Time estimation during motor activity

Abstract

Several studies on time estimation showed that the estimation of temporal intervals is related to the amount of attention devoted to time. This is explained by the scalar timing theory, which assumes that attention alters the number of pulses transferred by our internal clock to an accumulator that keeps track of the elapsed time. In a previous study, it was found that time underestimation during cognitive-demanding tasks was more pronounced while walking than while sitting, whereas no clear motor-induced effects emerged without a concurrent cognitive task. What remains unclear then is the motor interference itself on time estimation. Here we aim to clarify how the estimation of time can be influenced by demanding motor mechanisms and how different motor activities interact with concurrent cognitive tasks during time estimation. To this purpose, we manipulated simultaneously the difficulty of the cognitive task (solving arithmetic operations) and the motor task. We used an automated body movement that should require no motor or mental effort, a more difficult movement that requires some motor control, and a highly demanding movement requiring motor coordination and attention. We compared the effects of these three types of walking on time estimation accuracy and uncertainty, arithmetic performance, and reaction times. Our findings confirm that time estimation is affected by the difficulty of the cognitive task whereas we did not find any evidence that time estimation changes with the complexity of our motor task, nor an interaction between walking and the concurrent cognitive tasks. We can conclude that walking, although highly demanding, does not have the same effects as other mental tasks on time estimation.

Keywords: duration estimation; motor load; motor-cognitive interference; time estimation; walking.

FIGURE 1

Procedure and tasks. (A) Motor conditions. From left to right: forward regular-speed walking, forward irregular-speed walking, backward irregular-speed walking. (B) Easy task. (C) Hard task. In the instructions, the red word indicates the cognitive task to be performed. The green circle indicates the beginning of time estimation; the red circle indicated the stop of time estimation.

FIGURE 2

Time estimation in different motor and cognitive tasks. Each colored fit line is the average time estimate of the same participant. Dots indicate the forward regular-speed walking condition (left panels); triangles indicate the forward irregular-speed walking condition (middle panels); squares indicate the backward irregular-speed walking condition (right panels). Dashed black lines represent exact estimations. (A) Easy task. (B) Medium task. (C) Hard task.

FIGURE 3

Estimation uncertainty as a function of duration. Estimation uncertainty increases with duration, similarly for each cognitive task and motor condition. The graphs show the root mean square errors (RMSE), computed on 21 s intervals, for all cognitive and motor tasks with their best-fit curves. Solid lines and dots represent the forward regular-speed walking condition; dotted lines and triangles represent the forward irregular-speed walking condition; dashed lines and squares represent the backward irregular-speed walking condition. (A) Easy task (number of occurrences in each bin—forward regular walking: 67, 27, 26, 23, and 82; forward irregular walking: 63, 29, 44, 37, and 52; backward walking: 69, 35, 26, 42, and 53). Goodness of fit—forward regular walking: R² = 0.7; forward irregular walking: R² = 0.9; backward walking: R² = 0.9. (B) Medium task (number of occurrences in each bin—forward regular walking: 47, 17, 45, 49, and 66; forward irregular walking: 63, 36, 53, 24, and 49; backward walking: 48, 29, 53, 42, and 52). Goodness of fit – forward regular walking: R² = 0.8; forward irregular walking: R² = 0.9; backward walking: R² = 0.7. (C) Hard task (number of occurrences in each bin—forward regular walking: 72, 37, 15, 53, and 48; forward irregular walking: 54, 44, 30, 46, and 51; backward walking: 64, 22, 29, 44, and 66). Goodness of fit—forward regular walking: R² = 0.9; forward irregular walking: R² = 0.9; backward walking: R² = 0.7.

FIGURE 4

Averaged estimation bias in different motor and cognitive tasks. Underestimation increases with cognitive load but does not change with the difficulty of the motor task. The graphs show the estimation bias, computed on 21 s intervals, for all cognitive and motor tasks with their best-fit curves. Error bars are SE across participants. Solid lines and dots represent the forward regular-speed walking condition; dotted lines and triangles represent the forward irregular-speed walking condition; dashed lines and squares represent the backward irregular-speed walking condition. Dashed black lines represent exact estimations. (A) Easy task (forward regular-speed walking: slope = –0.04 ± 0.04, intercept = 4.4 ± 2.2; forward irregular-speed walking: slope = 0.004 ± 0.05, intercept = 1.7 ± 3.1; backward irregular-speed walking: slope = 0.01 ± 0.03, intercept = –0.6 ± 1.6). (B) Medium task (forward regular-speed walking: slope = –0.2 ± 0.03, intercept = 3.1 ± 1.3; forward irregular-speed walking: slope = –0.15 ± 0.05, intercept = 4.9 ± 3.2; backward irregular-speed walking: slope = –0.2 ± 0.03, intercept = 1.6 ± 1.8). (C) Hard task (forward regular-speed walking: slope = –0.3 ± 0.08, intercept = 5.6 ± 4.2; forward irregular-speed walking: slope = –0.2 ± 0.04, intercept = 4.2 ± 2.8; backward irregular-speed walking: slope = –0.3 ± 0.06, intercept = 3.6 ± 3.1).

FIGURE 5

Performance for math operations in the three motor conditions. Correct responses and reaction times change with cognitive load but do not change with the difficulty of the motor task. Green: medium task; orange: hard task. Solid bars correspond to the forward regular-speed walking; dotted bars correspond to the forward irregular-speed walking; striped bars correspond to the backward irregular-speed walking. (A) Percentage of correct responses. (B) Response times. Asterisks mark statistically significant differences with ANOVAs: ^**p < 0.01, ^***p < 0.001. Error bars are SE across participants.