Introduction

It has been reported that approximately 70–90% or more of patients with dementia show one or more behavioral and psychological symptoms of dementia (BPSD)1, which is an important problem because it increases the burden on not only the patient, but also their families and carers. Non-pharmacological therapy is the first choice of treatment for BPSD, with pharmacological therapy being the second choice while minimizing side effects2. The lack of established efficacy of pharmacological therapy may be partly due to the fact that the pathophysiology and neural substrate of BPSD have not been fully elucidated.

The Nakayama study is a complete enumeration survey of community-dwelling Japanese older persons in the town of Nakayama, located about 27 km south of the center of Matsuyama City, Ehime prefecture, Shikoku Island, Japan, with the aim of understanding the actual situation of dementia, which has been ongoing since 19973. We have reported the prevalence of dementia in older people3, its trends over time4, the prevalence of BPSD5, the association between caregiver burden and BPSD6, and factors associated with cognitive decline7,8. In the first Nakayama study in 1997, 88% of patients with dementia showed sone or more symptoms of BPSD, and the most common symptom was apathy/indifference5. In the same first survey participants, delusion, agitation/aggression, apathy/indifference, irritability/lability, and aberrant motor behavior showed correlations with caregiver burden6. Thus, we have long noted that patients with dementia present with a high frequency of BPSD, adding to their care burden.

The Neuropsychiatric Inventory (NPI) is an objective measure of behavioral and psychological symptoms in patients with dementia9. In recent years, several reports have examined the neurobiological underpinnings of BPSD in patients with dementia10,11,12,13,14. Some studies have focused on the high frequency of apathy in AD patients and examined it in detail, with one study based on MRI and positron emission tomography (PET) with11C-Pittsburgh compound-B image analysis showed that amyloid-β deposition in bilateral frontal cortex correlates with apathy in AD patients11, and one study based on PET with11C-pyrifinyl-butadienyl-benzothiazole 3 image analysis showed that apathy severity in AD patients was associated with tau pathology in the orbitofrontal cortex10. Furthermore, the association between apathy and the insular cortex has also been widely reported, including negative correlations between NPI scores and blood flow in the bilateral orbitofrontal cortex, left putamen, left nucleus accumbens, left thalamus, and bilateral insular cortices in AD patients with apathy14, as well as reduced connectivity between the left insula and right superior parietal cortex in AD patients with apathy13. However, one report indicated that apathy was not associated with cortical atrophy12; thus, the results have been inconsistent, because of differences in methods and covariates between studies. In addition, these previous studies were hospital- or clinic-based. The lack of consistent results applies not only to apathy, but also to other BPSD and is an area in need of further research.

Thus, the aim of this study was to examine the association between BPSD and brain volume as an exploratory study using cross-sectional data from an older Japanese community.

Methods

Study population

The fifth Nakayama study was conducted as the Japan Prospective Studies Collaboration for Ageing and Dementia (JPSC-AD)15. JPSC-AD is a multisite, population-based, prospective, cohort study designed to enroll 10,000 community-dwelling older people. This fifth Nakayama study was conducted between 2016 and 2018, and 927 individuals participated in the study. Of these, 222 with missing MRI data, 3 with missing FreeSurfer data values, and 579 with missing NPI data were excluded, resulting in a total of 90 individuals with cognitive impairment (52 with mild cognitive impairment (MCI) and 38 with dementia) eligible for the study (Fig. 1). The study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) checklist.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Flowchart of patient selection in this study. MRI magnetic resonance imaging, NPI Neuropsychiatric Inventory, MCI mild cognitive impairment, AD Alzheimer’s disease, VaD vascular dementia, NPH normal pressure hydrocephalus.

This study was approved by the Medical Ethics Review Board of Ehime University (approval number 1610004) and Kyushu University’s Institutional Review Board for clinical research (approval number 686–06). Written, informed consent was obtained from all participants. This study was performed in accordance with the Declaration of Helsinki.

