The authors have declared that no competing interests exist.
Conceived and designed the experiments: SS WvdH AM CG PH. Performed the experiments: SS WvdH AM CG. Analyzed the data: WvdH AM CG. Contributed reagents/materials/analysis tools: SS WvdH AM CG PH. Wrote the paper: SS WvdH AM CG PH.
There are concerns about the safety of texting while walking. Although evidence of negative effects of mobile phone use on gait is scarce, cognitive distraction, altered mechanical demands, and the reduced visual field associated with texting are likely to have an impact. In 26 healthy individuals we examined the effect of mobile phone use on gait. Individuals walked at a comfortable pace in a straight line over a distance of ∼8.5 m while; 1) walking without the use of a phone, 2) reading text on a mobile phone, or 3) typing text on a mobile phone. Gait performance was evaluated using a three-dimensional movement analysis system. In comparison with normal waking, when participants read or wrote text messages they walked with: greater absolute lateral foot position from one stride to the next; slower speed; greater rotation range of motion (ROM) of the head with respect to global space; the head held in a flexed position; more in-phase motion of the thorax and head in all planes, less motion between thorax and head (neck ROM); and more tightly organized coordination in lateral flexion and rotation directions. While writing text, participants walked slower, deviated more from a straight line and used less neck ROM than reading text. Although the arms and head moved with the thorax to reduce relative motion of the phone and facilitate reading and texting, movement of the head in global space increased and this could negatively impact the balance system. Texting, and to a lesser extent reading, modify gait performance. Texting or reading on a mobile phone may pose an additional risk to safety for pedestrians navigating obstacles or crossing the road.
Mobile phones are considered an essential part of everyday life, saturating all age groups and demographics. It is estimated that 77% of the world's population own a mobile phone and texting in particular, has emerged as a quick and cost effective method of communication. Although the dangers of typing text while driving have received considerable interest (e.g.
Typing and reading text on a mobile phone may modify walking as a result of the increased cognitive demand placed on working memory and executive control
To further explore the effects of mobile phone use on gait, we examined and compared the impact on gait performance and kinematics of typing and reading (without any manual input) text on a mobile phone when compared with walking without a mobile phone. We hypothesised that greater potential for cognitive distraction and modified mechanical demands associated with typing text would impact on gait performance to a greater degree than reading text.
All procedures were approved by The University of Queensland Medical Research Ethics Committee and conformed to the declaration of Helsinki. All participants provided written, informed consent.
Twenty-six healthy individuals (7 male; age 29±11 years; height 1.7±0.1 m; weight 71±13 kg, mean ± standard deviation) provided informed written consent to participate. Participants were excluded if they were less than 18 years of age, did not use a mobile phone with a touch screen and full QWERTY virtual keyboard, had less than 3 months experience with their current phone, did not use their phone on a daily basis, or if they had any neurological and/or musculoskeletal disorders that would interfere with gait. Participants were asked if they had experienced any previous accident while texting on their mobile phone and reported details regarding their typical mobile phone usage (
Variable | Data |
Handedness right ∶ left ∶ ambidextrous | 24∶1∶1 |
Typing method one handed ∶ two handed ∶ either method | 9∶15∶2 |
Phone orientation portrait ∶ landscape | 22∶4 |
Phone type iphone ∶ other | 21∶5 |
Usual use of autocorrect on ∶ off | 22∶4 |
Months of current phone use (mean ± SD) | 13.6±7.0 |
Number of minutes spent talking on a mobile phone per day (mean ± SD) | 17.7±15.9 |
Number of minutes spent texting on a mobile phone per day (mean ± SD) | 30.7±44.6 |
Number of subjects who reported prior texting related accidents | 9 |
SD – standard deviation
Three experimental conditions were included: 1) walking at a comfortable pace, 2) walking at a comfortable pace while reading a passage on a mobile phone screen with minimal manual input other than scrolling through text
In each condition participants walked in a straight line for ∼8.5 m. In the texting condition participants used their own mobile phone and their normal method of texting (one or two hands, phone held in portrait or landscape). No instruction was given regarding text accuracy and participants were free to correct their errors (or not) as they chose. However, autocorrect was turned off to allow the number of typing errors to be quantified
For movement registration, 8 cameras (T040, Vicon Motion Systems Ltd. Oxford, UK) were positioned at both sides of the walking path at ∼45 degree angle facing the direction of walking and placed ∼2 m apart. Clusters of three non-collinear reflective markers were attached to the back of the head using a head band and with double sided tape to the participant's body, at thorax (T6) and pelvis (posterior superior iliac spine). Single reflective markers were attached at the left and right heel. A reference measure, with the participant in the anatomical position facing the walking direction, allowed for alignment of cluster marker coordinate systems with the global coordinate system. The global coordinate system was defined with the positive
Right heel strikes were determined from the local vertical minima of the heel marker
Segment angles are reported as anatomically related movements (rotation (eq. 1), flexion-extension (eq. 2) and lateral flexion (eq. 3),
Relative motion between the thorax and head (neck motion), and between pelvis and thorax (trunk motion) were obtained by subtracting the time series of the relevant angles of the lower segment from the higher segment. Time series of segment angles were divided into stride cycles (from right heel strike to the following right heel strike). Within each stride cycle, the range of motion (ROM) was determined as the difference between the maximum and minimum angle, and was averaged across the stride cycles.
