Sex-Specific regional brain activity and cognitive function in mild cognitive impairment: An rs-fMRI study

Abstract

Mild cognitive impairment (MCI) is widely recognized as an early stage of dementia. Epidemiological studies suggest that MCI is more prevalent in females than in males. Notably, there are sex differences in MCI-related brain changes. Resting-state functional magnetic resonance imaging (rs-fMRI) offers a valuable method for assessing brain activity during rest. This study aims to explore sex-specific regional brain activity in participants with MCI during resting states. 86 MCI participants (21 males and 65 females) and 107 normal controls (NCs) (38 males and 69 females) were included in the present study. Regional homogeneity (ReHo), degree centrality (DC), amplitude of low frequency fluctuations (ALFF), and fractional ALFF (fALFF) were used to assess brain activity. MCI females showed increased ReHo values in the right cerebellum inferior compared to NC females and MCI males. However, MCI males exhibited increased ReHo values in the left hippocampus compared to NC males and MCI females. ReHo values in the right cerebellum inferior were associated with visuospatial skills in MCI males, and language function in MCI females. Additionally, ReHo values in the left hippocampus were associated with attention function in MCI females but not in MCI males. In MCI participants, sex moderated the relationship between ReHo values in the right cerebellum inferior and cognitive function (visuospatial skills and language function), as well as the association between ReHo values in the left hippocampus and attention function. In conclusions, this study revealed sex differences in ReHo of right inferior cerebellum and left hippocampus in MCI, and the association between ReHo and cognitive impairment in MCI differs by sex. These sex-specific patterns of regional brain activity can aid in the development of sex-specific precision medicine.

Introduction

Mild cognitive impairment (MCI) is characterized by significant cognitive decline, particularly in memory, but without impairing daily activities [1, 2]. Often considered as an early stage of dementia [3–5], it carries a heightened risk of progression to dementia [6, 7]. A meta-analysis revealed that more than 15% of participants living in community settings worldwide have MCI [8]. Furthermore, studies found that the prevalence of MCI tends to rise with age [9–11]. Dementia places a substantial burden on both patients and nations. It is projected that between 2020 and 2050, Alzheimer’s disease (AD) and other dementias will impose a staggering cost of 14,513 billion international dollars (INT$) on the global economy [12]. The success rate of dementia drug development has been alarmingly low, at just 0.4% between 2002 and 2012 [13]. Early prevention and treatment of MCI are crucial.

Numerous studies have found that MCI is more prevalent in females than in males [14, 15], with females often experiencing a faster progression from MCI to AD [16, 17]. MCI presents differently between the sexes. Females typically outperform males in verbal tasks, while males tend to have an advantage in spatial tasks relative to verbal tasks [18, 19]. Furthermore, females with MCI are observed to experience a more rapid cognitive decline. A study utilizing 8-year longitudinal data from the Alzheimer’s Disease Neuroimaging Initiative-1 (ADNI-1) found that MCI females consistently had higher Alzheimer’s Disease Assessment Scale – Cognitive Subscale (ADAS-Cog) scores at each follow-up year compared to MCI males, indicating poorer cognitive performance [20].These differences can be attributed to various factors. For instance, estrogen is known to have a protective effect on cognitive function in females. However, this protective effect diminishes after menopause, leading to cognitive decline [21, 22]. Additionally, genetic factors like the apolipoprotein E (APOE) epsilon gene can have different impacts depending on sex. Cognitive decline tends to be more pronounced in females carrying the APOE gene compared to their male counterparts [23, 24].

Research has shown that MCI females tend to exhibit greater rates of brain atrophy compared to MCI males [25]. Interestingly, a less pronounced hippocampal atrophy has been associated with decreased odds of MCI development in females, a correlation that is not significantly evident in males [26]. In a study investigating longitudinal sex differences using serial MRI over 2-3 years, it was found that in the amnestic MCI group, males showed distinctly lower grey matter volumes over time in the bilateral thalamus, the bilateral caudate nucleus, and the right middle temporal gyrus, while females showed less precuneus grey matter volume over time [27]. Zhao et al. proposed that the functional connectivity between the sensory-motor network (SMN) and white matter (WM), as well as the dorsal attention network (DAN) and white matter, is more severely impaired in females with MCI compared to males [28]. Jordan et al. found that males with MCI demonstrated stronger connectivity between the hippocampus and the precuneus cortex, as well as the brain stem [29]. However, research on sex differences in regional brain activity in MCI is scarce. By understanding sex differences, more precise interventions can be developed for both males and females with MCI.

