|
EVALUATION OF
BRAINAGE
PREDICTION PERFORMANCE IN REFERENCE SAMPLES
|
| Performance of the BrainAGE model for brain maturation during childhood & adolescencea
|
CTR |
394 [47%] |
10.7 ± 3.8 [5 – 19] |
1.5T [6] |
– |
Brain age estimation was highly accurate (r = 0.93; p < 0.001). The 95% confidence interval for the prediction of brain age was stable across the entire age range (±2.6 years). MAE was 1.1 years. BrainAGE
model for brain maturation during childhood and adolescence explained 87% of the individual variations in brain structures. |
| Performance of the BrainAGE model for brain aging from early into late adulthoodb
|
CTR
CTR |
547 [56%]
108 [37%] |
48 ± 17 [19 – 86]
32 ± 10 [20 – 59] |
1.5T [2], 3T [1]
1.5T [1] |
–
– |
Brain age estimation was highly accurate (r = 0.92; p < 0.001). The 95% confidence interval for the prediction of age was stable along the age range, with no broadening at old age (cf. age = 20 ± 11.6 years, age = 80 ± 11.7 years). Correlation between MAE and the true age indicated no systematical bias in the age estimations as a function of true ages (r = −0.015). MAE was 4.9 years. Results did not differ between genders (MAE: 5.0 years for males, 4.9 years for females; r = 0.9 for both genders). BrainAGE
model for brain aging during adulthood explained 85% of the individual variations in brain structures. |
| Performance of the BrainAGE model in baboonsC
|
CTR |
29 [52%] |
9.5 ± 4.9 [4 – 22] |
3T [1] |
– |
Strong correlation between estimated brain age and chronological age (r = 0.80; p < 0.0001) MAE was 2.1 years. Best fit between chronological and estimated brain age was linear (R2 = 0.64; p < 0.0001). With only 29 MRI data in the baboon sample, the baboon–specific
BrainAGE
framework showed very good performance, certainly improving with additional data |
| Performance of the BrainAGE model in rodentsd
|
CTR |
24 (up to 13 scans; n = 273) |
life span: 734 ± 110 days |
3T [1] |
– |
Brain age estimation was highly accurate (r = 0.95; p < 0.0001). MAE was 49 days, which equates to an estimation error of 6% in relation to the age range Best fit between chronological and estimated brain age was linear (R2 = 0.91; p < 0.0001). Analyses of individual brain aging trajectories showed increasing variance at old ages. Rodent–specific
BrainAGE
model showed excellent performances, explaining 91% of the individual variations in brain structures. |
|
RELIABILITY OF
BRAINAGE
ESTIMATIONS
|
| Scan-rescan-stability of BrainAGE estimations (same scanner)e
|
CTR, double-scanned on same scanner |
20 [60%] |
23.4 (4.0) [19 – 34] |
1.5T [1] |
1st scan: 13.8 (6.1) 2nd scan: 12.8 (5.6) |
BrainAGE estimations from 1st and 2nd scan were strongly correlated (r = 0.93***) and showed ICC of 0.93***. BrainAGE scores linearly adjusted for the offset at each scanning time point strongly correlated with raw scores (r = 0.996***). BrainAGE
estimations within the same subjects proved to be stable across a short delay between two scans. |
| Effect of MRI field strengths on stability of BrainAGE estimationse
|
CTR, double-scanned on 1.5T & 3T scanners |
60 [63%] |
75.2 (4.8) [60 – 87] |
1.5T/3T [26/26] |
1.5T scan: −5.9 (7.0) 3T scan: −9.1 (6.6) |
BrainAGE estimations from 1.5T and 3T scan were strongly correlated (r = 0.91***) and showed ICC of 0.90***. BrainAGE scores, linearly adjusted for the scanner–specific offset, did not differ between scanners***. BrainAGE
estimations within the same subjects proved to be stable across scanners with different field strengths. |
| Short-term changes of BrainAGE during the menstrual cycle f |
CTR (naturally cycling women) |
7 [100%] |
[21 – 31] |
1.5T [1] |
Difference to scan at menses:
Ovulation: −1.3 (1.2) Midluteal: 0.0 (1.6) Next menses: 0.1 (0.6)