The average flexion angle of the head was determined as the mean of the flexion-extension time series.
Relative phase angle is a frequency domain measure, and provides information (in degrees) regarding the coordination between two segments' main component of motion (in this study; rotations at the same frequency as stride frequency) averaged over time. If two segments rotate in opposite directions, the relative phase angle is 180°, i.e. the coordination between two segments is ‘out-of-phase’. If two segments rotate together, the relative phase angle is 0°, i.e. the coordination is ‘in-phase’. The standard deviation of the relative phase angle is a measure of the spread of the relative phase angle around the mean phase. Relative phase between two segments (head – thorax and pelvis – thorax) of rotation, flexion-extension and lateral flexion movements was calculated as follows
The average position in space of the pelvis cluster markers was used to determine deviations from a straight walking path. The straight line was defined by the position of two reflective markers placed at the beginning and end of the walking path, parallel to the walls of the room. All marker positions were rotated about the
To measure the motion of the phone/arm and to verify whether changes in kinematics were related to phone use and/or could be explained by the expected reduction in walking speed with phone use, 5 participants volunteered for an extra measurement on a separate day. To control for walking speed, this experiment was performed on a treadmill (Pioneer Pro, BH Fitness Products, California, USA) at two different speeds. The speeds were matched to that selected by the participant during the control and texting conditions of the main experiment. Kinematic data were collected as per the first experiment at both speeds while participants performed 3 tasks: control walking, reading and texting (randomised). To verify whether texting requires an additional mechanical demand above that required for reading (e.g. maintenance of phone position with respect to the head, fixation of the arms with the thorax), data of phone and elbow position were also collected using additional markers.
To test control of the position of the phone with respect to the head, the total distance moved by the phone (path) in three-dimensions per second was calculated: 1) in the global reference frame and 2) with respect to the head (after the coordinates of the phone were transformed into the head reference frame). To test whether arm movement was more constrained with respect to trunk motion (to hold phone still), the relative phase between the forward-backward arm movement and thorax rotations was calculated.
All outcome variables were averaged across the three repetitions within each condition (walk; text; read). To ensure normal distribution, data were log transformed if Shapiro-Wilk test for normality was significant (
Demographic and mobile phone usage data are presented in
Participants walked at a slower speed during reading and texting than when walking without the mobile phone, and walked slower during texting than reading (
ANOVARM | Post hoc analyses | Mean (±SD) | ||||||
Outcome measure | F ratio | P-value | Walk vs. Read | Walk vs. Text | Read vs. Text | Walk | Read | Text |
Walking speed (m/s) | 85.12 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 1.33 (0.15) | 1.16 (0.14) | 1.01 (0.17) |
Stride length (m) | 110.94 | 0.0000 | 0.0000 | 0.0000 | 0.0000 | 1.35 (0.12) | 1.23 (0.09) | 1.15 (0.09) |
Stride frequency (Hz) | 49.14 | 0.0000 | 0.0002 | 0.0000 | 0.0000 | 0.99 (0.06) | 0.95 (0.08) | 0.88 (0.11) |
Abs path lateral direction (m) | 13.23 | 0.0000 | 0.874 | 0.0000 | 0.0011 | 0.07 (0.02) | 0.08 (0.02) | 0.10 (0.03) |
Delta right foot position (m/stride) | 14.12 | 0.0000 | 0.0041 | 0.0000 | 0.2111 | 0.03 (0.01) | 0.04 (0.01) | 0.04 (0.01) |
Participants deviated more from a straight line during reading and texting than during the walking task (
The left hand side depicts the absolute medial-lateral deviations from the straight line. The right hand side depicts the absolute change in lateral foot position from one stride to the next of the right foot. The absolute change in lateral foot position per stride was greater during reading and texting than walking, but did not differ between the two phone tasks.
