We included n = 133 subjects (n = 97 individuals with PD (n = 67 with FOG) and n = 36 HC). Participant characteristics are shown in Tables 1, 2 comparing PD and HC and in Table 3 comparing Freezers and Non-freezers. Individuals with PD and HC did not differ significantly regarding age and sex (p > 0.05).

Table 1 Participant characteristics (n = 133)Table 2 Comparison of gait- and turning asymmetries (PD vs. HC)Table 3 Congruency between symptom- and gait asymmetry

The degree of gait and turning asymmetry and the distribution of the lesser versus the more affected side did not significantly differ between the ST and the DT condition (Supplementary Table 1). In the following, only the data of the ST condition is presented. Data regarding the DT condition can be found in the supplementary material (Tables 2, 3, 4).

Table 4 Congruency between symptom- and turning asymmetryAsymmetry in symptom laterality (MDS-UPDRS-III)

Eighty-eight (90.7%) of the 97 individuals with PD presented with symptom asymmetry according to the MDS-UPDRS-IIITOTAL. Fifty-two (59,0%) had a more affected left side and 36 (40.9%) had a more affected right side (and 9 had no asymmetry). Seventy-six (87.4%) participants presented with disease asymmetry based on the scores of the lower extremity items (MDS-UPDRS-IIILL).

Asymmetry in people with PD versus HC

Step length asymmetry differed significantly between people with PD and HC (p < 0.001, Table 2), whereby PD had more asymmetry. Similarly, temporal gait asymmetry for swing- (p < 0.001) and stance time (p = 0.003) was greater in PD. However, turning time asymmetry was not significantly different between PD and HC (p = 0.077). Figure 1 shows the step length asymmetry and turning time asymmetry values of PD and HC. The number of people with PD who had step length asymmetry and turning time asymmetry above the abnormal threshold (mean + 1 SD of HC values) was n = 41 (42.3%) and n = 28 (28.9%), and for HC n = 7 (19.4%) and n = 4 (11.1%), respectively.

Fig. 1

Asymmetry of step length and turning time during single task conditions. Dashed lines represent the mean + 1 standard deviation for HC. *Statistically significant difference

Asymmetry in Freezers versus Non-Freezers

Freezers did not differ significantly from Non-Freezers regarding age, sex and MoCA but performed worse for the Mini-BESTest and MDS-UPDRS III and had longer disease duration (p < 0.05, Table 5 supplemental material online). Gait asymmetry did not differ significantly between Freezers and Non-Freezers (step length asymmetry: p = 0.200, swing time asymmetry: p = 0.127, stance time asymmetry: p = 0.175). Similarly, turning time asymmetry was not significantly different between the groups (p = 0.102). However, turning step asymmetry was significantly higher in the Freezers compared to Non-Freezers (p = 0.010). For asymmetry of motor symptoms, the Non-Freezers showed higher asymmetry values (p = 0.029). For details see Table 6 in the supplemental material online.

Table 5 Congruency of gait- and turning asymmetryAsymmetry in early versus advanced PD

When comparing the early versus advanced participants with PD, we did not find any significant differences in gait, turning or symptom asymmetries (p < 0.05, for details see Table 7 in the supplemental material online).

Congruency between asymmetry domainsSymptom and gait asymmetry

Considering only the 41 (step length), 36 (stance time), 32 (swing time) participants with abnormal gait asymmetry, 53,7% showed a shorter step length, 61,1% a shorter stance time and 71,9% a longer swing time on the more affected side according to the MDS-UPDRS-IIITOTAL. When considering only the MDS-UPDRS-IIILL, the agreement tended to decrease, with only about 50% with congruent asymmetry (Table 5). The agreement between asymmetry domains is visualized in Fig. 2A, where the green points represent the participants with congruent asymmetries and the red dots indicate the discrepancies. The agreement for the subgroup with abnormal gait asymmetry is presented in Fig. 2D, showing that the majority of participants show congruent gait and symptom asymmetry.

Fig. 2

A asymmetry of step length and MDS-UPDRS-IIITOTAL asymmetry, n = 94. B asymmetry of turning time and MDS-UPDRS-IIITOTAL asymmetry, n = 95. C asymmetry of turning time and asymmetry of step length, n = 93. green points: participants with consistent asymmetric sides, red points: participants with no consistent asymmetric sides, gray points: represent participants with no asymmetry in MDS-UPDRS-IIITOTAL; D, E and F represent percentages based on the ‘pathological’ sample according to the cut-off (D and F) n = 41; E n = 28). For F the side participants are turning toward is the inner leg; the side participants are turning away from is the outer leg is the ST single task, LL lower limb

Symptom- and turning asymmetry

A higher number of patients (54%) needed more time for turns when turning toward the more affected side. When considering only the MDS-UPDRS-IIILL, the results did not change substantially. The results can be found in detail in Table 4. Figure 2B illustrates the agreement, where the green points represent the participants with congruent asymmetries and the red dots indicate the discrepancies. When looking at the subgroup with abnormal turning asymmetry, relatively more people took longer for turns toward the less affected side (hence, when the more affected side is the outer leg during the turn, see Fig. 2E).

Gait- and turning asymmetry

Forty-two percent of participants with PD had a longer turning time when the side with the shorter step length was at the outer side of the turn. Yet, 55% had a longer turning time when the side with the shorter step length was at the inner side while turning (see Fig. 2C). Regarding the number of steps, 29% of participants needed more steps when the leg with the shorter step length was at the inner side of the turn. Detailed results are presented in Table 5.

When we adjusted the analysis, by looking at the ‘abnormal asymmetry’ values only, the results changed slightly. Figure 2F shows that only 37 percent of participants took more time for turns, when the leg with the shorter step length was the outer leg during the turn, whereas the absolute majority of participants (58%) took more time when the side with the longer step length was the outer leg during the turn.

Correlations between the asymmetry indexes of gait, turning, and overall symptoms

There were no significant correlations between symptom laterality (MDS-UPDRS-IIITOTAL and MDS-UPDRS-IIILL) and the gait asymmetry index or the turning asymmetry index. The results of the correlation analysis are presented in Table 8 in the supplementary material.

Sub-analysis of Freezers

The sub-analysis of congruency between asymmetry domains in Freezers also showed inconsistent results. Depending on the outcome congruency between gait asymmetry and symptom laterality ranged between 40 and 71%. For turning asymmetry and symptom asymmetry, we found more participants, taking more time for turns, when turning toward the more affected side (for details, see Tables 9, 10, 11 in supplemental material online).