Table 3.
Relationship between different amyloid positivity thresholds and longitudinal performance on a memory task
| Threshold | Amyloid status | Estimated annual change | AIC | |
|---|---|---|---|---|
| Visual assessment | Negative | 0.19 (0.05) | 2933.7 | |
| Positive | − 0.28 (0.09)** | |||
| BPND | 0.19 | Negative | 0.19 (0.05) | 2938.5 |
| Positive | − 0.22 (0.08)** | |||
| 0.23 | Negative | 0.17 (0.05) | 2935.8 | |
| Positive | − 0.28 (0.11)** | |||
| 0.29 | Negative | 0.15 (0.05) | 2938.2 | |
| Positive | − 0.28 (0.12)** | |||
| SUVr | 1.28 | Negative | 0.16 (0.05) | 2941.2 |
| Positive | − 0.21 (0.10)** | |||
| 1.34 | Negative | 0.14 (0.05) | 2940.0 | |
| Positive | − 0.28 (0.12)** | |||
| 1.43 | Negative | 0.14 (0.05) | 2943.0 | |
| Positive | − 0.29 (0.13)** |
Values given are beta (SE), as estimated by linear mixed models (predictor: amyloid status, outcome: score on RAVLT delayed recall). Numbers reflect annual change in raw score points. Models are adjusted for age, sex, education and scanner type
AIC Akaike information criterion, SE standard error, RAVLT Rey auditory verbal learning task, BPND binding potential, SUVr standardized uptake value ratio
**p value < 0.01. p value represents the significance of the difference between a positive amyloid status compared with a negative amyloid status