In the following section, only step width, step length, and stride length recorded during single- and dual-task walking as well as their respective CVs are shown. Further gait parameters (double support time, stride duration, stance duration, swing duration, stance phase, swing phase, cadence) can be found in the supplementary data (Tab. S4 & S5). In addition, only main and/or interaction effects related to time are listed. All further statistical results of the ANOVA can also be found in the supplementary data (Tab. S10).

Single-task gait performance

A significant main effect of time was found for step width (P = 0.015, F1,44 = 6.417, ƞp2 = 0.127). For the older adults, post-hoc analysis revealed a significant decrease in step width after performing the Stroop intervention task (P = 0.001, diff.: −0.010 m (−0.016 m to −0.005 m), d = 0.807) and no reductions after reading (P = 0.133, diff.: −0.003 m (−0.007 m to 0.001 m), d = 0.267). Young adults showed statistical tendencies towards a significant decrease in step width after the Stroop intervention task (P = 0.093, diff.: −0.005 m (−0.010 m to 0.001 m), d = 0.407) as well as after reading (P = 0.092, diff.: −0.004 m (−0.008 m to 0.001 m), d = 0.496).

A significant main effect of time was found for step length (P = 0.012, F1,44 = 6.932, ƞp2 = 0.136). Furthermore, a time × group interaction (P = 0.018, F1,44 = 6.033, ƞp2 = 0.121) and a statistical tendency for condition × time interaction (P = 0.068, F1,44 = 3.494, ƞp2 = 0.074) for step length was observed. Post-hoc analysis indicated significant increases in step length after the Stroop intervention task in the older participants (P = 0.013, diff.: 0.005 m (0.001 m to 0.009 m), d = 0.466) as well as after the reading control task (P = 0.000, diff.: 0.015 m (0.008 m to 0.023 m), d = 0.632) (Fig. 4a). For the CVstep length a statistical tendency for a condition × time × group interaction (P = 0.091, F1,44 = 2.995, ƞp2 = 0.064) was found. For the older adults, post-hoc analysis revealed no significant reduction of CVstep length after performing the Stroop intervention task (P = 0.418, diff.: 0.092% (−0.135% to 0.320%), d = 0.140) but after reading (P = 0.007, diff.: −0.352% (−0.603% to −0.101%), d = 0.492) (Fig. 5a).

Fig. 4

Spatio-temporal gait parameters (step width, step length, and stride length) for Single-Task (a) and Dual-Task: Word list generation task (b), Arithmetic task (c), and Stroop-Task (d) for the young and old participants recorded before (PRE) and after (POST) the interventions (Stroop intervention task (Stroop), reading control task (Reading)). * P ≤ 0.050, † P ≤ 0.100

Fig. 5

Coefficient of variation (CV) of the spatio-temporal gait parameters (step width, step length, and stride length) for Single-Task (a) and Dual-Task: Word list generation task (b), Arithmetic task (c), and Stroop-Task (d) for the young and old participants recorded before (PRE) and after (POST) the interventions (Stroop intervention task (Stroop), reading control task (Reading)). * P ≤ 0.050, † P ≤ 0.100

A significant time × group interaction was found for stride length (P = 0.024, F1,44 = 5.452, ƞp2 = 0.110). Post-hoc analysis showed, similar to the step length, significant increases in stride length after the Stroop intervention task (P = 0.034, diff.: 0.010 m (0.001 m to 0.018 m), d = 0.419) and after reading (P = 0.000, diff.: 0.022 m (0.011 m to 0.034 m), d = 0.659) only for the older participants (Fig. 4a).

Dual-task gait performanceWord list generation task (cognitive flexibility)

A statistical tendency was found for the main effect time for step width (P = 0.084, F1,44 = 3.130, ƞp2 = 0.066). For the young adults, post-hoc analysis indicated a significant reduction after performing the Stroop intervention task (P = 0.000, diff.: −0.009 m (−0.014 m to −0.005 m), d = 0.888) as well as after the reading control task (P = 0.010, diff.: −0.007 m (−0.013 m to −0.002 m), d = 0.619). Older adults showed a similar reduction in step width after the Stroop intervention task (P = 0.001, diff.: −0.009 m (−0.014 m to −0.004 m), d = 0.782) and after the reading task (P = 0.035, diff.: −0.006 m (−0.011 m to 0.000 m), d = 0.433) (Fig. 4b). A significant main effect of time (P = 0.040, F1,44 = 4.490, ƞp2 = 0.093) and a statistical tendency for a condition × time interaction (P = 0.076, F1,44 = 3.311, ƞp2 = 0.070) was found for the CVstep width. Post-hoc analysis revealed a statistical tendency towards an increased CVstep width after the Stroop intervention task (P = 0.099, diff.: 2.943% (−0.573% to 6.459%), d = 0.249) as well as after the reading control task (P = 0.055, diff.: 1.789% (−0.038% to 3.617%), d = 0.489) for the older adults (Fig. 5b).