|
BrainAGE decreased by −1.3 years* from menses to ovulation. Classification analyses of data whether acquired at menses or ovulation is much more precise when based on BrainAGE (accuracy: 86%/AUC: 0.88) as compared to GM (57% 0.55), WM (43%/0.51), and CSF (64%/0.55) volumes*. Lower BrainAGE were correlated to higher estradiol levels (r = −0.42*), whereas progesterone levels did not correlate with individual BrainAGE. The
BrainAGE
method proved to recognize short-term effects of hormones on individual brain structure. |
|
BrainAGE
MODEL FOR BRAIN MATURATION DURING CHILDHOOD AND ADOLESCENCE
|
| Effects of being born preterm on brain maturationa
|
Born preterm, before 27 weeks of gestation
Born preterm, after 29 weeks of gestation |
10
15 |
14.3 (1.4) [12 – 16]
14.7 (1.5) [12 – 16] |
1.5T (1) |
−2.0 (0.7)
−0.4 (1.5) |
Scanned between the ages of 12–16 years, BrainAGE were about 1.5 years lower in subjects who were born before the end of the 27th week of gestation vs. subjects who were born after the end of the 29th week of gestation**. Although the mean difference in gestational age between both groups was only 5 weeks, results show a systematically lower
BrainAGE
in adolescents who were born extremely preterm, implying delayed brain maturation. |
|
BRAINAGE
IN MILD COGNITIVE IMPAIRMENT AND ALZHEIMER'S DISEASE
|
| Premature brain aging in ADb
|
CTR
AD |
232 [49%]
102 [54%] |
76.0 (5.1) [60 – 90]
75.8 (8.2) [55 – 88] |
1.5T [26] |
0
10 |
For people with mild AD, the mean
BrainAGE
score was 10 years, implying a systematically higher estimated than chronological age based on structural MRI data***. BrainAGE estimations differed significantly between CTR/sMCI vs. pMCI/AD at baseline* and follow-up*. Over the follow-up period of up to 4 years, BrainAGE remained stable for CTR (annual changing rate: 0.12) & sMCI (0.07), but increased in the pMCI (1.05) and AD (1.51), thus suggesting additional acceleration in brain aging*. Higher BrainAGE were related to worse cognitive functioning and more severe clinical symptoms at baseline (ADAS: r = 0.45***; CDR: r = 0.39***; MMSE: r = −0.46***) and at follow up (ADAS: r = 0.55***; CDR: r = 0.46***; MMSE: r = −0.55***). |
| Longitudinal changes of individual brain aging in CTR, MCI, ADe
|
CTR
sMCI
pMCI
AD |
108 [43%]
36 [17%]
112 [40%]
150 [49%] |
Baseline: 75.6 (5.0) follow-up: 78.9 (5.0)
Baseline: 77.0 (6.1) follow-up: 80.1 (6.0)
Baseline: 74.5 (7.4) follow-up: 77.2 (7.6)
Baseline: 74.6 (7.6) follow-up: 76.3 (7.7) |
1.5T (26) |
Baseline: −0.3 follow-up: −0.1
Baseline: −0.5 follow-up: −0.4
Baseline: 6.2 follow-up: 9.0
Baseline: 6.7 follow-up: 9.0 |
Changes in BrainAGE from baseline to last follow-up scan were related to worsening of cognitive functioning and clinical symptoms (ADAS: r = 0.30***; CDR: r = 0.27***; MMSE: r = −0.33***). Results suggest structural changes that show the pattern of accelerated brain aging in pMCI and AD, accelerating even more, at the speed of 1 additional year in
BrainAGE
estimation per follow-up year in pMCI and 1.5 additional years in AD. |
| Effects of APOE–genotype on longitudinal changes in CTR, MCI, ADg
|
CTRC [APOE ε4 carriers]
sMCIC [APOE ε4 carriers]
pMCIC [APOE ε4 carriers]
ADC [APOE ε4 carriers]
CTRNC [APOE ε4 non-carriers]
sMCINC [APOE ε4 non-carriers]
pMCINC [APOE ε4 non-carriers]
ADNC [APOE ε4 non-carriers] |
26
14
78
101
81
22
34
49 |
Baseline: 75.0 (5.1) follow-up: 78.2 (5.1)
Baseline: 77.3 (5.6) follow-up: 80.4 (5.4)
Baseline: 74.1 (6.5) follow-up: 76.7 (6.7)
Baseline: 74.1 (6.8) follow-up: 75.8 (6.9)
Baseline: 75.9 (4.9) follow-up: 79.1 (5.0)
Baseline: 76.8 (6.5) follow-up: 79.9 (6.5)
Baseline: 75.5 (9.3) follow-up: 78.1 (9.4)
Baseline: 75.7 (8.9) follow-up: 77.4 (9.1) |
