This study was a 12-month, randomized controlled trial conducted in community-dwelling older adults aged 65-85 years. The intervention was carried out between October 2015 and March 2017 in Iinan Town, Shimane Prefecture, Japan. Although the study was conducted prior to trial registration, it was retrospectively registered in the UMIN Clinical Trials Registry (registration no. UMIN000053587, date: February 9th, 2024) before submission. The delay in registration occurred due to a lack of awareness of prospective trial registration requirements at the time of study initiation. We acknowledge this as a limitation and affirm that all future clinical trials will be prospectively registered in accordance with international standards. The study was approved by the Ethics Committee of the Shimane University School of Medicine (approval no. 1940-2504). The study adhered to the ethical principles outlined in the Declaration of Helsinki. Prior to study participation, all volunteers provided written informed consent. In this study, 54 healthy volunteers participated, and underwent physical examinations including anthropometry, medical interview by a physician and blood biochemical tests. Volunteers completed a lifestyle questionnaire covering medical and medication history. The following volunteers were excluded from the study: volunteers with any medical disorders, including cardiac, hepatic, renal, gastrointestinal, respiratory disease, diabetes mellitus, neurological disorders, and osteoporosis, metabolic, cancer, endocrine, or hematological disorders; those eating BR daily; those consuming medications/supplements to treat bone disorders or osteoporosis and/or improve cognitive and mental function that could affect the results of the study; those with allergies and hypersensitivity; smokers. Of the 54 volunteers screened for eligibility, 44 participants (mean age, 73.1±5.6 years) were enrolled into the study and assigned to either the WR-intake group (n=22) or the UBR intake group (n=22) (Fig. 1). Participants were diagnosed by a clinician to ensure they had no neuropsychiatric or other disorders. No volunteers had participated in any other clinical studies within the past year. Group allocation was conducted by stratified random assignment, as described previously (24,25). The randomization code lists were prepared by an independent clinical research advisor and concealed from the investigators, participants, outcome assessors, and data analysts. All assessments and data analyses were performed in a blinded manner, and group assignments were disclosed only after completion of the study. Although participants were not informed which intervention (UBR or WR) was considered active, the noticeable differences in color, texture, and flavor between the two made complete blinding of participants infeasible. Participants in the WR-intake group received 200 g of WR daily for 12 months; those in the UBR-intake group received 100 g of UBR and 100 g of WR daily for 12 months. The rice intake was voluntary for the participants. During the intervention period, participants were advised to avoid other BR products and functional foods potentially influencing cognitive or bone health. This request was made to reduce confounding factors, and adherence was encouraged without disrupting participants' usual lifestyle. Participants' adherence to rice consumption was daily investigated using a self-administered questionnaire.

Both UBR and WR were supplied by NPO Satoyama Commission (Iinan-cho). To prepare UBR, BR was exposed to water at a hydrostatic pressure of 600 MPa for 5 s using a hydrostatic pressurizer (21). UBR and WR were both cultivated in Iinan-cho, Shimane, Japan, and produced from the same variety of rice. The nutrient content of WR and UBR was measured at Shimane Environment & Health Public Corporation and Shimane Institute for Industrial Technology and is summarized in Table I. Compared with WR, UBR was richer in lipids and dietary fiber and has a higher content of minerals, including magnesium (Mg), calcium (Ca), inorganic phosphorus (Pi), and iron (Fe). In addition, UBR is rich in vitamins B₁, B₆, and niacin, as well as bioactive substances such as GABA, inositol, and ferulic acid (Table I).

Height, waist circumference, and blood pressure were measured by trained nurses. Body weight and body fat were measured using a bioelectrical impedance analyzer (WB-150; TANITA Co.), as described previously (26). After these measurements, participants completed two self-administered questionnaires: a general lifestyle questionnaire, which included items on educational background and medical/medication history, and a brief diet history questionnaire, as described previously (22,26).

At baseline and after 12 months of the intervention, blood samples were drawn by nurses in the morning after confirming whether participants had fasted. Serum was separated from whole blood samples and stored at -80˚C until use. Serum biochemical parameters were measured using an automated clinical chemistry analyzer TBA-c16000 (TOSHIBA); serum HbA1c level was determined with high-performance liquid chromatography (HPLC) using an HLC-723G9 (TOSOH); and blood sugar was measured using the GA08III automatic analyzer (A&T Co.), as described previously (7).

Adverse events were monitored throughout the 12-month study period using structured monthly interviews and questionnaires. Participants were asked about specific symptoms including gastrointestinal issues (e.g., bloating, flatulence, diarrhea), allergic reactions, appetite changes, and bowel habit alterations. Any reported events were recorded and evaluated by the study physician to determine their relationship to the intervention.