Arithmetic task (working memory)

A significant main effect of time was found for step width (P = 0.001, F1,44 = 13.841, ƞp2 = 0.239). For the young participants, post-hoc analysis showed a significant reduction after the Stroop intervention task (P = 0.008, diff.: −0.007 m (−0.013 m to −0.002 m), d = 0.737) as well as after reading (P = 0.001, diff.: −0.008 m (−0.012 m to −0.004 m), d = 0.949). These significant decreases were also observed for the older participants after the Stroop intervention task (P = 0.000, diff.: −0.011 m (−0.017 m to −0.006 m), d = 0.717) and after reading (P = 0.000, diff.: −0.008 m (−0.013 m to −0.004 m), d = 0.467) (Fig. 4c). A significant main effect of time (P = 0.001, F1,44 = 12.031, ƞp2 = 0.215) as well as a statistical tendency for condition × time × group interaction (P = 0.090, F1,44 = 2.998, ƞp2 = 0.064) was found for the CVstep width. For the young adults, post-hoc analysis revealed no significant increase after the Stroop intervention task (P = 0.135, diff.: 1.446% (−0.467% to 3.359%), d = 0.412) but a statistical tendency towards an increase of the CVstep width after reading (P = 0.052, diff.: 1.630% (−0.017% to 3.277%), d = 0.389). For the older adults, a significant increase after the Stroop intervention task (P = 0.006, diff.: 2.805% (0.850% to 4.760%), d = 0.472) but no significant increase after reading was found (P = 0.680, diff.: 0.347% (−1.336% to 2.030%), d = 0.041).

A statistical tendency for a condition × time interaction (P = 0.083, F1,44 = 3.144, ƞp2 = 0.067) was found for the CVstep length. Post-hoc analysis revealed, only for the older adults, a significant decrease in CVstep length after performing the Stroop intervention task (P = 0.008, diff.: −0.411% (−0.710% to −0.113%), d = 0.451) as well as after reading (P = 0.001, diff.: −0.468% (−0.746% to −0.191%), d = 0.564) (Fig. 5c).

Stroop-task (response inhibition)

A significant main effect of time was found for step width (P = 0.004, F1,44 = 9.295, ƞp2 = 0.174). Post-hoc analysis showed, for the young adults, a significant reduction in step width after performing the Stroop intervention task (P = 0.003, diff.: −0.006 m (−0.011 m to −0.002 m), d = 0.816) and after reading (P = 0.000, diff.: −0.009 m (−0.013 m to −0.004 m), d = 1.298). Significant reductions in step width were also observed for the older adults after the Stroop intervention task (P = 0.000, diff.: −0.011 m (−0.015 m to −0.007 m), d = 1.019) as well as after reading (P = 0.000, diff.: −0.011 m (−0.015 m to −0.006 m), d = 0.789) (Fig. 4d). A significant main effect of time (P = 0.023, F1,44 = 5.590, ƞp2 = 0.113) was found for the CVstep width. Post-hoc analysis showed for the young participants a significant increase after performing the Stroop intervention task (P = 0.043, diff.: 2.064% (0.071% to 4.057%), d = 1.151) bot not after reading (P = 0.165, diff.: 2.033% (−0.871% to 4.936%), d = 0.780). For the older participants, a significant increase after the Stroop intervention task (P = 0.003, diff.: 3.151% (1.114% to 5.188%), d = 0.485) and a statistical tendency toward an increase of the CVstep width after reading was found (P = 0.096, diff.: 2.504% (−0.464% to 5.472%), d = 0.236) (Fig. 5d).

A statistical tendency for a condition × time × group interaction was found for the step length (P = 0.051, F1,44 = 4.021, ƞp2 = 0.084). Post-hoc analysis showed, for the young adults, a significant increase in step length after performing the Stroop intervention task (P = 0.048, diff.: 0.006 m (0.000 m to 0.013 m), d = 0.854) but not after reading (P = 0.361, diff.: 0.003 m (−0.003 m to 0.008 m), d = 0.462). For the older adults, no significant increase in step length after the Stroop intervention task (P = 0.155, diff.: 0.005 m (−0.002 m to 0.011 m), d = 0.251) but a significant increase after reading (P = 0.000, diff.: 0.012 m (0.006 m to 0.017 m), d = 0.761) was observed (Fig. 4d). A significant time × group interaction was found for the CVstep length (P = 0.020, F1,44 = 5.839, ƞp2 = 0.117). Post-hoc analysis showed, only for the older adults, a significant reduction for the CVstep length after the Stroop intervention task (P = 0.002, diff.: −0.453% (−0.725% to −0.180%), d = 0.557) as well as after reading (P = 0.000, diff.: −0.535% (−0.806% to −0.265%), d = 0.649) (Fig. 5d).