1.5T [26] |
Baseline: −0.1 (6.8) follow-up: −0.2 (7.9)
Baseline: −0.9 (6.1) follow-up: 0.0 (6.0)
Baseline: 5.8 (6.4) follow-up: 8.7 (7.2)
Baseline: 5.8 (7.7) follow-up: 8.3 (8.0)
Baseline: −1.3 (6.4) follow-up: −1.4 (6.1)
Baseline: −0.9 (6.1) follow-up: −0.6 (4.8)
Baseline: 5.5 (9.7) follow-up: 7.3 (10.3)
Baseline: 6.2 (9.5) follow-up: 7.7 (10.1) |
BrainAGE estimations differed significantly between CTR/sMCI vs. pMCI/AD at baseline* and up to 4 years follow-up*, without significant effects regarding APOE ε4 status or interaction between diagnostic group and APOE ε4 status, nor particular allelic isoforms. Annual changing rates in BrainAGE differed significantly between CTR/sMCI vs. pMCI/AD as well as between APOE ε4 carriers vs. ε4 non-carriers*, with APOE ε4 carriers showing C NC C NC C increased changing rates (NO: 0.0; NO: 0.0; sMCI: 0.2; sMCI: −0.1; pMCI: 1.1; NC C NC pMCI: 0.6; AD: 1.7; AD: 0.9). Larger BrainAGE were significantly related to worse cognitive functioning and more sever clinical symptoms at baseline, being stronger in APOE ε4 non-carriers vs. ε4 carriers. Results suggest structural changes that show the pattern of accelerated brain aging in pMCI and AD, accelerating even more during follow-up in pMCI and AD, with APOE
ε4 carriers showing faster acceleration of brain aging. |
|
BRAINAGE–BASED PREDICTION OF CONVERSION TO ALZHEIMER'S DISEASE
|
|
BrainAGE–based prediction of conversion from MCI to ADh
|
(1) sMCI
(2) pMCI_early
(3) pMCI_late |
62 [21%]
58 [43%]
75 [36%] |
76.4 (6.2) [58 – 88]
73.9 (7.0) [55 – 86]
75.2 (7.3) [56 – 88] |
1.5T [26] |
0.75
8.73
5.62 |
Predicting future conversion to AD within 12-months follow-up based on baseline BrainAGE (accuracy: 81%/AUC: 0.83) was significantly more accurate than predictions based on chronological age (41%/0.59), hippocampus volumes (left: 66%/0.69; right: 61%/0.67), cognitive scores (ADAS: 66%/0.80; CDR–SB: 59%/0.71; MMSE: 57% /0.69), and CSF biomarkers (T-Tau: 60%/0.60; P-Tau: 57%/0.66; Aβ42: 57%/0.58; Aβ42/P-Tau: 69%/0.65). Predicting future conversion to AD within 36-months follow-up based on baseline BrainAGE (accuracy: 75%/AUC: 0.78) was significantly more accurate than predictions based on chronological age (52%/0.56), hippocampus volumes (left: 61%/0.69; right: 54%/0.67), cognitive scores (ADAS: 48%/0.75; CDR–SB: 38%/0.67; MMSE: 37%/0.67), and CSF biomarkers (T-Tau: 58%/0.61; P-Tau: 43%/0.63; Aβ42: 49%/0.56; Aβ42/P-Tau: 73%/0.62). Prognostic certainty for prediction of conversion to AD increased from 68% pre-test probability to 90% post-test probability when using
BrainAGE
(right hippocampus: 84%; left hippocampus: 85%; ADAS: 86%; CDR-SB: 68%; MMSE: 79%). Each additional year in
BrainAGE
was associated with a 10% greater risk of developing AD during 36-months follow-up. |
| Effects of APOE-genotype on BrainAGE-based prediction of conversion from MCI to ADg
|
sMCIC [APOE ε4 carriers]
pMCIC_early [APOE ε4 carriers]
pMCIC_late [APOE ε4 carriers]
sMCINC [APOE ε4 non-carriers]
pMCINC_early [APOE ε4 non- carriers]
pMCINC_late [APOE ε4 non- carriers] |
26 [12%]
33 [39%]
58 [38%]
36 [28%]
24 [46%]
16 [31%] |
76.5 (5.2)
72.9 (6.0)
75.0 (6.4)
76.2 (6.8)
75.3 (8.3)
76.4 (10.0) |
1.5T [26] |
0.0 (4.4)
9.0 (6.3)
5.7 (6.0)
1.2 (4.0)
8.0 (9.2)
5.0 (7.7) |