The concentrations of monoamines, i.e., epinephrine (Epi), norepinephrine (NE), dopamine (DA), and serotonin (5-HT), in the serum were measured using a previously described HPLC method (22). Briefly, the serum was pretreated with a clean EG column (EICOM). The HPLC equipment consisted of an EICOM HTEC-500 (EICOM) equipped with a data processor (EICOM EPC-500 PowerChrom) and an automatic injector (EICOM M-514). Chromatographic separation was performed using an EICOMPAK CA-5ODS column (2.1x150 mm ID) linked to a precolumn (EICOM PREPAK PC-03-CA). PowerChrom software was used for data collection and analysis (EICOM). The mobile phase consisted of 0.1 M phosphate buffer (pH 5.7) containing 700 mg/l sodium 1-octanesulfonate, 12% methanol, and 50 mg/l EDTA disodium salt. The flow rate was set to 0.23 ml/min, and the applied potential was +450 mV with respect to an Ag/AgCl reference electrode. The column temperature was maintained at 25˚C.

Cognitive function was assessed using the Mini-Mental State Examination (MMSE) (27), the Revised Hasegawa's Dementia Scale (HDS-R) (28), the Frontal Assessment Battery (FAB) (29), and the Cognitive Assessment for Dementia, iPad version (CADi) (30). The MMSE is a cognitive function test consisting of 11 subitems. MMSE can be used to screen cognitive impairment, estimate the severity of cognitive impairment at a given point in time, track an individual's cognitive changes over time, and document an individual's response to treatment (27). A total MMSE score of 24 to 27 points is considered indicative of mild cognitive impairment (MCI), while a score below 23 suggests possible dementia (27). The HDS-R is commonly used for dementia screening in Japan and includes 9 items assessing orientation, memory, attention, and verbal fluency (28). Generally, a score below 20 is considered suggestive of dementia (28). The FAB, consisting of 6 subtests, is a cognitive test that incorporates several clinical assessments to screen for frontotemporal dementia, including S-word generation, similarities, Luria's test, grasp reflex, and the Go-No-Go test (29). The CADi consists of 10 items and is a useful tool for population screening for dementia and is known to correlate significantly with MMSE scores (30). In CADi, the time spent performing a task is also evaluated and used as an indicator of cognitive processing ability (30). Furthermore, mental condition, i.e., apathy and depression, was assessed using the Japanese version of the Starkstein apathy scale (31) and the Zung Self-Rating Depression Scale (SDS) (32), respectively. The Starkstein apathy scale is a subjective rating scale with scores ranging from 0 to 42; higher scores indicate less motivation and a greater apathy severity (31). The cutoff for the apathy scale is 16 points, and apathy is suspected in those who score 16 points or higher. The SDS is also a self-rating scale, comprising responses to 20 questions on a 4-point scale; a higher total score indicates a tendency toward depression (32). These tests are commonly used in clinical and interventional studies as they are relatively quick and can easily assess an individual's cognitive and mental function (12-15,22,25,26). All assessments were conducted by trained clinicians.

To assess bone status, we used an ultrasonic bone densitometer (Venus-α; Nihon Kohden, Tokyo, Japan) to perform quantitative ultrasound (QUS) of the right calcaneus, as previously described (23). The device uses a combination of ultrasound pulse reflection and ultrasound pulse transmission methods to measure the speed of sound (SOS) and the bone width of the calcaneus and subsequently calculates the bone area ratio (BAR). QUS offers a quick and noninvasive method for evaluating bone health without the radiation exposure associated with dual-energy X-ray absorptiometry (DXA). The BAR values obtained with this device have been reported to correlate significantly with bone mineral density (BMD) measurements by DXA at the lumbar spine (r=0.77, P<0.01) and calcaneus (r=0.83, P<0.01) (33). For each participant, the BAR of the right calcaneus was measured at least three times and converted to the percentage of the young adult mean (%YAM), using reference data from Japanese adults aged 20-44, in which 100% represents the average bone density for that age group. %YAM is widely used in clinical practice in Japan as an index of bone status and allows for standardized comparisons across individuals (34,35).

The sample size was calculated using PS: Power & Sample Size Calculation Software Version 3.1.2 (Vanderbilt University, Nashville, TN, USA). The primary outcomes were the change in MMSE scores after 12 months of intervention. Based on data from a previous randomized controlled trial involving a similar population (15,22), a mean difference of 1.4 points (SD=2.0) between groups was expected. Assuming two-sided α=0.05 and 90% power, a minimum of 20 participants per group was required using an independent t-test. To account for an anticipated dropout rate of 10%, we enrolled 22 participants per group.