The objective of this study is to discern the sex-specific regional brain activity in participants with MCI during resting states, furthering our understanding of the role of sex in cognition and brain functionality, thereby aiding in the development of sex-specific precision medicine.

Methods

Participants

A total of 193 participants (21 male participants with MCI, 65 female participants with MCI, 38 male normal controls (NCs) and 69 female normal controls) were recruited from the Affiliated Brain Hospital of Guangzhou Medical University and from the community in Guangzhou. All the participants or their legal guardians provided written informed consent before taking part in our study. This study was approved by the ethics committees of The Affiliated Brain Hospital of Guangzhou Medical University.

The diagnostic criteria of MCI were based on the Peterson criteria [6] (with 1.5 SD below their agemates in cognitive scores and full score in activities of daily living). NC participants were age-matched, cognitively and physically healthy participants. The exclusion criteria were as follows: (1) neurodegenerative diseases such as Parkinson’s disease, frontotemporal dementia, and dementia with Lewy bodies; (2) other conditions that affect cognitive function, such as hypothyroidism, vitamin B12 or folic acid deficiency, syphilis infection; (3) mental disorders such as depressive disorders, bipolar disorders, schizophrenia, and substance dependence. Participants underwent structured interviews, comprehensive cognitive assessments and neuroimaging scanning on the same day.

Demographic and clinical measurements

Basic demographic information including sex, age, and years of education was collected for all participants. Additional clinical information such as Body Mass Index (BMI) and the 17-item Hamilton Depression Rating Scale (HAMD-17) [30] were recorded. The assessments were conducted by two trained professional psychiatrists who had successfully completed a concordance assessment.

Neuropsychological assessments

All participants underwent a comprehensive neuropsychological battery assessment. Their global cognitive functioning was assessed using the Mini-Mental State Examination (MMSE) [31]. Five specific cognitive domains were evaluated using a range of validated instruments including memory (the Auditory Verbal Learning Test recognition (AVLT-N6) [32] and the Logical Memory Test [LMT] [33]), language (the Boston Naming Test [BNT] [34], and the Verbal Fluency Test [VFT] [35]), executive function (the Stroop Color and Word Test [Stroop] [36]), visuospatial skills (the Rey-Osterrieth Complex Figure [ROCF] Test [37] and the Clock Drawing Test [CDT] [38]), and attention (the Trail-Making Test [TMT] [39], the Symbol-Digit Modality Test [SDMT] [40] and the Digit Span Test [DST] [41]).

MRI data acquisition

Magnetic resonance imaging (MRI) data were acquired by the Philips 3.0 T MR systems in The Affiliated Brain Hospital of Guangzhou Medical University (Philips, Achieva, Netherlands). Sagittal resting-state fMRI datasets of the whole brain were acquired in 8 min using a single-shot gradient echo-planar imaging (EPI) pulse sequence with the following parameters: TE = 30 ms, TR = 2000 ms, flip angle (FA) = 90°, number of slices= 33, slice thickness= 4 mm, matrix size = 64 × 64, and field of view (FOV) = 220 × 220 mm. During the entire duration of the scan, participants were instructed to lie down in the scanner, maintain their eyes closed without sleeping, try to hold their heads steady, and refrain from systematic thought.

Image preprocessing

Preprocessing of resting-state fMRI data was performed using the DPABI 4.3 software, which is based on Statistical Parametric Mapping (SPM12). The first 10 volumes were discarded to allow magnetization stabilization. The remaining images underwent slice-timing correction, realignment for head motion correction, and calculation of head motion parameters. Participants with > 2 mm maximum displacement, > 2^◦ rotation, or > 0.2 mm mean framewise displacement were excluded from further analysis. The realigned images were then spatially normalized to the Montreal Neurological Institute (MNI) EPI template and resliced to a voxel size of 3 × 3 × 3 mm³. The linear trend was removed, and nuisance covariates were then regressed out from each time series, including signals of white matter and cerebrospinal fluid as well as the Friston-24 parameters of head motion. To reduce the effect of low-frequency drifts and high-frequency noise, a bandpass filter (0.01 Hz < f < 0.1 Hz) was applied.

