| 1. |
- Dartora, Caroline, et al.
(författare)
-
A deep learning model for brain age prediction using minimally preprocessed T1w images as input
- 2023
-
Ingår i: Frontiers in Aging Neuroscience. - : Frontiers Media SA. - 1663-4365. ; 15
-
Tidskriftsartikel (refereegranskat)abstract
- Introduction: In the last few years, several models trying to calculate the biological brain age have been proposed based on structural magnetic resonance imaging scans (T1-weighted MRIs, T1w) using multivariate methods and machine learning. We developed and validated a convolutional neural network (CNN)-based biological brain age prediction model that uses one T1w MRI preprocessing step when applying the model to external datasets to simplify implementation and increase accessibility in research settings. Our model only requires rigid image registration to the MNI space, which is an advantage compared to previous methods that require more preprocessing steps, such as feature extraction. Methods: We used a multicohort dataset of cognitively healthy individuals (age range = 32.0–95.7 years) comprising 17,296 MRIs for training and evaluation. We compared our model using hold-out (CNN1) and cross-validation (CNN2–4) approaches. To verify generalisability, we used two external datasets with different populations and MRI scan characteristics to evaluate the model. To demonstrate its usability, we included the external dataset’s images in the cross-validation training (CNN3). To ensure that our model used only the brain signal on the image, we also predicted brain age using skull-stripped images (CNN4). Results: The trained models achieved a mean absolute error of 2.99, 2.67, 2.67, and 3.08 years for CNN1–4, respectively. The model’s performance in the external dataset was in the typical range of mean absolute error (MAE) found in the literature for testing sets. Adding the external dataset to the training set (CNN3), overall, MAE is unaffected, but individual cohort MAE improves (5.63–2.25 years). Salience maps of predictions reveal that periventricular, temporal, and insular regions are the most important for age prediction. Discussion: We provide indicators for using biological (predicted) brain age as a metric for age correction in neuroimaging studies as an alternative to the traditional chronological age. In conclusion, using different approaches, our CNN-based model showed good performance using one T1w brain MRI preprocessing step. The proposed CNN model is made publicly available for the research community to be easily implemented and used to study ageing and age-related disorders.
|
|
| 2. |
- Lindberg, Olof, et al.
(författare)
-
Atrophy of the posterior subiculum is associated with memory impairment, Tau- and Aβ pathology in non-demented individuals
- 2017
-
Ingår i: Frontiers in Aging Neuroscience. - : Frontiers Media SA. - 1663-4365. ; 9:SEP
-
Tidskriftsartikel (refereegranskat)abstract
- Alzheimer's disease (AD) is associated with atrophy of the cornu ammonis (CA) 1 and the subiculum subfield of the hippocampus (HC), and with deficits in episodic memory and spatial orientation. These deficits are mainly associated with the functionality of the posterior HC. We therefore hypothesized that key AD pathologies, i.e., β-amyloid and tau pathology would be particularly associated with the volume of the posterior subiculum in non-demented individuals. In our study we included 302 cognitively normal elderly participants (CN), 183 patients with subjective cognitive decline (SCD) and 171 patients with amnestic mild cognitive impairment (MCI), all of whom underwent 3T magnetic resonance images (MRI). The subicular subfield was segmented using Freesurfer 5.3 and divided into 10 volumetric segments moving from the most posterior (segment 1) to the most anterior part along the axis of the hippocampal head and body (segment 10). Cerebrospinal fluid (CSF) Aβ42 and phosphorylated tau (P-tau) were quantified using ELISA and were used as biomarkers for β-amyloid and tau pathology, respectively. In the total sample, tau-pathology and Aβ-pathology and (measured by elevated P-tau and low Aβ42 levels in CSF) and mild memory dysfunction were mostly associated with the volume changes of the posterior subiculum. Both SCD and MCI patients with elevated P-tau or low Aβ42 levels displayed predominantly posterior subicular atrophy in comparisons to control subjects with normal CSF biomarker levels. Finally, there was no main effect of Aβ42 or P-tau when comparing SCD with abnormal P-tau or Aβ42 with SCD with normal levels of these CSF-biomarkers. However, in the left subiculum there was a significant interaction revealing atrophy in the left posterior but not the anterior subiculum in participants with low Aβ42 levels. The same pattern was observed on the contralateral side in participants with elevated P-tau levels. In conclusion, AD pathologies and mild memory dysfunction are mainly associated with atrophy of the posterior parts of the subicular subfields of the HC in non-demented individuals. In light of these findings we suggest that segmentation of the HC subfields may benefit from considering the volume of the different anterior-posterior subsections of each subfield.
|
|
| 3. |
- Voevodskaya, Olga, et al.
(författare)
-
The effects of intracranial volume adjustment approaches on multiple regional MRI volumes in healthy aging and Alzheimer's disease
- 2014
-
Ingår i: Frontiers in Aging Neuroscience. - : Frontiers Media SA. - 1663-4365. ; 6, s. 264-
-
Tidskriftsartikel (refereegranskat)abstract
- In neurodegeneration research, normalization of regional volumes by intracranial volume (ICV) is important to estimate the extent of disease-driven atrophy. There is little agreement as to whether raw volumes, volume-to-ICV fractions or regional volumes from which the ICV factor has been regressed out should be used for volumetric brain imaging studies. Using multiple regional cortical and subcortical volumetric measures generated by Freesurfer (51 in total), the main aim of this study was to elucidate the implications of these adjustment approaches. Magnetic resonance imaging (MRI) data were analyzed from two large cohorts, the population-based PIVUS cohort (N = 406, all subjects age 75) and the Alzheimer disease Neuroimaging Initiative (ADNI) cohort (N = 724). Further, we studied whether the chosen ICV normalization approach influenced the relationship between hippocampus and cognition in the three diagnostic groups of the ADNI cohort (Alzheimer's disease, mild cognitive impairment, and healthy individuals). The ability of raw vs. adjusted hippocampal volumes to predict diagnostic status was also assessed. In both cohorts raw volumes correlate positively with ICV, but do not scale directly proportionally with it. The correlation direction is reversed for all volume-to-ICV fractions, except the lateral and third ventricles. Most gray matter fractions are larger in females, while lateral ventricle fractions are greater in males. Residual correction effectively eliminated the correlation between the regional volumes and ICV and removed gender differences. The association between hippocampal volumes and cognition was not altered by ICV normalization. Comparing prediction of diagnostic status using the different approaches, small but significant differences were found. The choice of normalization approach should be carefully considered when designing a volumetric brain imaging study.
|
|