All statistical analyses were conducted using IBM SPSS Statistics for Windows, version 26.0 (IBM Corp.). A per-protocol approach was adopted for efficacy analyses. Data were first assessed for normality using the Shapiro-Wilk test. All variables met normality assumptions and are presented as mean ± standard error of the mean (SEM). The primary outcomes were the MMSE score and Starkstein apathy scale score at 12 months. Secondary outcomes included other cognitive measures (CADi, HDS-R, FAB), additional mental health indicators (SDS), and bone health index (%YAM). To assess between-group differences at 12 months while controlling for potential baseline imbalances, analysis of covariance (ANCOVA) was employed. In the ANCOVA models, the 12-month score was entered as the dependent variable, group (UBR vs. WR) as the fixed factor, and baseline score and age as covariates. To examine time-by-group interaction effects across baseline and 12 months, two-way repeated measures ANOVA were used. When a significant interaction was found, post hoc comparisons were performed using Tukey's test. Between-group comparisons of change scores (Δ=12 months-baseline) were conducted using independent samples t-tests as supplementary analyses. While change scores are reported for descriptive purposes, interpretation of primary effects was based on ANCOVA results. Effect sizes for between-group comparisons were calculated using Cohen's d, with thresholds of 0.2, 0.5, and 0.8 representing small, medium, and large effects, respectively. Partial correlation analyses were conducted to explore the associations among cognitive, mental, and bone-related outcomes at 12 months, adjusting for age and baseline scores. All tests were two-tailed, and statistical significance was set at P<0.05.

Fig. 1 shows a flow diagram of this intervention study. During the 12-month intervention, four subjects (two in the WR-intake group and two in the UBR-intake group) retired for personal reasons (Fig. 1). Therefore, the 12-month intervention trial was completed by twenty participants both in the WR-intake group and UBR-intake group (Fig. 1). Participants demonstrated excellent adherence to the intervention protocol for 12 months (UBR: 95.5%, WR: 96.8%). Data from general questionnaires on lifestyle habits and medical/medication history showed no significant differences during 12 months of intervention (data not shown). Adverse effects disturbing participants' daily lives such as palpitation, anorexia, diarrhea, constipation, allergic reactions, and irritated stomach were not observed in both groups. The BDHQ survey showed no significant difference in mean dietary nutrient intake between the UBR-intake and WR-intake groups at 12 months (data not shown).

Table II shows the values for body composition, blood pressure, blood biochemistry, and serum monoamine levels of the participants at baseline and at 12 months after the intervention, as well as the changes during the intervention period. At baseline, there were no significant differences between the WR-intake and UBR intake groups in body weight, height, fat mass, BMI, waist circumference, or blood pressure (Table II). Similarly, no significant differences were observed between the groups for blood biochemical variables, aspartate transaminase (AST), alanine transaminase (ALT), γ-glutamyl transpeptidase (γ-GTP), albumin (ALB), total cholesterol (T-cho), triglyceride (TG), blood urea nitrogen (BUN), creatinine (CRE), LDL-cholesterol (LDL-C), HDL-cholesterol (HDL-C), blood sugar, or HbA1c levels at baseline (Table II). After 12 months of the intervention, the two groups did not differ in height, body weight, body fat mass, BMI, waist circumference, or blood pressure (Table II), and there were no significant differences in blood biochemistry parameters, blood sugar, or HbA1c. Furthermore, for each parameter, the change from baseline to 12 months was not significantly different (Table II). Serum EPi, NE, DA, and 5-HT levels did not differ between the groups at baseline, and long-term UBR intake had no effect. From baseline to 12 months, D5-HT levels were slightly increased in the UBR intake group, but the change was not statistically significant (P=0.081, Table II).

Adverse events were monitored monthly using structured interviews. Participants were asked about common symptoms potentially associated with dietary fiber intake, such as bloating, diarrhea, and allergic reactions. No participants reported any adverse symptoms during the 12-month study period. These findings, together with the absence of unfavorable changes in clinical parameters (e.g., liver enzymes, kidney function, blood pressure), support the safety of long-term UBR intake.