Regional homogeneity measurement

The regional homogeneity (ReHo) value was obtained by using the Kendall’s coefficient of concordance (KCC) between the unsmoothed time series of a given voxel and its nearest neighboring 26 voxels [42]. Z scores were applied to each ReHo map to derive the regional homogeneity (zReHo). Finally, the zReHo maps were spatially smoothed using a 6 mm FWHM Gaussian kernel.

Degree centrality measurement

The BOLD signal of each voxel was extracted and collected with every other voxel. The degree centrality (DC) value was determined by the number of correlations that exceeded 0.25 (r > 0.25) [43]. Then the DC metric maps were z-transformed and smoothed (6 mm × 6 mm × 6 mm full width at half the maximum Gaussian kernel).

Amplitude of low frequency fluctuations measurement

Previous preprocessed results before filter were used for ALFF analysis. For a given voxel, the time sequences were transformed to the frequency domain using a fast Fourier transform and then the square root of the power spectrum was computed. The amplitude of low frequency fluctuations (ALFF) value was defined as the average square root [44]. The standard ALFF value was calculated by subtracting the average ALFF value from each voxel’s ALFF value and then dividing this result by the standard deviation of the entire brain’s ALFF map. Then the standard ALFF maps were smoothed using a 6 mm FWHM Gaussian kernel.

Fractional ALFF measurement

The fractional ALFF (fALFF) is a refinement of the ALFF approach. The fALFF is calculated as the ratio of the power at each frequency in the low-frequency range (0.01–0.08 Hz) to the power across the entire frequency range (0–0.25 Hz) [45]. After the removal of the linear trend, the time series for each voxel was transformed into a frequency domain without the application of band-pass filtering. The square root of the power spectrum was then computed at each frequency. The subsequent procedures followed the same steps as those used in the above-mentioned ALFF measurement.

Statistical analyses

Statistical Package for Social Sciences 25.0 (SPSS, Chicago, IL, USA), R version 4.3.2 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria) and RStudio 2023.09.1 (RStudio Team. RStudio Inc., Boston, MA, USA) were used to perform statistical analysis and visualization. In demographic, clinical, and neurophysiological characteristics, continuous variables were expressed as mean ± standard deviation (SD). Analysis of Covariance (ANCOVA) was used to test for the differences of the four groups of normally distributed continuous variables, while the Kruskal-Wallis H test was used to analyze the continuous variables that did not obey the normal distribution. All tests were performed in a two-tailed format with a set significance threshold of 0.05. The post hoc analysis was performed using Bonferroni correction.

DPABI 4.3 software package was used to establish a statistical model to analyze the differences in ReHo, DC, ALFF, and fALFF of the four groups (MCI male, MCI female, NC male, and NC female). Mixed effect analysis was utilized among the four groups to detect areas of difference, by using age and framewise displacement (FD) as covariates. Multiple comparison correction was conducted using Gaussian Random-Field Theory (GRF), setting a voxel level threshold at p < 0.001 and a cluster level threshold at p < 0.05.

Inter-group differences in neuroimaging indicators were investigated using a univariate general linear model (GLM), with age and FD as control variables. Then, partial correlation analysis was used to compute the associations between neuroimaging indicators and neuropsychological scores, controlling variables included age, years of education, and FD. Furthermore, stepwise multiple linear regression analyses were used to explore which neuropsychological tests were most associated with neuroimaging indicators. Moderation analyses using PROCESS 4.0 in SPSS were performed to explore relationships between cognitive function and neuroimaging indicators, with sex as the moderator and age, years of education, and FD as covariates. A 95% confidence interval and 5000 bootstrap samples were used.

Results

Demography, clinic, and neuropsychological tests

The demographic, clinical, and neuropsychological information are shown in Table 1. Multi-group comparisons found differences in years of education, MMSE, AVLT-N6, LMT, BNT, VFT, Stroop-C, ROCF-copy, CDT, SDMT, DST (p < 0.05). There were no statistical differences in terms of age, BMI, HAMD, and TMT-A among the participants (p > 0.05). The post hoc analysis with Bonferroni’s correction was performed to confirm differences between groups. Differences between MCI males and NC females, or between MCI females and NC males, are not presented. NC males exhibited higher years of education than MCI males, and NC females exhibited higher years of education than MCI females. Significant differences were observed between MCI males and NC males in the following neuropsychological tests: AVLT-N6, BNT, and CDT, with MCI males scoring lower in all these tests. Similarly, significant differences were noted between MCI females and NC females in the neuropsychological tests: MMSE, AVLT-N6, LMT, BNT, VFT, ROCF-copy, CDT, and DST, with MCI females scoring lower in all these tests. A significant difference was also observed between NC males and NC females in the BNT, with NC males scoring higher. However, no significant differences were found between MCI males and MCI females.

