56 CHAPTER 3 Results One healthy control was excluded from further analyses, as his overall ER of 25% was more than three standard deviations higher than the mean ER. Data of 21 NA patients and 20 healthy controls were used in the final analyses. For four NA patients, we did not have the SRQ-DLV scores. There were no significant group differences in age or sex (Table 1). The groups did differ significantly in force exerted by the serratus anterior muscles, as evidenced by the significant GROUP x SIDE interaction effect (F(1,39) = 9.3, p = 0.004, part. η2 = 0.19). Post-hoc comparisons showed that NA patients exerted significantly less force with the serratus anterior on the affected, right side compared to the contralateral unaffected left side (F(1,20) = 14.8, p = 0.001, part. η2 = 0.43) and compared to healthy controls on the right side (F(1,39) = 11.3, p = 0.002, part. η2 = 0.23). This confirms that NA patients had lateralized symptoms of the right upper limb. Group means and standard error of the means are displayed in Table 1. Hand Laterality Judgment Task Error rates (ER) Participants performed the task accurately, reflected by low overall ER across both groups (NA patients: 3.2 ± 2.8%; healthy controls: 2.4 ± 2.3%; no effect of GROUP: F(1, 39) = 0.6, p = 0.46, part. η2 = 0.01). In both groups, subjects made more errors for more extreme rotations (main effect of ROTATION: F(7, 273) = 11.9, p < 0.001, part. η2 = 0.23; no interaction effects involving ROTATION and GROUP: F ≤ 1.6, p ≥ 0.17, part. η2 = 0.04; Figure 2). However, NA patients made relatively more errors with their affected right limb than healthy controls (GROUP x LATERALITY interaction: F(1,39) = 13.1, p = 0.001, part. η2 = 0.25). Specifically, post-hoc comparisons showed that NA patients had a higher ER for the affected, right limb than healthy controls (F(1,39) = 8.9), p = 0.005, part. η2 = 0.19) (see Figure 2C). Furthermore, in healthy controls the ER was significantly lower on their dominant right side than on their non-dominant left side (F(1,19) = 19.4, p < 0.001, part. η2 = 0.51), whereas for NA patients ER tended to be higher on their affected, dominant right side than on their non-dominant left side (F(1,20) = 4.0, p = 0.06, part. η2 = 0.17). (The interaction between GROUP and LATERALITY remained significant when correcting ERs for RTs by running the same analysis on efficiency scores (F(1,39) = 5.4, p = 0.03, part. η2 = 0.12), confirming that group differences in speed-accuracy trade-off did not influence our results (see supplementary materials). Across groups, participants’ own hand posture did not significantly influence ER (no main effect of POSTURE: F(1,39) = 2.2, p = 0.14, part. η2 = 0.05; no interaction effects involving POSTURE and GROUP: F(1, 39) ≤ 1.1, p ≥ 0.30, part. η2 ≤ 0.03). Reaction times (RT) Overall, NA patients and healthy controls performed the task equally fast (no main effect of GROUP: F(1,39) = 0.04, p = 0.84, part. η2 = 0.001; Figure 3). All participants were faster on their dominant, right side than on their non-dominant, left side (significant main effect of LATERALITY: F(1,39) = 20.6, p < 0.001, part. η2 = 0.35; no significant interaction effects involving LATERALITY and GROUP: F ≤ 0.5, p ≥ 0.50, part. η2 ≤ 0.01). Participants were slower for stimuli that were rotated at more extreme angles (main effect of ROTATION (F(3.1, 119.5) = 67.8, p < 0.001, part. η2 = 0.64). Importantly, in both groups RTs were
RkJQdWJsaXNoZXIy MjY0ODMw