Cox regression showed higher baseline BrainAGE being associated with a higher risk of converting to AD independent of APOE status, with BrainAGE above median of 4.5 years indicating a nearly 4 times greater risk of converting to AD as compared to BrainAGE below median***#. Including APOE status into Cox model, the accuracy of the prediction tended to improve. APOE ε4 carriers: predicting future conversion to AD within 12-months follow-up based on baseline BrainAGE (accuracy: 85%/AUC: 0.88) was significantly more accurate than predictions based on chronological age (39%) or cognitive scores (ADAS: 69%; CDR-SB: 49%; MMSE: 46%). APOE ε4 carriers: predicting future conversion to AD within 36-months follow-up based on baseline BrainAGE (accuracy: 75%/AUC: 0.82) was significantly more accurate than predictions based on chronological age (54%) or cognitive scores (ADAS: 43%; CDR-SB: 26%; MMSE: 23%). APOE ε4 non-carriers: predicting future conversion to AD within 12-months follow-up based on baseline BrainAGE (accuracy: 78%/AUC: 0.75) was significantly more accurate than predictions based on chronological age (50%) or cognitive scores (ADAS: 68%; CDR SB: 67%; MMSE: 60%). APOE ε4 non-carriers: predicting future conversion to AD within 36-months follow-up based on baseline BrainAGE (accuracy: 74%/AUC: 0.71) was significantly more accurate than predictions based on chronological age (47%) or cognitive scores (ADAS: 64%; CDR SB: 51%; MMSE: 47%). From diagnosis at study baseline onwards, APOE ε4 carriers showed the tendency to take to convert to AD (560 ± 280 days) as compared to APOE ε4 non-carriers (471 ± 233 days)#. Prediction of conversion was most accurate using
BrainAGE
as compared to neuropsychological test scores, even when including the APOE
ε4-status. |
| EFFECTS OF PSYCHIATRIC DISORDERS ON BRAIN AGING |
| Effects of schizophrenia and bipolar disorder on brain agingi
|
CTR
SZ
BD |
70 [43%]
45 [36%]
22 [55%] |
33.8 (9.4) [22 - 58]
33.7 (10.5) [21 – 65]
37.7 (10.7) [24 – 58] |
3T [1] |
−0.2 (5.6)
2.6 (6.0)
−1.2 (4.6) |
BrainAGE scores were significantly higher in SZ by about 3 years*, but not BD patients. Structural brain aging in bipolar disorder is comparable to healthy brain aging. Structural brain aging is significantly advanced in schizophrenia. |
| Brain age in early stages of bipolar disorders or schizophreniak
|
CTR
SZ (FES)
CTR
Unaffected, high- risk for BD
BD |
43 [40%]
43 [40%]
60 [60%]
48 [60%]
48 [69%] |
27.0 (4.4)
27.1 (4.9)
23.4 (4.9)
20.9 (4.1)
23.1 (4.5) |
3T [1]
1.5T [2] |
−0.01 (4.1)
2.6 (4.1)
0.2 (5.3)
−1.0 (5.0)
−1.0 (5.2) |
BrainAGE scores were significantly higher in SZ by about 3 years**. The proportion of participants who had a greater biological than chronological age was higher in SZ (74%) than CTR (46%)**. BrainAGE was not associated with duration of illness or duration of untreated psychosis. No differences in BrainAGE between the SZ diagnoses. BrainAGE in SZ was negatively associated with GM volume diffusely throughout the brain***. Structural brain aging is significantly advanced in schizophrenia BrainAGE scores were comparable between unaffected, high-risk for BD, BD, and CTR participant's#. BrainAGE scores were not associated with number of episodes or hospitalizations, as we as duration of illness. Structural brain aging in bipolar disorder and unaffected, high-risk subjects for BD is comparable to healthy brain aging. |
| Obesity, dyslipidemia and brain age in first-episode psychosisl
|
CTR
FEP |
114 [45%]
120 [38%] |
33.8 (9.4) [18 – 35]
33.7 (10.5) [18 – 35] |
3T [1] |
−0.2 (5.6)
2.6 (6.0) |
BrainAGE scores were significantly associated with FEP**, obesity**, and BMI*. BrainAGE was highest in participants with a combination of FEP and obesity (3.8 years) and lowest in normal weight CTRs (−0.3 years) *. Even among only FEP participants, BMI remained significantly associated with BrainAGE. As compared to CTRs, BrainAGE scores in non-medicated FEP participants were greater than in CTRs**, comparable to previously medicated FEP individuals, and not associated with cumulative exposure to antipsychotics (with non-medicated FEP participants not differing from the previously medicated ones in relevant clinical variables). Medication dosage at the time of scanning was not associated with BrainAGE or BMI. BrainAGE was not associated with duration of illness, duration of untreated psychosis, another health markers. Brain structural aging is significantly advanced in medicated as well as non- medicated patients with psychosis (FEP). Obesity added to advanced structural brain aging in controls as well as psychosis. |
