Improved prediction of brain age using multimodal neuroimaging data

Abstract

Brain age prediction based on imaging data and machine learning (ML) methods has great potential to provide insights into the development of cognition and mental disorders. Though different ML models have been proposed, a systematic comparison of ML models in combination with imaging features derived from different modalities is still needed. In this study, we evaluate the prediction performance of 36 combinations of imaging features and ML models including deep learning. We utilize single and multimodal brain imaging data including MRI, DTI, and rs‐fMRI from a large data set with 839 subjects. Our study is a follow‐up to the initial work (Liang et al., 2019. Human Brain Mapping) to investigate different analytic strategies to combine data from MRI, DTI, and rs‐fMRI with the goal to improve brain age prediction accuracy. Additionally, the traditional approach to predicting the brain age gap has been shown to have a systematic bias. The potential nonlinear relationship between the brain age gap and chronological age has not been thoroughly tested. Here we propose a new method to correct the systematic bias of brain age gap by taking gender, chronological age, and their interactions into consideration. As the true brain age is unknown and may deviate from chronological age, we further examine whether various levels of behavioral performance across subjects predict their brain age estimated from neuroimaging data. This is an important step to quantify the practical implication of brain age prediction. Our findings are helpful to advance the practice of optimizing different analytic methodologies in brain age prediction.

Multimodal brain-age prediction and cardiovascular risk: The Whitehall II MRI sub-study

“…Despite poorer performance accuracy of the functional connectivity model, the fMRIbased brain-age deltas showed associations with the biomedical variables that were similar to the other modalities. In line with recent findings from UK Biobank [14,15], positive associations were found between brain-age deltas and diastolic blood pressure, alcohol intake, and stroke risk, concurring with previous WHII studies [32,50], and demonstrating that the brain age-delta measure reflects individual variation in neural aging processes [30]. The associations with biomedical variables were consistent across models (see Figure 5), indicating that while modality-specific brain age models may be informative in patient groups where tissue types are differently affected by disease [26,27,54,67,68], such models may be more closely related in healthy cohorts [14].…”

“…Although previous studies have suggested better prediction with multiple imaging modalities [13,14,30,54], the current study showed equivalent prediction accuracy between the multimodal model and the gray and white matter models. The exclusion of low-quality data improved the performance of the multimodal model, suggesting that established procedures for data quality control may have implications for model performance [39,55].…”

“…To obtain a direct comparison of β values, both the brain-age deltas and the clinical variables were standardized (subtracting the mean and dividing by the SD) before a series of multiple regressions were run. In order to adjust for a frequently observed bias in brain-age prediction leading to an overestimation of brain age in younger subjects and an underestimation of brain age in older subjects [14,15,30,51,52], chronological age was included as a covariate. Correction for multiple comparisons was performed using false-discovery rate correction [53].…”

“…Figure 4 shows the correlations between the multimodal and modality-specific brain-age deltas. The delta values were first corrected for age-bias [14,15,30,51,52] using linear models, and the residuals were used in the correlation analyses. Excluding low-quality data improved the multimodal model (z = 2.07, p = 0.04), but none of the other models, as shown in Table 4 Figure 2: The association between predicted and chronological age for each of the models is shown by the red line.…”

“…Functional MRI (fMRI) measures are indicative of brain network connectivity, which may change with advancing age [2,3]. Cardiovascular risk factors may influence these neural aging processes differently [26,27,30], and in a recent Whitehall II (WHII) MRI study using voxelwise analyses, allostatic load, metabolic syndrome, and multifactorial stroke risk predicted gray matter density measured decades later, while only cumulative stroke risk measured by the Framingham stroke risk score [31] predicted white matter integrity in terms of fractional anisotropy and mean diffusivity [32].…”