NPI

Behavioral and psychological symptoms in people with dementia were measured using the Japanese version of the NPI-Brief Questionnaire Form (NPI-Q)16,17, the questionnaire version of the NPI, a caregiver-based clinical scale that assesses 12 domains of neuropsychiatric symptoms (delusions, hallucinations, agitation/aggression, depression/confusion, anxiety, euphoria/illusion, apathy/indifference, loss of control, irritability/stability, abnormal motor behavior, abnormal night-time behavior disturbances, and appetite and eating disturbance). Caregivers were asked whether the behavior represented a change from that shown by the participant before the onset of dementia and was present during the past month. The informant then rated the behaviors. Scores from 1 to 4 were obtained for the frequency and from 1 to 3 for the severity of each behavior (the NPI score for each domain was the product of the frequency and severity subscores, maximum 12).

Measurement of other risk factors

Each participant completed a self-administered questionnaire including educational background, medical history, alcohol consumption, smoking, medical treatments (antihypertensive drugs, diabetes medications), and physical activity. Blood pressure was measured three times with an automated sphygmomanometer after at least five minutes of rest, and the mean of the three measurements was used for analysis18. All blood samples were analyzed at the central laboratory (LSI Medience Corporation, Tokyo, Japan) using the same equipment. Each risk factor was categorized based on the following clinical criteria and references. Low educational level was defined as ≤ 9 years19,20. Hypertension was defined by a blood pressure of 140/90 mmHg or higher or current use of antihypertensive medication. Diabetes mellitus was defined as follows: fasting glucose level ≥ 7.0 mmol/L, casual glucose levels ≥ 11.1 mmol/L, or HbA1c ≥ 6.5%, using the 2010 American Diabetes Association (ADA) criteria21, and/or current use of glucose-lowering agents. Serum total cholesterol levels were measured by an enzymatic method. The body mass index (kg/m2) was measured as an indicator of obesity. Obesity was defined as a body mass index ≥ 25 kg/m222. Chronic kidney disease was defined as an estimated glomerular filtration rate < 60 ml/min/1.73 m223, which was calculated using the modified Chronic Kidney Disease Epidemiology Collaboration Study equation multiplied by the Japanese coefficient of 0.81324,25. ECG abnormalities were defined as ST depression (Minnesota codes, 4–1, 2, 3), left ventricular hypertrophy (3–1), or atrial fibrillation (8–3)15. Regular exercise was defined as any physical activity performed for at least 30 min twice a week over the most recent year or longer. Alcohol consumption and smoking were classified as current or not. The apolipoprotein E (APOE) ε4 carrier state was defined as the presence of the ε2/ε4, ε3/ε4, or ε4/ε4 alleles. Depression was diagnosed according to the 'Diagnostic and Statistical Manual of Mental Disorders, Third Revised Edition26.

Diagnosis of dementia

Dementia was diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, Third Revised Edition26. The diagnosis of AD was made based on the following criteria: the National Institute of Neurological and Communicative Disorders and Stroke-Alzheimer’s Disease and Related Disorders Association criteria27. The diagnosis of vascular dementia (VaD) was based on the following criteria: the National Institute of Neurological Disorders and Stroke-Association Internationale pour la Recherche et l’Enseignement en Neurosciences criteria28. MCI was diagnosed according to Petersen’s criteria29. The diagnosis of dementia was made by the JPSC-AD study team’s expert psychiatrists and neurologists using a standardized diagnostic system. If the members of the endpoint determination committee agreed, the diagnosis was confirmed; if not, an endpoint determination committee meeting was held to confirm the diagnosis through discussion. The diagnostic procedure has been previously reported15.

MRI analysis

Brain MRI equipment was set up according to the Alzheimer’s Disease Neuroimaging Initiative (ADNI) study30 and T1-weighted imaging (T1WI) parameters. Details of MRI are given in Supplementary Table 1.

Segmentation and volumetric measurements of cortical and subcortical brain structures were performed automatically using FreeSurfer software (http://surfer.nmr.mgh.harvard.edu), version 5.3. Total brain volume (TBV) was calculated from brain segment volumes excluding ventricles. Cortical segmentation was performed using the Desikan-Killiany atlas31. The estimated total intracranial volume (eTIV) of each subject was also calculated using the standard FreeSurfer processing pipeline by exploiting the relationship between the intracranial volume and the linear transformation to the atlas template32. Specifically, FreeSurfer estimates of each brain parameter were fitted with a linear regression model adjusted for age, age squared, and sex. If the residuals of an individual participant’s data were less than − 4 standard deviation (SD) units or greater than + 4 SD units from the linear regression, these data were considered extreme outliers. Extreme outliers in total eTIV and volumes of at least five brain regions were excluded.