Table 1.

Demographic, clinical and neurophysiological data of all participants.

	MCI Male (n = 21)	MCI Female (n = 65)	NC Male (n = 38)	NC Female (n = 69)	F/H	p	Posthocb
Age	67.67 ± 8.56	66.97 ± 6.95	67.03 ± 5.77	66.43 ± 5.97	0.384	0.944	—
Years of education	8.83 ± 3.71	8.92 ± 3.09	11.97 ± 3.08	10.64 ± 2.61	24.678	<0.001	Cå A; Då B
BMI	22.62 ± 2.13	23.43 ± 3.06	22.81 ± 2.49	22.13 ± 2.74	2.445	0.065	—
HAMD	2.33 ± 2.08	3.58 ± 4.45	1.87 ± 2.42	3.48 ± 4.30	4.237	0.237	—
Global cognition
MMSE	25.71 ± 2.31	25.23 ± 2.63	27.08 ± 2.23	27.17 ± 1.84	5.235	0.002	Då B
Memory
AVLT-N6	3.18 ± 2.07	4.56 ± 2.71	6.14 ± 2.21	7.05 ± 2.03	14.752a	<0.001	Cå A; Då B
LMT	5.12 ± 2.96	4.14 ± 2.29	5.66 ± 1.95	5.68 ± 2.23	3.684a	0.014	Då B
Language
BNT	20.59 ± 3.43	18.91 ± 3.09	24.54 ± 1.73	22.37 ± 2.43	25.859a	<0.001	Cå A,D; Då B
VFT	12.24 ± 3.23	12.63 ± 3.49	14.27 ± 3.53	16.38 ± 3.39	3.623a	0.014	Då B
Executive function
Stroop-C	101.41 ± 33.15	91.14 ± 29.36	81.33 ± 24.62	77.83 ± 19.10	3.453a	0.018	—
Visuospatial skill
ROCF-copy	23.62 ± 4.04	24.1 ± 6.91	28.33 ± 3.13	27.84 ± 3.84	3.804a	0.011	Då B
CDT	3.41 ± 0.62	3.11 ± 0.82	4.0 ± 0.00	3.97 ± 0.17	27.864a	<0.001	Cå A; Då B
Attention
SDMT	26.9 ± 10.88	30.37 ± 10.66	34.32 ± 10.25	36 ± 9.02	2.913a	0.036	—
DST	5.18 ± 1.63	5.46 ± 1.96	7.11 ± 2.50	6.75 ± 2.17	4.652a	0.004	Då B
TMT-A	54.35 ± 17.88	56.66 ± 18.61	48.29 ± 15.68	50.12 ± 15.50	1.363a	0.256	—

Bold means that the significant p values. P value meant the comparison among MCI males, MC females, NC males and NC females by one-way ANOVA or Kruskal-Wallis H test. Continuous variables are reported as mean ± standard deviation. Post-hoc multiple comparisons used the Bonferroni correction. Any differences between MCI males and NC females, or between MCI females and NC males, are not presented in Table 1.

MCI mild cognitive impairment, NC normal control, BMI Body Mass Index, HAMD Hamilton Depression Scale, MMSE Mini-Mental State Examination, AVLT-N6 Auditory Verbal Learning Test Recognition, LMT Logical Memory Test, BNT Boston Naming Test, VFT Verbal Fluency Test, Stroop The Stroop Color and Word Test, TMT Trail-Making Test, ROCF Rey-Osterrieth Complex, CDT Clock Drawing Task, SDMT Symbol-Digit Modality Test, DST Digit Span Test.

^a Age, years of education were included as control variables.

^b A: MCI male; B: MCI female; C: NC male; D: NC female.