Participants looked at their phone for reading and texting with a flexed head position, and the angle of flexion did not differ between these conditions (see
Note the increase in range of head rotation in relation to the global reference frame during reading and texting with reduction of phase angle and phase variability between thorax and head. The dashed vertical grey lines denote right heel strikes.
ANOVARM | Post hoc analyses | Mean (±SD) | ||||||
Outcome measure | F ratio | P-value | Walk vs. Read | Walk vs. Text | Read vs. Text | Walk | Read | Text |
Head flexion position | 168.06 | 0.0000 | 0.0000 | 0.0000 | 0.1881 | 0.47 (5.63) | 29.22 (9.12) | 31.80 (10.76) |
Head flexion-extension | 44.3 | 0.0000 | 0.0001 | 0.0000 | 0.0001 | 6.65 (2.04) | 5.20 (1.46) | 4.20 (1.47) |
Head lateral flexion | 8.35 | 0.0015 | 0.0005 | 0.1609 | 0.1196 | 4.51 (1.74) | 6.06 (2.57) | 5.23 (2.15) |
Head rotation | 12.47 | 0.0003 | 0.0000 | 0.0048 | 0.3849 | 4.75 (1.67) | 6.57 (2.75) | 5.96 (2.56) |
Thorax flexion-extension | 19.59 | 0.0000 | 0.0018 | 0.0000 | 0.0411 | 3.78 (1.18) | 3.38 (1.11) | 3.06 (0.74) |
Thorax lateral flexion | 13.34 | 0.0000 | 0.2776 | 0.0000 | 0.0045 | 5.51 (2.11) | 4.90 (1.38) | 4.20 (1.38) |
Thorax rotation | 2.96 | 0.0762 | 6.33 (1.56) | 7.26 (2.41) | 7.00 (2.40) | |||
Pelvis flexion-extension | 5.05 | 0.0132 | 0.7079 | 0.0083 | 0.1711 | 6.68 (4.25) | 5.90 (2.76) | 5.35 (2.17) |
Pelvis lateral flexion | 31.29 | 0.0000 | 0.0002 | 0.0000 | 0.0029 | 12.14 (4.05) | 10.79 (3.97) | 9.70 (3.23) |
Pelvis rotation | 10.62 | 0.0008 | 0.0037 | 0.0002 | 1.0000 | 15.73 (8.11) | 11.59 (5.20) | 10.49 (3.89) |
Neck flexion-extension | 45.35 | 0.0000 | 0.0000 | 0.0000 | 0.0001 | 7.05 (2.37) | 5.10 (1.68) | 3.92 (1.64) |
Neck lateral flexion | 28.55 | 0.0000 | 0.0005 | 0.0000 | 0.0033 | 5.63 (1.97) | 4.07 (1.54) | 3.11 (1.04) |
Neck rotation | 21.38 | 0.0000 | 0.0212 | 0.0000 | 0.0016 | 5.41 (1.50) | 4.60 (1.30) | 3.68 (1.06) |
Trunk flexion-extension | 6.7 | 0.0045 | 0.2720 | 0.0018 | 0.1767 | 7.64 (4.66) | 6.63 (2.85) | 5.99 (2.45) |
Trunk lateral flexion | 23.35 | 0.0000 | 0.0098 | 0.0000 | 0.0014 | 15.72 (4.69) | 14.08 (3.87) | 12.10 (3.68) |
Trunk rotation | 40.28 | 0.0000 | 0.0003 | 0.0000 | 0.0001 | 16.89 (8.99) | 13.77 (6.44) | 11.01 (4.86) |
ROM of anterior and posterior tilt of the pelvis (sagittal plane) and flexion-extension ROM between the pelvis and thorax were similar during reading and walking, but reduced during texting. In the global frame of reference, pelvic lateral flexion (frontal plane) ROM was lower during reading, and further reduced during texting, when compared to walking. Pelvic rotation (transverse plane) ROM was lower during reading and texting than walking. The rotation ROM of the trunk was lower during reading and texting than walking and ROM during texting was lower than reading, consistent with tighter mechanical constraint between these segments when manipulating the phone in the hands. The flexion-extension phase angle between the pelvis and thorax was less during reading than walking, and lateral flexion phase angle was reduced to a greater extent during texting than reading or walking (