Statistical analyses

For the clinical characteristics of the study participants, the values are shown as the means (standard deviation) or frequencies, as appropriate. Two-group comparisons of the NPI scores between the MCI and dementia patient groups and between the AD and VaD groups were assessed with Mann–Whitney U tests. Multiple regression analysis was performed for the association between scores of each psychological and behavioral symptom of the NPI and brain volume. Multivariable adjustment was performed for age, sex, low educational level, hypertension, diabetes mellitus, serum total cholesterol level, BMI, chronic kidney disease, ECG abnormalities, history of stroke, current smoker status, current drinker status, exercise habits, APOE ε4 carrier status, and depression (Tables 1,2). The same analyses were also performed with Mini Mental State Examination (MMSE) scores added as a covariate to these variables due to its strong correlation with brain volume (Supplemental Tables 2,3). Of the 90 participants selected for the analysis, 2 (2.2%) were excluded from the multivariable-adjusted analysis because of missing data for any covariate. Subgroup analyses were performed for each of dementia and dementia subtypes (AD (isolated) or VaD). Heterogeneity in the associations of BPSD score of the NPI with brain volumes between subgroups was tested by adding a multiplicative interaction term in the relevant statistical model. P values of less than 0.05 were considered significant. All analyses were conducted using the Statistical Package for the Social Sciences (version 23; SPSS Inc., Chicago, IL, USA).

Table 1 Apathy/indifference and total brain volume and each regional brain volume to estimated total intracranial volume ratios in all participants.
Full size table
Table 2 Appetite and eating disturbances and total brain volume and each regional brain volume to estimated total intracranial volume ratios in all participants.
Full size table

Results

Baseline characteristics of the study population

The clinical characteristics of the study population are summarized in Table 3. The age of the study population was 82.1 (5.9) years, and 55.6% were female; the MMSE score was 22.4 (4.5); of the 90 study patients, 52 (57.8%) had MCI, and 38 (42.2%) had dementia; of the 52 MCI patients, 12 had MCI due to AD, 22 had MCI due to cerebral vascular disease (CVD), 1 had MCI due to AD with CVD, 17 had other/unclassified MCI; of the 38 dementia patients, 25 had AD (isolated type), 9 had VaD, 2 had AD with VaD, 1 had AD with normal pressure hydrocephalus (NPH), and 1 had dementia caused by head trauma (Table 4).

Table 3 Clinical characteristics of the study participants (n = 90).
Full size table
Table 4 Mild cognitive impairment and dementia subtypes (n = 90).
Full size table

Frequency of BPSD of the NPI

The frequency of BPSD of the NPI in the study population was 44.4%. The frequency of BPSD in patients with dementia was 63.2%, significantly higher than that with MCI (30.8%). Apathy was also the most frequent BPSD of the NPI seen in patients with dementia. The frequency of apathy/indifference was significantly higher in patients with dementia than with MCI (Table 5, Fig. 2). There were no statistically significant differences in the frequency of each BPSD between MCI due to AD and MCI due to cerebral vascular disease and between AD and VaD. (Table 6, Figs. 3, 4).

Table 5 Frequency and percentage of behavioral and psychological symptoms of the study participants (n = 90).
Full size table
Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Frequency of behavioral and psychological symptoms in patients with dementia and MCI. Each NPI score for behavioral and psychological symptoms is shown separately. ap < 0.05.

Table 6 Frequency and percentage of behavioral and psychological symptoms of patients with MCI due to Alzheimer’s disease and MCI due to cerebral vascular disease and Alzheimer’s disease (isolated) and vascular dementia.
Full size table
Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Frequency of behavioral and psychological symptoms of dementia in patients with Alzheimer’s dementia and vascular dementia. Each NPI scores for behavioral and psychological symptoms of dementia is shown separately.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Frequency of behavioral and psychological symptoms of dementia in patients with MCI due to Alzheimer’s disease and MCI due to cerebral vascular disease. Each NPI scores for behavioral and psychological symptoms of dementia is shown separately.