Neuroimaging indicators differences among MCI males, MCI females, NC males, and NC females

ReHo was significantly different between the four groups. The mixed effect analysis showed that the differential brain regions were located in the right cerebellum inferior and left hippocampus (GRF voxel P < 0.001, cluster P < 0.05) (Table 2, Fig. 1). However, there was no significant difference in DC, ALFF, fALFF between the four groups.

Table 2.

Differences of ReHo among MCI male, MCI female, NC male, and NC female groups.

Brain regions	Cluster size (mm3)	Peak MNI coordinate			Peak intensity
		X	Y	Z
right cerebellum inferior	54	45	−63	−39	20.49
left hippocampus	74	−24	−30	−3	26.75

GRF corrected (voxel P < 0.001, cluster P < 0.05).

GRF Gaussian random field, ReHo regional homogeneity, MCI mild cognitive impairment, NC normal control, MNI Montreal Neurological Institute.

Fig. 1. Significant differences in ReHo among MCI male, MCI female, NC male, and NC female groups, as well as interactions of ReHo values among MCI, NC, male, and female groups.

(A) There were significant differences of ReHo in the right cerebellum inferior among the four groups; An interaction among the four groups was observed in the ReHo values of the right cerebellum inferior. Sex differences in ReHo values were found in the right cerebellum inferior. Within the female group, significant differences in ReHo values were identified between the NC and MCI groups. (B) There were significant differences of ReHo in the left hippocampus among the four groups; An interaction among the four groups was detected in the ReHo values of the left hippocampus. Sex differences in ReHo values were present in the left hippocampus for both the MCI and NC groups. Within the male group, significant differences in ReHo values were observed between the NC and MCI groups. GRF corrected (voxel P < 0.001, cluster P < 0.05). The color bar represents F-values in mixed effect analyses. GRF, Gaussian random field. ReHo: regional homogeneity; MCI: mild cognitive impairment; NC: normal control.

The univariate GLM showed that in the MCI group, there were sex differences in ReHo values in the right cerebellum inferior, but no such differences were observed in the NC group. Within the female group, differences in ReHo values in the right cerebellum inferior were identified between the NC and MCI groups, whereas no such differences were found within the male group. MCI females showed increased ReHo in the right cerebellum inferior compared to both MCI males and NC females. Additionally, both in the MCI and NC groups, there were sex differences in ReHo values in the left hippocampus. Within the male group, differences in ReHo values of the left hippocampus were identified between the NC and MCI groups, whereas no such differences were found within the female group. MCI males showed increased ReHo in the left hippocampus compared to both MCI females and NC males. Moreover, NC females showed increased ReHo in the left hippocampus compared to NC males (Fig. 1).

Correlation analyses

In MCI males, correlations were observed between the ReHo values in the right cerebellum inferior and both ROCF-copy (r = −0.733, p = 0.003) and CDT (r = 0.541, p = 0.043). However, in MCI females, these ReHo values showed a correlation with VFT (r = −0.290, p = 0.033) and CDT (r = 0.315, p = 0.020). In NC females, the ReHo values in the right cerebellum inferior were found to be correlated with SDMT (r = −0.312, p = 0.012). When considering the ReHo values in the left hippocampus, there was a correlation with AVLT-N6 (r = 0.614, p = 0.020) in MCI males, and with MMSE (r = −0.261, p = 0.040) and DST (r = −0.368, p = 0.006) in MCI females. In NC females, a correlation was observed with LMT (r = 0.328, p = 0.020). No correlation was found between ReHo values and any cognitive scores in NC males (Fig. 2). According to the stepwise multiple linear regression analysis, significant correlations were found between the ReHo values in the right cerebellum inferior and both ROCF-copy (B = −0.058 p < 0.001) and CDT (B = 0.175 p = 0.047) in MCI males (Adjusted R² = 0.631, p < 0.001), as well as with VFT (B = −0.031 p = 0.028) in MCI females (Adjusted R² = 0.068, p = 0.028). Additionally, the ReHo values in the left hippocampus were significantly correlated with DST (B = −0.034 p = 0.013) in MCI females (Adjusted R² = 0.090, p = 0.039) (Table 3).

Fig. 2. Partial correlation between cognition and ReHo values (control variables included age, years of education, and FD).