ANOVARM | Post hoc analyses | Mean (±SD) | ||||||
Outcome measure | F ratio | P-value | Walk vs. Read | Walk vs. Text | Read vs. Text | Walk | Read | Text |
Flexion-extension | 13.09 | 0.0000 | 0.0006 | 0.0001 | 1.0000 | 90.11 (37.17) | 57.09 (32.86) | 51.35 (32.42) |
Lateral flexion | 18.92 | 0.0000 | 0.0000 | 0.0000 | 1.0000 | 51.21 (33.00) | 22.28 (19.66) | 20.21 (13.55) |
Rotation | 9.24 | 0.0010 | 0.0121 | 0.0004 | 0.7707 | 34.99 (28.02) | 18.97 (17.91) | 14.06 (12.47) |
Flexion-extension | 3.77 | 0.0342 | 0.0548 | 0.0754 | 1.0000 | 76.41 (20.07) | 64.74 (17.98) | 65.37 (19.96) |
Lateral flexion | 19.56 | 0.0000 | 0.0000 | 0.0000 | 1.0000 | 28.68 (21.75) | 12.88 (10.16) | 13.04 (11.38) |
Rotation | 17.58 | 0.0000 | 0.0000 | 0.0000 | 1.0000 | 21.53 (9.48) | 13.46 (12.72) | 11.96 (8.26) |
Flexion-extension | 4.44 | 0.0185 | 0.0137 | 0.6256 | 0.2882 | 97.69 (47.16) | 70.90 (41.51) | 86.87 (38.77) |
Lateral flexion | 11.58 | 0.0002 | 0.3228 | 0.0001 | 0.0095 | 133.06 (31.32) | 127.03 (35.41) | 115.32 (42.23) |
Rotation | 3.46 | 0.0691 | 98.19 (44.70) | 96.00 (50.48) | 82.05 (48.50) | |||
Flexion-extension | 4.78 | 0.0128 | 0.4817 | 0.0098 | 0.3068 | 51.92 (17.05) | 57.09 (18.87) | 63.13 (18.65) |
Lateral flexion | 5.6 | 0.0089 | 0.1570 | 0.0050 | 0.5605 | 8.49 (5.44) | 9.33 (4.58) | 10.47 (4.89) |
Rotation | 0.34 | 0.6484 | 18.32 (11.93) | 17.56 (11.43) | 16.49 (11.16) |
When the phone was used for reading or texting with walking speed controlled on a treadmill to match that used during the individuals' overground walking and texting conditions, the following variables were found to be altered as a result of mobile phone use and not walking speed (
Normal walking speed | Walking speed while texting | |||||
Outcome measure | Walk | Read | Text | Walk | Read | Text |
Delta right foot position (m/stride) | 0.015 (0.002) | 0.018 (0.003) | 0.020 (0.003) | 0.017 (0.003) | 0.017 (0.002) | 0.019 (0.003) |
Head flexion position (°) | 2.67 (1.96) | 27.22 (6.72) | 31.92 (8.68) | 2.38 (2.25) | 27.46 (7.62) | 32.37 (10.44) |
Head flexion-extension | 4.59 (1.35) | 4.22 (0.61) | 3.46 (0.40) | 4.64 (0.79) | 3.98 (0.47) | 3.50 (0.58) |
Head rotation | 3.77 (1.40) | 6.94 (2.00) | 6.40 (1.38) | 4.80 (1.36) | 7.03 (2.04) | 6.54 (1.18) |
Neck rotation | 5.98 (1.99) | 4.79 (1.32) | 4.45 (1.41) | 6.24 (2.42) | 4.13 (1.11) | 4.26 (0.92) |
Thorax head rotation | 28.04 (7.95) | 15.85 (11.49) | 15.15 (2.60) | 21.51 (5.75) | 10.56 (1.51) | 10.41 (5.42) |
Arm swing thorax rotation | 45.96 (27.02) | 17.36 (6.45) | 15.37 (7.47) | 29.45 (20.57) | 10.92 (5.36) | 12.02 (5.17) |
irt Global frame | 0.079 (0.012) | 0.073 (0.009) | 0.068 (0.012) | 0.059 (0.005) | ||
irt Head frame | 0.039 (0.011) | 0.036 (0.015) | 0.033 (0.008) | 0.032 (0.011) |
Data (mean ± standard deviation) are shown for the additional treadmill experiment. The outcome measures that were affected by mobile phone use and not walking speed are shown.