Associations of NPI scores with brain volume

Next, the associations between the scores of each BPSD of the NPI and each brain volume/eTIV ratio were examined, and the significant associations are shown in Tables 5 and 6. The temporal cortical volume/eTIV ratio decreased significantly with increasing apathy/indifference scores (β = − 0.054, 95% CI − 0.104 to − 0.004, p value = 0.04). The insular cortical volume/eTIV ratio also decreased significantly with increasing apathy/indifference scores (β = − 0.013, 95% CI − 0.021 to − 0.005, p value = 0.002). And, the total brain volume/eTIV ratio decreased significantly with increasing appetite and eating disturbance scores (β = − 0.534, 95% CI − 1.005 to − 0.063, p value = 0.03). The parietal cortical volume/eTIV ratio also decreased significantly with increasing appetite and eating disturbance scores (β = − 0.107, 95% CI − 0.172 to − 0.041, p value = 0.002). The temporal cortical volume/eTIV ratio also decreased significantly with increasing appetite and eating disturbance scores (β = − 0.100, 95%CI = − 0.176 to − 0.023, p value = 0.012). The hippocampal cortical volume/eTIV ratio also decreased significantly with increasing appetite and eating disturbance scores (β = − 0.014, 95%CI = − 0.025 to − 0.003, p value = 0.016). The associations between apathy/indifference scores and the insular cortical volume/eTIV ratio and between appetite and eating disturbance scores and the parietal cortical volume/eTIV ratio, temporal cortical volume/eTIV ratio, and hippocampal volume/eTIV ratio remained significant even after FDR correction (q-value = 0.01, q-value = 0.01, q-value = 0.04, q-value = 0.04, respectively). The significant association between apathy/indifference scores and the insular cortical volume/eTIV ratio was also observed with MMSE scores included as an additional covariate (p value = 0.01) (Supplemental Table 2). The significant associations between appetite and eating disturbance scores and the parietal cortical volume/eTIV ratio, and temporal cortical volume/eTIV ratio were also observed with MMSE scores included as an additional covariate (p value = 0.007, p value = 0.047, respectively) (Supplemental Table 3). No other brain regions were significantly associated with each BPSD score of the NPI (data not shown).

In addition, subgroup analyses of the associations of apathy/indifference scores and appetite and eating disturbance scores of the NPI with brain volumes for dementia (MCI or dementia) and dementia subtypes (AD (isolated) or VaD) were performed. There was significant heterogeneity in the associations of appetite and eating disturbance scores with total brain volumes/eTIV ratio, parietal cortical volume /eTIV ratio, and insular cortical volume /eTIV ratio, by dementia (all p values for heterogeneity < 0.05) (Supplemental Tables 4, 5). There was no significant heterogeneity in the associations of apathy/indifference scores and appetite and eating disturbance scores of the NPI with brain volumes by dementia subtype (Supplemental Tables 6, 7).

Discussion

This study has three major findings.

First, the present study showed that apathy was a significant behavioral and psychological symptom in patients with dementia (Table 5, Fig. 2). The frequency of apathy in the NPI was the highest and is consistent with our previous study5. The fact that apathy was significantly more common in dementia than in MCI in the present study may be supported by the meta-analysis showing that apathy was associated with an increased risk of conversion to AD and all-cause dementia of patients with MCI33. There was no difference in the frequency of BPSD between AD and VaD (Table 6, Fig. 3). The fact that apathy was more frequent in VaD than AD in this study, but the difference was not significant, is partly due to the small sample size of VaD. In our previous study5, delusions and aberrant motor behaviors were significantly more frequent in AD than in VaD, but no significant differences were found for these in the present study. Differences between that study (about 25 years ago)5 and the present study include the availability of rehabilitation and other long-term care services, and the introduction of medications effective for BPSD, such as cholinesterase inhibitors and an NMDA antagonist. These changes may reduce delusions and aberrant motor behaviors in AD. We are continuing our community-based cohort study on dementia and BPSD. In the future, for example, if BPSD such as apathy improves with anti-dementia drugs such as cholinesterase inhibitors, it will be necessary to analyze anteriorly whether changes in the structure and function of these brain regions will occur.