(A) VFT was associated with the ReHo values of the right cerebellum inferior in MCI females (r = −0.290, p = 0.033); (B) ROCF-copy was associated with the ReHo values of the right cerebellum inferior in MCI males (r = −0.733, p = 0.003); (C) CDT was associated with the ReHo values of the right cerebellum inferior in MCI males (r = 0.541, p = 0.043) and MCI females (r = 0.315, p = 0.020); (D) MMSE was associated with the ReHo values of the left hippocampus in MCI females (r = −0.261, p = 0.040); (E) AVLT-N6 was associated with the ReHo values of the left hippocampus in MCI males (r = 0.614, p = 0.020); (F) DST was associated with the ReHo values of the left hippocampus in MCI females (r = −0.368, p = 0.006). ReHo: regional homogeneity; FD: framewise displacement; MCI: mild cognitive impairment; NC: normal control; VFT, Verbal Fluency Test; ROCF, Rey-Osterrieth Complex; CDT: Clock Drawing Task; MMSE, Mini-Mental State Examination; AVLT-N6, Auditory Verbal Learning Test Recognition; DST, Digit Span Test.

Table 3.

Results of the stepwise multiple linear regression between cognition and ReHo values in MCI participants.

	B	Standard Error	t	P value	Lower	Upper	β
					95% CI	95% CI
right cerebellum inferior
MCI Male (n = 21)
ROCF-copy	−0.058	0.012	−4.764	<0.001	−0.085	−0.032	−0.726
CDT	0.175	0.080	2.180	0.047	0.003	0.347	0.332
Constant	0.933	0.417	2.235	0.042	0.038	1.828
Adjusted R2				0.631
p				<0.001
MCI Female (n = 65)
VFT	−0.031	0.014	−2.255	0.028	−0.058	−0.003	−0.291
Constant	0.783	0.178	4.396	<0.001	0.426	1.139
Adjusted R2				0.068
p				0.028
left hippocampus
MCI Female (n = 65)
DST	−0.034	0.013	−2.563	0.013	−0.061	−0.007	−0.327
Constant	−0.539	0.077	−7.013	<0.001	−0.693	−0.385
Adjusted R2				0.090
p				0.039

ReHo regional homogeneity, MCI mild cognitive impairment, ROCF Rey-Osterrieth Complex, CDT Clock Drawing Task, VFT Verbal Fluency Test, DST Digit Span Test.

Moderation analyses

Moderation analyses were conducted, with cognitive scores serving as predictor variables, ReHo values as outcome variables, and sex as the moderator. Age, years of education, and FD were included as covariates. In MCI participants, three moderation models were found to be significant. The relationships between VFT and the ReHo values in the right cerebellum inferior, as well as between ROCF-copy and the ReHo values in the right cerebellum inferior, were both moderated by sex (β = −0.058, p = 0.042 and β = 0.062, p = 0.011 respectively). Moreover, the relationship between TMT-A and the ReHo values in the left hippocampus was also moderated by sex (β = 0.007, p = 0.022). No significant moderating effect of sex was found in NCs (Fig. 3).

Fig. 3. Sex modulated the relationship between cognition and ReHo values in MCI participants.

(A) Conceptual and statistical model of the relationship between VFT and right cerebellum inferior moderated by sex; (B) Conceptual and statistical model of the relationship between ROCF-copy and right cerebellum inferior moderated by sex; (C) Conceptual and statistical model of the relationship between TMT-A and left hippocampus moderated by sex. MCI: mild cognitive impairment; VFT, Verbal Fluency Test; ROCF, Rey-Osterrieth Complex; TMT, Trail-Making Test.

Discussion

In this study, we observed sex differences in rs-fMRI among participants with MCI. Further, we investigated the relationship between cognitive function and brain regions exhibiting altered neuroimaging indicators, and delved into the influence of sex on this relationship. Our primary findings include: (1) Significant differences were observed in ReHo between the four groups. (2) Compared to NC females, MCI females exhibited increased ReHo values in the right cerebellum inferior, while compared to NC males, MCI males exhibited increased ReHo values in the left hippocampus. (3) Within the MCI group, females showed higher ReHo values in the right cerebellum inferior compared to males, while males showed higher ReHo values in the left hippocampus compared to females. (4) There were sex differences in the correlations between ReHo values and specific cognitive functions. (5) Sex moderates the relationship between ReHo values and cognitive functions in MCI participants.

