Research Article

Association between Community Ambulation Walking Patterns and Cognitive Function in Patients with Parkinson’s Disease: Further Insights into Motor-Cognitive Links

Table 3

(a) Associations between the attention and 3-day acceleration measures in the entire cohort (). (b) Acceleration derived 3-day measures of the two groups (using Student’s -tests) (low and high quartiles).
(a)

Adjusted univariate modelAdjusted multivariate model

Quantity measures (activity count)
Total percent of activity duration [%]
 AP−0.03 (−2.00–1.44)
Total number of steps for 3 days [#]
 AP0.02 (−0.0003–0.0004)
Cadence [steps/minute]
 AP0.45 (0.35–0.82)  0.36 (0.24–0.70)  

Quality of activity measures
Phase coordination index [%]
 AP−0.10 (−0.61–0.18)
Amplitude of dominant frequency [prs]
 V0.34 (12.93–46.51)  0.21 (3.46–34.80)  
 AP0.11 (−9.62–35.44)
 ML−0.23 (−48.77–−3.18)  
Stride regularity [g2]
 V0.27 (9.22–55.70)  
 AP0.12 (−11.13–52.74)
 ML0.01 (−25.63–28.36)
Harmonic ratio
 V0.19 (−0.04–12.28)
 AP0.22 (0.97–14.03)  
 ML−0.19 (−24.88–20.50)

Entries are the values, 95% confidence intervals, and the associated value. Univariate and multivariate models were adjusted for age, gender, and disease duration.
(b)

MeasurePD low attentionPD high attention value

Quantity measures (activity count)
Total percent of activity duration [%]
 AP2.27 ± 2.252.20 ± 2.010.957
Total number of steps for 3 days [#]
 AP9537.29 ± 9612.5410187.55 ± 10479.790.789
Cadence [steps/minute]
 AP99.47 ± 16.73110.33 ± 11.940.009

Quality of activity measures
Phase coordination index [%]
 AP7.78 ± 8.977.53 ± 8.960.278
Amplitude of dominant frequency [prs]
 V0.57 ± 0.200.70 ± 0.160.013
 AP0.57 ± 0.160.56 ± 0.140.809
 ML0.19 ± 0.150.16 ± 0.120.393
Stride regularity [g2]
 V0.50 ± 0.120.58 ± 0.130.054
 AP0.53 ± 0.100.54 ± 0.120.707
 ML0.38 ± 0.140.38 ± 0.130.896
Harmonic ratio
 V2.13 ± 0.492.28 ± 0.560.308
 AP1.96 ± 0.482.07 ± 0.550.446
 ML0.60 ± 0.160.63 ± 0.130.581

Measures which were not distributed normally according to the Kolmogorov-Smirnov test and therefore were analyzed with the Mann-Whitney test.
Measures which were significantly different in the two groups. We performed the Hochberg-Benjamini method for multiple comparison analysis for each of the 3 locomotor constructs separately: vertical (V), anterior-posterior (AP), and mediolateral (ML). values less than or equal to 0.013 (V), 0.05 (AP), and 0.05 (ML) were considered statistically significant in the 3 different constructs in the present analyses.