Table 2.
Dysfunctions in specific cognitive domains in idiopathic parkinsonism
| PD | MSA | PSP | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Mean | SD | P-value (PD versus HC) | Mean | SD | P-value (MSA versus PD) | Mean | SD | P-value (PSP versus PD) | |
| (A) Baseline differences | |||||||||
| Working memory | −0.47 | 0.85 | 0.017* | −0.01 | 0.92 | 0.103 | −0.88 | 0.67 | 0.052 |
| Psychomotor speed and attention | −2.08 | 2.25 | <0.001* | −2.04 | 2.45 | 0.816 | −4.52 | 3.57 | 0.003* |
| Visuospatial function | −0.65 | 1.31 | 0.008* | −0.72 | 0.97 | 0.696 | −1.58 | 2.36 | 0.189 |
| Language | −0.47 | 0.93 | 0.005* | −0.52 | 0.69 | 0.979 | −1.76 | 1.32 | <0.001* |
| Episodic memory | −0.81 | 1.04 | <0.001* | −0.95 | 0.97 | 0.641 | −1.96 | 0.89 | <0.001* |
| Executive function | −0.97 | 1.35 | <0.001* | −0.67 | 1.01 | 0.725 | −1.52 | 1.19 | 0.072 |
| Diff. | P-value (PD versus HC) | Diff. | P-value (MSA versus HC) | Diff. | P-value (PSP versus HC) | ||||
| (B) Longitudinal differences (estimated by mixed models) | |||||||||
| Working memory | −0.49 | 0.002* | −0.30 | 0.632 | −1.15 | 0.002* | |||
| Psychomotor speed and attention | −2.19 | <0.001* | −2.56 | <0.001* | −4.44 | <0.001* | |||
| Visuospatial function | −0.67 | 0.003* | −0.46 | 0.257 | −1.74 | <0.001* | |||
| Language | −0.50 | 0.004* | −0.50 | 0.068 | −1.65 | <0.001* | |||
| Episodic memory | −0.87 | <0.001* | −0.84 | 0.033* | −1.85 | <0.001* | |||
| Executive function | −1.13 | <0.001* | −0.98 | 0.003* | −1.56 | <0.001* | |||
The performance in cognitive domains is showed in z-scores, where the mean of all test results in a domain is in SDs from the mean of the HCs at baseline. At the baseline cross-sectional comparison, Parkinson disease was compared with the HCs, while PSP and MSA were only compared with Parkinson disease, as this was determined to be clinically most relevant. The difference in table section (B) (Diff.) is estimated marginal means, given as the total difference from HCs throughout all follow-ups, after adjustment for age and sex. Mixed models were calculated with intercept as a random effect, autoregressive 1.
The variable was significant after controlling for false discovery rate with α = 0.05, by the Benjamini–Hochberg procedure. TMT A is the measure for psychomotor speed and attention.