Sex differences are evident in the progression of AD, with females exhibiting a higher prevalence of both MCI and AD [15, 46]. According to Holland et al., females demonstrated greater rates of brain atrophy compared to males at every stage, from healthy aging through to AD dementia [25]. Females diagnosed with MCI exhibit a more rapid decline in cognitive functions during follow-up [20]. Therefore, it is crucial to consider sex differences in research when studying MCI. Sex differences in brain ReHo have been documented in healthy participants [47–49]. Previous studies reported altered ReHo in cerebellum and hippocampus in MCI participants [50, 51], which is consistent with our results. Our study found that MCI females showed increased ReHo value in the right cerebellum inferior compared with NC females and MCI males. MCI males showed increased ReHo value in left hippocampus compare with NC males and MCI females. These findings indicate that there are sex-based differences in brain function among MCI participants, which may be associated with the prevalence and clinical characteristics of MCI males and females. This could potentially result in sex-specific dysfunctions in MCI.

The primary functions of the cerebellum include maintaining body balance and regulating muscle tone [52, 53]. Additionally, recent studies have discovered that the cerebellum also plays a significant role in cognitive function (visuospatial skills, language, and other cognitive domains) [54–56]. The cerebellum is rich in visuospatial information, and lesions in this region can lead to multidimensional visuospatial disturbances. Visuospatial and visuomotor deficits in preterm infants have been associated with cerebellar dysfunction [57]. Similarly, these deficits have been confirmed in adults with focal cerebellar infarctions [58]. The cerebellum also contributes to language comprehension [59]. For instance, a patient with a right cerebellar lesion was prone to making errors in all verbal fluency tasks [60]. Sex differences in cognitive performance and cerebellum have been observed [61, 62]. Males generally exhibiting superior spatial cognition and females demonstrating stronger verbal cognition [19]. In participants with MCI, females tend to better retain verbal abilities, whereas males perform significantly better on visuospatial tasks compared to verbal tasks [18]. A study found that After normalizing lobular volume to total cerebellar gray-matter volume females demonstrated significantly larger relative volume in Crus II while males had larger relative volume in H V and VIIIA/B [63]. Stuart et al. found that in adult human brain females showed stronger weighted degree than males in some cerebellar regions [64].Our study found a significant association between ReHo in the right inferior cerebellum and visuospatial skills in males with MCI. Additionally, a significant association was observed between ReHo in the same region and language abilities in females with MCI. These results suggest that the right inferior cerebellum may serve as a biomarker for diagnosing and treating MCI. Further research on the right inferior cerebellum is warranted.

The hippocampus is a crucial brain region responsible for memory encoding and storage, and its functional activity is closely related to attention. Focused attention contributes to the efficient encoding and consolidation of information [65]. Mariam et al. reported that attention can stabilize activity patterns in the human hippocampus [66]. Consistent with this, animal studies have shown that the persistence of firing fields in mouse hippocampal location cells depends on the animal’s attention to task-relevant markers in the test environment [67], further supporting the important modulatory role of attention in hippocampal activity. This study found that MCI males exhibited higher ReHo values in the left hippocampus compared to MCI females, however, such increased activity was not associated with attention performance. We speculate that higher ReHo values in the left hippocampus may represent a compensatory mechanism to counteract early functional decline, rather than directly contributing to attentional capacity. In contrast, MCI females showed a significant association between ReHo values in the left hippocampus and attention scores, indicating a closer link between hippocampal functional fluctuations and attentional regulation. Attention regulation relies on the coordinated activity of multiple brain networks rather than a single region. Prior studies have demonstrated that, in adults, females show stronger weighted degree in the hippocampus than males [64], reflecting stronger overall connectivity between the hippocampus and other brain regions. The hippocampus, together with the prefrontal cortex and thalamus, plays a critical role in attention regulation [68, 69], and females generally exhibit higher activity in the prefrontal cortex and limbic system, potentially enhancing this network’s function. Moreover, sex hormones may influence hippocampal function; for example, estradiol protects hippocampal neurons from Aβ-induced toxicity in animal models [70]. Postmenopausal and perimenopausal females experience reduced ovarian hormone levels, increasing their vulnerability to Alzheimer’s disease compared to age-matched males [71, 72]. These factors may explain the observed significant correlation between ReHo values in the left hippocampus and attention performance in MCI females, whereas in MCI males, higher ReHo values in the left hippocampus were not directly associated with attention.