A statistical tendency for condition × time × group interaction was found for the stride length (P = 0.085, F1,44 = 3.098, ƞp2 = 0.066). Post-hoc analysis revealed, for the older adults, no significant increase in stride length after performing the Stroop intervention task (P = 0.240, diff.: 0.008 (−0.005 to 0.021), d = 0.225) but after reading (P = 0.000, diff.: 0.021 m (0.011 m to 0.030 m), d = 0.742) (Fig. 4d). A significant time × group interaction (P = 0.015, F1,44 = 6.379, ƞp2 = 0.127) was observed for the CVstride length. For the older participants, post-hoc analysis indicated a significant decrease of the CVstride length after the Stroop intervention task (P = 0.021, diff.: −0.271% (−0.499% to −0.042%), d = 0.391) and after reading (P = 0.000, diff.: −0.396% (−0.603% to −0.189%), d = 0.659) (Fig. 5d).

Cognitive interference task performanceSingle task performance

A significant time × group interaction (P = 0.010, F1,44 = 7.205, ƞp2 = 0.141) and condition × time interaction (P = 0.042, F1,44 = 4.389, ƞp2 = 0.091) was found for cognitive performance during the word list generation task under single-task condition. Post-hoc analysis showed, for the young adults, a significant increase in the number of words after the Stroop intervention task (P = 0.000, diff.: 3.097 (1.791 to 4.404), d = 0.683) but not after reading (P = 0.272, diff.: 0.535 (−0.435 to 1.505), d = 0.325). For the older adults, a significant increase in the number of words was found after the Stroop intervention task (P = 0.012, diff.: 1.746 (0.411 to 3.082), d = 0.304) and a significant decrease after the reading control task (P = 0.008, diff.: −1.362 (−2.354 to −0.371), d = 0.370).

Dual-task performance

During dual-task walking, no significant changes in cognitive interference task performance associated with the main factor time were found. The results for the other comparisons can be found in the supplementary data (Tab. S10).

Cognitive performance measureTrail Making Test (TMT)

A significant time × group interaction (P = 0.015, F1,44 = 6.450, ƞp2 = 0.130) and main effect of group (P = 0.000, F1,44 = 42.087, ƞp2 = 0.495) was found for the TMT Part B. Post-hoc analysis showed, for the older participants, a significant reduction of time to complete Part B of the TMT after the Stroop intervention task (P = 0.040, diff.: −6.001 s (−11.702 s to −0.301 s), d = 0.257) as well as after reading (P = 0.000, diff.: −11.964 s (−18.184 s to −5.744 s), d = 0.671). For the young participants, no significant reduction was observed after the Stroop intervention task (P = 0.583, diff.: −1.496 s (−6.945 s to 3.954 s), d = 0.185) but after reading (P = 0.046, diff.: −6.062 s (−12.009 s to −0.116 s), d = 0.920). No other main effects or interactions were found for the TMT Part B (supplementary data, Tab. S10). A statistical tendency for time × group interaction was found for the TMT (B-A) (P = 0.062, F1,44 = 3.662, ƞp2 = 0.078). Post-hoc analysis showed no significant reduction after the Stroop intervention task (P = 0.427, diff.: −2.491 s (−8.754 s to 3.772 s), d = 0.055) but after reading (P = 0.002, diff.: −10.085 s (−16.399 s to −3.772 s), d = 0.563) for the older adults. For the young adults, no changes were found after both the Stroop intervention task as well as after the reading control task. No other main effects or interactions were observed for the TMT (B-A) (supplementary data, Tab. S10).

Stroop intervention task performance

A statistical tendency for a time × group interaction (P = 0.068, F5,150 = 2.102, ƞp2 = 0.065) was found for the number of tasks. Post-hoc analysis showed, for the young participants, a significant increase in number of tasks in the period from 10 to 15 min in (P = 0.016, diff.: 18.135 (2.223 to 34.047), d = 0.642). No other significant changes were observed for the young and old participants. For the reaction time during the Stroop intervention task, a significant main effect of time (P = 0.005, F5,150 = 3.467, ƞp2 = 0.104) was found. For the young adults, post-hoc analysis revealed a significant decrease in reaction time in the period from 10 to 15 min (P = 0.036, diff.: −0.071 s (−0.138 s to −0.003 s), d = 0.577), corresponding to the increased number of tasks during this time period. No further main effects nor interactions were found for reaction time as well as accuracy during the sustained Stroop intervention task (supplementary data, Tab. S10).