Table 2.
Model parameter estimates summarizing growth processes in five neuropsychological factors by diagnostic status: Results from ADNI (n=626)
| Memory |
Executive function/processing speed |
Visuospatial |
Language |
Attention |
||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Intercept (Mean) | Trajectory (Mean) | Intercept (Mean) | Trajectory (Mean) | Intercept (Mean) | Trajectory (Mean) | Intercept (Mean) | Trajectory (Mean) | Intercept (Mean) | Trajectory (Mean) | |
| Means | ||||||||||
| Healthy control | 56.6* | –0.5* | 53.8* | 0.0 | 52.7* | –0.1 | 54.9* | 0.1 | 52.8* | 0.2 |
| MCI | 45.5* | –1.1* | 47.4* | –1.6* | 48.6* | –1.1* | 47.6* | –1.1* | 48.7* | –1.0* |
| Between-diagnostic status, within-domain differences | ||||||||||
| –11.1* | –0.6* | –6.4* | –1.7* | –4.1* | –0.9* | –7.3* | –1.2* | –4.1* | –1.1* | |
| Within-diagnostic status, between-domain (memory) differences | ||||||||||
| Healthy control | NA | NA | 2.8* | –0.5* | 3.9* | –0.3 | 1.7* | –0.6* | 3.8* | –0.7* |
| MCI | NA | NA | –1.9* | 0.5* | –3.0* | –0.1 | –2.1* | –0.1 | –3.2* | –0.2 |
| Magnitude of Differences | NA | NA | 4.6* | –1.0* | 6.9* | –0.3 | 3.8* | –0.5* | 7.0* | –0.5* |
| Residual variances, variance (SE) | ||||||||||
| Growth parameters | 34.8* (2.6) | 1.5* (0.3) | 59.1* (4.2) | 2.6* (0.6) | 31.4* (3.4) | 2.4* (0.8) | 37.4* (2.3) | 0.0 (0.0) | 57.3* (3.5) | 0.0 (0.0) |
| Indicator residual variances | 1.5* (0.4) | 0.1 (0.4) | 4.2* (1.3) | 0.3 (0.4) | 0.4 (1.1) | |||||
| Model fit statistics | ||||||||||
| RMSEA | 0.057 | 0.053 | 0.061 | 0.058 | ||||||
| CFI | 0.891 | 0.905 | 0.884 | 0.914 | ||||||
p<0.05
Results from four multiple indicator parallel process latent growth models used to explore domain-specific growth in cognitive domains in ADNI healthy control and MCI participants. Means of latent intercept and trajectory parameters representing growth processes are shown in the first two rows for each cognitive domain. Between-diagnostic group, within-domain differences provide tests of differences in parameters between healthy control and MCI participants, which was accomplished by regressing growth parameters on an indicator for diagnostic group (see Fig. 1 and the Methods). Within-diagnostic group, between-domain differences provide tests of differences between parameters for a cognitive domain and memory. Because the memory growth process was part of all four models, fit statistics are not provided