Second, the present study showed a negative correlation between the severity of apathy and the volume of the insular cortex. The insular cortex is the brain region responsible for the expression of feelings and emotions34. There have been many recent reports of the neural basis of apathy including the structure and function of brain regions in patients with dementia10,11,12,13,14. Of them, the correlation between apathy and the insular cortex has been widely reported, including reports of a negative correlation between NPI scores and blood flow in several tissues, including bilateral insular cortices12, and reduced connectivity between the left insula and right superior parietal cortex13 in AD patients with apathy. On the other hand, some studies have found no association between apathy and grey matter volume10,11. In other words, the present study differs in that it correlated with brain volume, whereas previous studies suggested a link with changes in brain function and connectivity networks. From these reports and the present study, structural or functional changes of the insular cortex appear to be associated with apathy in AD patients. In order to clarify the mechanism of the association between apathy and the insular cortex, it is necessary to examine not only structural changes, as indicated by changes in volume, but also functional changes in brain regions in the future. The insular cortex has been shown to undergo subtle structural changes due to systemic inflammation35, and it is also a brain region that serves as a hub for immunoreception in the brain among brain regions affected by inflammation34. If there is a relationship between apathy and the insular cortex, the insular cortex in AD patients may be involved in immunoreception36. This is the reason that the adjacent temporal cortex is known to be vulnerable to systemic inflammation35,37. Whereas previous studies on the relationship between apathy in patients with cognitive decline and the volume of the insular cortex were hospital or clinic-based12,13, the present study was population-based and included potential confounders as adjustment factors that have not been addressed in depth in other studies.

Third, the more severe the eating behavioral abnormalities, the lower the total brain volume, parietal, temporal, and hippocampal volumes in MCI and dementia. The temporal and parietal cortices and hippocampus are brain regions whose atrophy is a useful finding in the diagnosis of AD38. The hippocampus is also a region closely involved in the regulation of food intake39. However, how these regions influence BPSD in people with dementia remains unclear. These brain regions (parietal cortex, temporal cortex, and hippocampus) that were associated with appetite and eating disturbance in the present study are distinct from the orbitofrontal plane and insula, which have been reported as the neural basis of abnormal eating behavior in FTD40,41. A previous study reported that eating-related problems are common across conditions causing dementia42, and in the present study, appetite and eating disturbance was found to a similar extent in each of the different disease groups, AD and VaD. Appetite and eating disturbance of the NPI can be broken down into a variety of symptoms, including decreased appetite, fluctuations in appetite, changes in eating behavior and altered preferences. In the present study, a change in the type of food he/she likes and gained weight were the most frequent symptoms of the appetite and eating disturbance sub-category of NPI. The next most frequent symptoms were fluctuations in appetite and the habitual eating behavioral abnormalities such as eating the same food for every meal. 7 out of 8 patients with appetite and eating disturbance of the NPI presented with one of these symptoms, mainly hyperorality. Therefore, the result in this study may indicate a relationship between hyperorality and specific brain regions. The breakdown of eating disturbance is more frequent in hyperorality, and amnesia is thought to be the main cause. It therefore makes sense that the parietal cortex, temporal cortex and hippocampus are involved. In the future, it is necessary to accumulate systematic research that carefully captures and organizes a wide variety of BPSD.

In the subgroup analyses, the severity of apathy was negatively correlated with the volume of the insular cortex and the hippocampus, irrespective of MCI or dementia (Supplemental Table 4). In contrast, there was significant heterogeneity in the associations of the severity of appetite and eating disturbance with brain volumes across between MCI and dementia (Supplemental Table 5). The severity of apathy and appetite and eating disorders were also negatively correlated with hippocampal volume, irrespective of dementia subtype (Supplemental Table 6, 7). These analyses suggest that the association between apathy and atrophy of brain regions (e.g. insula cortex and hippocampus) may be found regardless of dementia subtype or stage. However, the lack of a relationship between apathy and insular cortex in the MCI group, for example, makes it difficult to conclude that BPSD is associated with brain atrophy alone by this study and is one reason why the scientific community needs to conduct further research.

The strength of the present study is that it was community-based and that adjustment for potential confounders could be performed. However, there are some limitations. First, because it was a cross-sectional study, it was not possible to address the causal relationship between BPSD and brain volumes. Selection bias was also assumed because the analysis population was limited to those whose information could be obtained from their caregivers who accompanied them in order to take the NPI, as this was a venue survey. Importantly, due to the small sample size of this study, the analysis did not take into account the stage of the disease, which may affect brain atrophy, for each type of dementia. In addition, to generalize the results of the present study to other ethnic groups may be limited due to differences in lifestyles and backgrounds.

Conclusions

In patients with cognitive decline, there are correlations between apathy and atrophy of the insular cortex and between eating-related problems and atrophy of the whole brain, parietal cortex, temporal cortex, and hippocampus.