ReHo values in the right inferior cerebellum and left hippocampus were found to be negatively correlated with performance on several neuropsychological tests in MCI females or MCI males, including the ROCF-copy which assesses visuospatial skills, the VFT which assesses language function, and the DST which assesses attention function. Performance across these cognitive domains declined as ReHo values in the right inferior cerebellum or left hippocampus increased, suggesting that heightened local neural activity homogeneity in these regions may reflect worsening brain damage. Interestingly, in males with MCI, higher scores on the CDT were observed with increased ReHo values in the right inferior cerebellum. Both the ROCF-copy and CDT are commonly used to assess visuospatial skills, but their associations with ReHo values in the right inferior cerebellum differ in MCI males. This discrepancy may be attributed to the ROCF-copy’s requirement for participants to memorize and replicate a complex set of patterns within a short timeframe, necessitating higher visuospatial skills and greater reliance on the cerebellum’s coordination and integration functions. In contrast, the CDT is relatively simpler, relying more on long-term memory and basic visuospatial skills.

The relationship between visuospatial skills and the ReHo values in the right inferior cerebellum was moderated by sex, as was the relationship between language function and the ReHo values in the same region. Additionally, the association between attention function and the ReHo values in left hippocampus was also moderated by sex. These suggests that visuospatial skills, language function and attention function interact with sex to influence ReHo values. Therefore, when implementing neuromodulation, it is essential to adopt different regulatory targets and parameters based on the sex and specific cognitive impairments of individuals with MCI to enhance the precision and effectiveness of the treatment.

It is crucial to recognize certain limitations when interpreting the current findings. First, this study employed a cross-sectional design, which precludes the observation of changes in the relationships between cognitive function and ReHo values in the right inferior cerebellum and left hippocampus in males and females with MCI over time. Future longitudinal studies are necessary to elucidate these relationships further. Second, the limited number of participants constrained our ability to conduct more granular subgroup analyses, such as examining sex differences by age group or distinguishing between amnestic and non-amnestic MCI subtypes. Third, pathological and other markers are lacking; therefore, our sample is based on the entire population, which may lead to a limited representation of subjects with MCI. Future research should aim to increase the sample size and include examination of pathological biomarkers to provide a more comprehensive understanding.

In summary, we investigated differences in brain activity between males and females with MCI. Our findings indicate sex-specific patterns of regional brain activity in MCI patients, with the right inferior cerebellum and left hippocampus identified as the primary regions showing these differences. MCI females demonstrated increased ReHo values in the right inferior cerebellum, while MCI males exhibited increased ReHo values in the left hippocampus. Additionally, we found that sex moderated the associations between visuospatial skills and language with the ReHo values in the right inferior cerebellum, as well as between attention function and the ReHo values in the left hippocampus. Mapping the abnormal ReHo patterns in the right inferior cerebellum and left hippocampus in MCI males and females provides deeper insights into the underlying pathology of MCI. Further detailed investigations into sex differences in MCI are warranted to enhance our understanding of these disparities.

Author contributions

QL conducted the study, collected and analyzed the data, and drafted the initial manuscript. BC conceptualized the study, interpreted the data, and contributed to manuscript revision. TS provided technical support. QW, DX, MY, GL, YZ, JL, SL, JL, KY, ZX, PG, and XZ assisted with data collection. XZ and YN critically reviewed and revised the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Guangzhou Municipal Key Discipline in Medicine (2025-2027), Guangdong Province Key Areas Research and Development Programs-Brain Science and Brain-Inspired Intelligence Technology (2023B0303010003), National Natural Science Foundation of China (No. 82371428,No. 82171533), Natural Science Foundation of Guangdong Province, China (2024A1515011035; 2025A1515011872), The Science and Technology Program of Guangzhou Liwan District (No.202201003), Brain Science and Brain-Like Intelligence Technology (2021ZD0201800), and Guangzhou Key ClinicaSpecialty (Clinical Medical Research Institute). The funding source had no role in the study design, analysis, or interpretation of data or in the preparation of the report or decision to publish. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the funders.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate

The study adhered to the Declaration of Helsinki and was approved by the Ethics Committee of the Affiliated Brain Hospital of Guangzhou Medical University. All participants provided written informed consent prior to participation. All methods were performed in accordance with the relevant guidelines and regulations.

Competing interests

The authors declare no competing interests.