| EFFECTS OF INDIVIDUAL HEALTH ON BRAIN AGING |
| Effects of type 2 diabetes mellitus on brain agingm
|
CTR
DM2 |
87 [53%]
98 [46%] |
65.3 (8.5)
64.6 (8.1) |
3T [1] |
0.0 (6.7)
4.6 (7.2) |
Brain ages in DM2 were estimated 4.6 years higher than their chronological age***. Diabetes duration correlated positively with BrainAGE scores (r = 0.31*). BrainAGE scores in whole sample were related to fasting blood glucose (r = 0.34*; BrainAGE 1st vs. 4th quartile: 5.5 years*), TNFα levels (r = 0.29**), smoking duration (r = 0.20**; BrainAGE 1st vs. 4th quartile: 3.4 years**), alcohol consumption (r = 0.24***; BrainAGE 1st vs. 4th quartile: 4.1 years**). BrainAGE scores in whole sample were related to verbal fluency (r = −0.25**; BrainAGE 1st vs. 4th quartile: 5.6 years***). BrainAGE scores in whole sample were related to depression scores (r = 0.23*; BrainAGE 1st vs. 4th quartile: 5.4 years**). BrainAGE scores were higher in males than females**. Type 2 DM is associated with structural brain changes that reflect advanced brain aging. |
| Longitudinal effects of type 2 diabetes mellitus on brain agingm
|
CTR
DM2 |
13 [61%]
12 [67%] |
Baseline: 69.9 (5.5) follow-up: 73.9 (5.7)
Baseline: 63.3 (6.9) follow-up: 66.8 (6.7) |
3T [1] |
Baseline: 0.0 follow-up: 0.0
Baseline: 5.1 follow-up: 5.9 |
At baseline BrainAGE scores in DM2 subjects were 5.1 years higher than in CTR*. BrainAGE scores in CTR did not change during 3.8 ± 1.5 years follow-up. BrainAGE scores in DM2 subjects after 3.8 ± 1.5 years follow-up were 5.9 years higher than in CTR*. BrainAGE
in DM2 is increasing by 0.2 years per follow-up year. |
| Gender-specific effects of health parameters on brain agingn
|
male CTR
female CTR |
118
110 |
75.8 (5.3) [60 – 88]
76.1 (4.8) [62 – 90] |
1.5T [26] |
0
0 |
39% of variance within BrainAGE scores were attributed to health parameters, with BMI, uric acid, GGT, DBD contributing most***. BrainAGE scores were related to BMI (r = 0.35***; BrainAGE 1st vs. 4th quartile: 7.5 years***), uric acid (r = 0.25**; BrainAGE 1st vs. 4th quartile: 5.6 years*), GGT (r = 0.20*; BrainAGE 1st vs. 4th quartile: 7.5 years**), DBD (r = 0.19*; BrainAGE 1st vs. 4th quartile: 6.6 years**). BrainAGE scores in “healthy” men (values below the medians of BMI, DBD, GGT, uric acid; n = 9) vs. men with “risky” health markers (values above the medians of BMI, DBD, GGT, and uric acid; n = 14): −8.0 vs. 6.7 years*. In cognitively healthy elderly men, markers of the metabolic syndrome, and impaired liver and kidney functions were associated with subtle structural changes that reflect accelerated brain aging, whereas protective effects on brain aging were observed for markers of good health. 32% of variance within BrainAGE scores were attributed to health parameters, with GGT, ALT, AST, vitamin B12 contributing most**. BrainAGE scores were related to GGT (r = 0.25*; BrainAGE 1st vs. 4th quartile: 6.1 years**), ALT (r = 0.23*; BrainAGE 1st vs. 4th quartile: 5.1 years*), AST (r = 0.20*; BrainAGE 1st vs. 4th quartile: 3.1 years), vitamin B (r = −0.17; BrainAGE 1st vs. 4th quartile: 4.8 years*). 12 BrainAGE scores in “healthy” women (values below the medians of GGT, ALT, AST, vitamin B12; n = 14) vs. women with “risky” health markers (values above the medians of GGT, ALT, AST, vitamin B12; n = 13): −1.0 vs. 3.8 years.* In cognitively fit elderly women, protective effects on brain aging were observed for markers of good health. |
