MMSE Cutoff Discriminates Hippocampal Atrophy: Neural Evidence for the Cutoff of 24 Points

The cutoff of 24 points on the Mini-Mental State Examination (MMSE) is used to detect cognitive impairment (CI).1, 2 However, the underlying neural mechanism of the value is poorly understood. A previous quantitative study revealed that older adults who scored less than 24 points on the MMSE show whole-brain atrophy and extensive white-matter lesions.3 Assessing detailed structural brain changes in CI older community-dwelling adults with no dementia may help to explain their high risk of progression to dementia and can provide evidence of the cutoff for screening early stages of CI. This study aimed to evaluate cerebral structural differences between older adults with and without CI defined by the traditional MMSE cutoff in a sample of community-dwelling non-demented older adults. This was a cross-sectional analysis of baseline data from the Takashimadaira Study, an ongoing cohort study.4 The Tokyo Metropolitan Institute of Gerontology Ethical Review Board approved the study and written informed consent was obtained at enrollment. Among older adults living in community independently who participated in a health checkup, those who scored below 24 points of the MMSE underwent an additional magnetic resonance imaging (MRI) assessment and were considered as participants with significant CI (CI group). The participants who scored above 23 points were randomly recruited for MRI assessment as the controls (non-CI group). The participants were free of dementia, depression, and activities of daily living problems. Magnetic resonance structural image data (3D T1-turbo field echo images) were processed using Statistical Parametric Mapping software 12 in MATLAB R2020a (Mathworks, Natick, MA) to perform voxel-based morphometry. To detect differences in gray matter volume between the CI and non-CI groups, voxel-wise two-sample t-test adjusted for age, gender, education level, number of comorbidities, presence of instrumental activities of daily living (IADL) disability, and scores of the Geriatric Depression Scale were performed. We performed family-wise error (FWE) correction at the voxel level using a significance threshold of P less than .05 for corrected results and applied a cluster-level threshold of k > 38 voxels. We also analyzed regions of interest (ROIs) to identify the degree of atrophy. ROIs were obtained with FreeSurfer software5 (version 6.0.0). Fifty-six and 94 older adults (mean age 79.5 ± 4.7 years and 79.2 ± 4.8 years, P = .733; 58.9% and 53.2% females, P = .494) with high functionality (5.4% of CI and 1.1% of non-CI adults showed IADL disabilities, P = .114) were included in the CI and non-CI groups. The CI group showed significantly lower gray-matter volume (476.9 ± 51.7 and 510.8 ± 55.9 mL, P < .001) and fewer years of education (11.3 ± 3.1 and 12.4 ± 3.1 years, P < .001) than the non-CI group. Voxel-wise two-sample t-test revealed clusters of significantly lower volume in the cortical region in the medial temporal lobe (Brodmann area 35) of the left and right hemispheres in the CI group compared to the non-CI group. This region includes the parahippocampus and the hippocampus (Figure 1). Furthermore, the cluster of the right hemisphere was larger than that of the left hemisphere. ROI analysis revealed 10.1% and 11.5% smaller parahippocampal volumes for the left and right hemispheres in the CI group compared to the non-CI group. Similarly, the hippocampal volumes of the left and right hemispheres in the CI group were 8.1 and 7.6% smaller than in the non-CI group. Voxel-based morphometry analysis identified a significantly decreased hippocampal volume of the left and right hemispheres in the CI group. There were no significant differences in other brain areas, indicating that structural changes in participants with cognitive impairments were limited to the hippocampus. Although hippocampal atrophy may be an early sign of mild CI and pathological aging (e.g., Alzheimer's disease),6, 7 execution of complex goal-directed activities (i.e., IADL) may rely on the integrity of specific areas of the frontal lobe.8, 9 This finding may thus explain why many older adults scoring less than 24 points on the MMSE live independently in the community, despite being considered a high-risk population of progression to dementia. Because of possible sampling bias, caution should be exercised when interpreting the results, although the neural difference detected in the present study is plausible. In conclusion, our results indicate that the cutoff of 24 points on the MMSE discriminates hippocampal atrophy, an early sign of pathological aging (e.g., neurodegenerative dementia and cerebral small vessel disease resulting in vascular CI). The present findings provide support to epidemiological evidence of screening for cognitive impairment using the cutoff value and indicate its capability for capturing pathological changes in the brain. This study was supported by the Grand from Bureau of Social Welfare and Public Health (Tokyo Metropolitan Government), Grand from Health and Welfare Bureau for the Elderly (Ministry of Health, Labour and Welfare), the Grant-in-Aid for JSPS fellows (26–7,168) and the Grant-in-Aid for Young Scientists (A). The authors have no conflicts of interest. None.