheimers Dis 82, 361–371.

37.

Perry

BAL

, Lomi

, Mitchell

(2021) Thalamocortical interactions in cognition and disease: The mediodorsal and anterior thalamic nuclei. Neurosci Biobehav Rev 130, 162–177.

38.

Roy

, Zhang

, Aida

, Choi

, Chen

, Hou

, Lea

, Skaggs

, Quay

, Liew

, Maisano

, Le

, Jones

, Xu

, Kong

, Sullivan

, Saunders

, McCarroll

, Wickersham

, Feng

(2021) Anterior thalamic dysfunction underlies cognitive deficits in a subset of neuropsychiatric disease models. Neuron 109, 2590–2603.e13.

39.

Ahmadian

, van Baarsen

, van Zandvoort

, Robe

(2019) The cerebellar cognitive affective syndrome-a meta-analysis. Cerebellum 18, 941–950.

40.

Kim

, Hsu

, Talmasov

, Joutsa

, Soussand

, Wu

, Rost

, Morenas-Rodríguez

, Martí-Fàbregas

, Pascual-Leone

, Corlett

, Fox

(2021) Lesions causing hallucinations localize to one common brain network. Mol Psychiatry 26, 1299–1309.

41.

Capogna

, Manca

, De Marco

, Hall

, Soininen

, Venneri

(2019) Understanding the effect of cognitive/brain reserve and depression on regional atrophy in early Alzheimer’s disease. Postgrad Med 131, 533–538.

42.

, Wang

, Lyu

, Wang

, Hu

, Guo

, Xu

, Hu

(2020) Alterations of white matter microstructure in subcortical vascular mild cognitive impairment with and without depressive symptoms. J Alzheimers Dis 73, 1565–1573.

43.

Joo

, Lee

, Lim

(2016) Apathy and intrinsic functional connectivity networks in amnestic mild cognitive impairment. Neuropsychiatr Dis Treat 13, 61–67.

44.

Chaudhary

, Zhornitsky

, Chao

, van Dyck

, Li

(2022) Hypothalamic functional connectivity and apathy in people with Alzheimer’s disease and cognitively normal healthy controls. J Alzheimers Dis 90, 1615–1628.

45.

Kong

, Qiao

, Liu

, Zhang

, Li

, Wang

, Li

, Su

, Li

, Yan

, Mitchell

, Si

(2018) Aberrant intrinsic functional connectivity in thalamo-cortical networks in major depressive disorder. CNS Neurosci Ther 24, 1063–1072.

46.

, Chen

, Cui

, Guo

, Pang

, Luo

, Yu

, Chen

, Gao

, Sheng

, Tang

, Zeng

, Jiang

, Gao

, He

, Chen

(2023) Shared and distinct patterns of dynamic functional connectivity variability of thalamo-cortical circuit in bipolar depression and major depressive disorder. , bhac. Cereb Cortex 14, 534.

47.

Wingo

, Liu

, Gerasimov

, Vattathil

, Wynne

, Liu

, Lori

, Faundez

, Bennett

, Seyfried

, Levey

, Wingo

(2022) Shared mechanisms across the major psychiatric and neurodegenerative diseases. Nat Commun 13, 4314.