The additional analysis revealed that the phase angle between the forward-backward motion of the elbow and rotation of the thorax was smaller (moved almost ‘in-phase’), when participants manipulated the phone for reading or texting while walking on a treadmill than walking without a phone (Fcondition = 7.55,
This study is the first to compare the impact of typing text on a mobile phone on gait performance and kinematics against that associated with reading text on a phone and walking without constraint, and without any additional restriction of field of view. Evaluation of gait performance revealed that individuals walk slower, demonstrate greater absolute medial-lateral step deviation, increase rotation ROM of the head with respect to the global reference frame, walk with a flexed head position, reduce neck ROM, and move the thorax and head more in-phase with reduced phase variability, during texting and reading than unconstrained walking. Differences between typing and reading text were less pronounced, but typing text was associated with slower walking speed, greater deviation from a straight line, more ‘in-phase’ lateral flexion motion between the thorax and pelvis and generally reduced ROM of the neck compared to reading text on a mobile phone. Furthermore, while reading, phase angle between pelvis and thorax flexion-extension was reduced. These findings are similar to those observed in previous studies. For instance, Lamberg and Muratori
As participants walked slower while reading and reduced speed further while texting, some changes in gait kinematics may be explained by reduced speed. The additional experiment performed on a treadmill was conducted to evaluate this confounding effect. Participants walked at their normal (control) and texting speed, derived from the overground walking experiment. The following variables were less likely to be affected by reduced speed, and more likely to be related to the effect of dual tasking with a phone: 1) phone movement closely related to head movement, which likely makes it easier to read or type a message on a phone and 2) motion of the arms was closely related to thorax rotation, which is likely to reduce the number of degrees of freedom controlled by central nervous system. The resultant coupling of motion of the arms (and phone), thorax and head would maintain the phone in a steady position in the visual field. Although the reduced phase angle and almost in-phase coordination between head and thorax rotation would facilitate steadiness of the phone for reading, this has negative consequences, as head stability in the global reference frame is compromised. This strategy to optimise the phone task, may compromise the accuracy of head control and impact on balance performance. This hypothesis is supported by increased medial lateral head motion of ∼1.5 degrees during texting and reading in the current study which, although small, exceeds the threshold for detection of sway with proprioceptive, visual and vestibular systems in humans
A key finding was reduced neck ROM (head relative to thorax) in all planes during reading, and to a greater extent with typing text. The head moved more ‘in-phase’ with the thorax, and coordination between segments was less variable (lower phase angle variability) in lateral flexion and rotation directions. These findings imply the head is controlled in a manner that constrains its relationship with the thorax, most likely to optimize the relationship between the eyes, trunk/arm and phone. This is supported by our observation that the arms were ‘locked’ to the thorax, such that the phone moved together with the thorax, in the overground experiment and confirmed in the treadmill experiment where forward-backward arm swing shifted to an almost ‘in-phase’ relationship with thorax rotation. Motion of the arm was more ‘out-of-phase’ when walking on a treadmill without the phone. Phone movement with respect to the head frame of reference was lower than in the global frame of reference. Reduced arm swing can negatively impact on walking balance. For instance, arm swing reduces angular momentum about the vertical axis
Changes in gait associated with mobile phone use may undermine functional walking and impact on safety in common pedestrian environments. Individuals with constrained movement patterns
The gait kinematic most likely to impact on safety was the deviation from a straight walking path during typing and reading text on a mobile phone. In a pedestrian environment inability to maintain a straight path would be likely to increase potential for collisions, trips and traffic accidents. There are two plausible mechanisms for the inability to maintain a straight walking path during texting and reading. First, reduced awareness of the visual field would limit use of external cues to guide path, and second the greater head motion relative to the global reference frame (but greater constraint to the trunk) may reduce the utility of vestibular information. Vestibular input is essential for accurate navigation during walking
This study is the first to compare the impact of typing and reading text on a mobile phone on gait performance. We demonstrate slower walking speed, greater deviation from a straight path and increase absolute lateral step deviation in conjunction with increased rotation ROM of the head in global space, reduced relative motion and greater ‘in-phase’ motion of the head during typing, and to a lesser extent, reading text on a mobile phone than normal walking. These altered gait parameters may have an impact on the safety of pedestrians who type or read text on a mobile phone while walking.