MMSE, CADi, HSD-R, and FAB were used to assess participants' cognitive function, and the results were summarized in Table III. In this study, all participants had MMSE scores of 24 or higher, Apathy scores of 16 or lower, and SDS scores of 40 or lower during the intervention period, suggesting that they were not suffering from dementia, apathy, or depression. At baseline, there were no significant differences between the groups for all measures of cognitive outcomes (Table III). At 12 months, MMSE and CADi total scores in the UBR-intake group were higher than in the WR-intake group. A two-way ANOVA revealed a significant group x time interaction for MMSE total scores, F(1,78)=7.67, P=0.021. Although the mean change in MMSE from baseline to 12 months was not statistically significant between the groups (P=0.142), ANCOVA adjusting for baseline value and age revealed that the 12-month MMSE score was significantly higher in the UBR group than in the WR group [F(1,36)=4.21, P=0.046, Cohen's d=0.72]. Sensitivity analyses using a simple comparison without covariates (P=0.034) and an alternative ANCOVA adjusting for age and sex (P=0.040) confirmed these results (Table SI). The absolute difference of 1.4 points, while statistically significant, was below the commonly cited in Minimal Clinically Important Difference (MCID) of 2-3 points, suggesting cognitive preservation rather than true improvement. Subitem analysis of the MMSE revealed no significant between-group difference in the change in ‘Recall 3 words’, a measure of recent memory (P=0.098, Table III). Regarding CADi, significant group x time interactions were observed for total score [F(1,78)=5.59, P=0.048] and execution time [F(1,78)=16.46, P=0.014]. At 12 months, CADi total score increased (P=0.038) and execution time decreased (P=0.008) in the UBR group compared to WR (Table III). The change in execution time (ΔCADi time) was significantly greater in the UBR group (Fig. 2A, P=0.026), and subitem analysis revealed a significant difference in ‘Delayed Recognition’ (Fig. 2B, P=0.027). For apathy, ANCOVA adjusting for baseline score and age showed a significantly lower 12-month score in the UBR group [F(1,36)=7.92, P=0.008; Cohen's d=0.89]. A two-way ANOVA also revealed a significant group x time interaction [F(1,78)=11.13, P=0.007], and Δapathy scores were significantly lower in the UBR group compared to the control (Fig. 2C, P=0.015), suggesting a potential benefit of UBR intake on motivational state. SDS values showed a non-significant trend toward reduction in the UBR group at 12 months (P=0.142), and between-group differences in change scores were not significant (P=0.294). HDS-R and FAB scores showed no significant differences. Sensitivity analyses for apathy and CADi confirmed consistent results across models (P=0.010-0.012 for apathy; P<0.05 for CADi), with moderate to large effect sizes (apathy: d=0.89; CADi: d=0.75-1.02). For HDS-R, FAB, and SDS, exploratory sensitivity analyses revealed small to moderate effect sizes (e.g., HDS-R: d=0.38) despite non-significant P-values (Table SII).

Table IV presents the values of SOS and %YAM at baseline and after the 12-month intervention, along with the corresponding changes over time. At baseline, no significant differences were observed between the UBR and WR groups in SOS or %YAM. After 12 months, both SOS and %YAM values were higher in the UBR group compared to the WR group. A two-way ANOVA revealed a significant main effect of group for both SOS [F(1,78)=4.22, P=0.046] and %YAM [F(1,78)=4.64, P=0.043], indicating that participants in the UBR group maintained significantly greater bone status than those in the WR group. No significant group x time interaction effects were found for either SOS or %YAM (P>0.05). Post hoc comparisons showed that the %YAM at 12 months was significantly higher in the UBR group compared to the WR group (P=0.032). Similarly, SOS was significantly greater in the UBR group at 12 months (P<0.05). No significant difference in calcaneal bone width was detected between groups (data not shown). Fig. 3 illustrates the changes in SOS (ΔSOS) and %YAM (Δ%YAM) from baseline to 12 months. Both ΔSOS (P=0.035) and Δ%YAM (P=0.034) were significantly greater in the UBR group compared to the WR group, further supporting a potential benefit of UBR intake on bone health in older adults.

Partial correlation analyses adjusting for age and baseline values revealed significant associations between cognitive and mental health indicators. After 12 months of intervention, MMSE score had significant negative correlations with the apathy score (Fig. 4A, r=-0.513, P=0.001) and SDS (Fig. 4B, r=-0.472, P=0.002). A strong positive correlation was observed between apathy and depression levels (Fig. 4C, r=0.769, P<0.0001). Moreover, %YAM had a significant negative correlation with CADi time (Fig. 4D, r=-0.283, P=0.046) and a positive correlation with CADi total (Fig. 4E, r=0.369, P=0.021). The %YAM had significant negative correlations with SDS (Fig. 4F, r=-0.439, P=0.005) and apathy score (Fig. 4G, r=-0.331, P=0.039) after 12 months of intervention. The apathy score showed a significant negative correlation with MMSE subitem ‘Recall 3 words’ (Fig. 4H, r=-0.499, P<0.001), and a significant positive correlation with CADi time subitem ‘Delayed Recognition’ (Fig. 4I, r=0.372, P=0.019). A full matrix of partial correlation coefficients is presented in Table SIII.