| PROTECTING INTERVENTIONS FOR BRAIN AGING |
| Effects of long-term meditation practice on brain agingo
|
CTR [no meditation practice]
Meditators |
50 [44%]
50 [44%] |
51.4 (11.8) [24 – 77]
51.4 (12.8) [24 – 77] |
1.5T [1] |
0
−7.53 |
Brains of meditators (4–46 years practice, mean = 20 years) were estimated to be 7.5 years younger at age 50 than those of CTRs*. For every additional year over age fifty, meditators' brains were estimated to be an additional 1 month, 22 days younger than their chronological age*. Female brains were estimated to be 3.4 years younger than male brains**. Meditation is beneficial for brain preservation, effectively protecting against age–related atrophy with a consistently slower rate of brain aging throughout life. |
| Effects of making music on brain agingp
|
CTR [non-musicians]
Amateur musicians
Professional musicians |
38 [39%]
45 [40%]
42 [48%] |
25.2 (4.8)
24.3 (3.9)
24.3 (3.9) |
1.5T [1] |
0.48 (6.85)
−4.51 (5.60)
−3.70 (6.57) |
Musicians had younger brains than non-musicians**. Small positive correlation between years of music making and BrainAGE score in professional musicians (r = 0.32*), suggesting that with increasing number of years of music making, the age-delaying effect (in professionals) might lessen. Making music has an protecting effect on brain aging, with a stronger effect when it is not performed as a main profession, but as a leisure or extracurricular activity. |
| EFFECTS OF PRENATAL UNDERNUTRITION ON BRAIN AGING IN HUMANS AND NON-HUMAN PRIMATES |
| Gender-specific effects of prenatal under nutrition on brain aging in humansq
|
Men born before Dutch famine
Men exposed to Dutch famine in early gestation
Men conceived after Dutch famine
Women born before Dutch famine
Women exposed to Dutch famine in early gestation
Women conceived after Dutch famine |
14
19
19
21
22
23 |
68.6 (0.4)
67.4 (0.1)
66.7 (0.4)
68.7 (0.5)
67.4 (0.2)
66.7 (0.4) |
3T [1] |
−1.8 (3.5)
2.5 (5.2)
0.5 (4.6)
−0.1 (4.3)
0.9 (4.0)
−0.1 (5.3) |
In men, the variance in individual BrainAGE scores was best explained by birth characteristics, late–life health characteristics, chronological age, and famine exposure*. In women, the variance in individual BrainAGE scores was best explained by birth characteristics, chronological age at MRI data acquisition, and famine exposure*. Premature brain aging by about 4 years in male offspring who had been exposed to Dutch famine during early gestation, as compared to men born before the famine. BrainAGE did not differ in the female sample. Cognitive and neuropsychiatric test scores in late adulthood did not differ between the famine exposure groups. Exposure to prenatal under nutrition is associated with premature brain aging during late adulthood. |
| Gender–specific effects of prenatal undernutrition on brain aging in non– human primatesC
|
CTR
MNR |
12 [42%]
11 [45%] |
4.9 (1.1) [4–7 (equiv. to human 14–24)]
5.0 (1.1) [4–7 (equiv. to human 14–24)] |
3T [1] |
−0.2 (1.9) [males: 0.9 (1.5)] [females: −1.6 (1.4)]
1.0 (1.8) c[males: 0.9 (2.4)] [females: 1.2 (0.8)] |
Baboon BrainAGE based on species-specific preprocessed GM images, were significantly increased by 2.74 years in young adult female MNR subjects as compared to young adult female CTR offspring**, suggesting premature brain aging in female MNR offspring as a result of developmental programming due to fetal undernutrition. In males, BrainAGE did not differ between MNR and CTR offspring. The effects of moderate MNR on individual brain aging occurred in the absence of fetal growth restriction or marked maternal weight reduction at birth. |
