DATASET - Smartphone use does not influence postural control variability in young adult university students: a cross-sectional study using nonlinear measures
Description
Dataset used to assess the association between smartphone use profile and postural control variability.
Files
Steps to reproduce
Smartphone use Smartphone use intensity was characterized by two continuous measures. The subjective measure was the total score of the Portuguese version of the Smartphone Addiction Scale (SAS-PT), a 33-item instrument covering six dimensions (daily-life disturbance, positive anticipation, withdrawal, overuse, tolerance and cyberspace-oriented relationships; range 33–198). The objective measure was screen time recorded by the built-in usage-monitoring applications of participants' own devices. Records were accessed by participants in the presence of one investigator, who verified the displayed values and excluded data from synchronized devices (e.g. tablets); all values were converted into minutes. Three variables were extracted: the previous day, the average of the preceding complete week, and the average of the three most recent complete weeks. The assessment week was excluded because it was incomplete and potentially unrepresentative. The previous-month average was used as the primary objective measure, as it best reflects habitual use. Postural control assessment Participants stood barefoot on a tri-axial force platform, which was reset before each participant. Ground reaction forces were sampled at 1000 Hz and CoP coordinates were extracted along the medio-lateral (ML) and antero-posterior (AP) axes. Four standing postures reproducing the cervical flexion angles typically adopted during smartphone use were tested: neutral (0°), 30°, 45° and 60° of cervical flexion. Cervical angles were measured with a smartphone clinometer application, previously shown to be valid and reliable for cervical range of motion, by the same examiner throughout the study. Each posture was performed once with eyes open and once with eyes closed, giving eight conditions presented in an order randomized per participant with a web-based tool. Each condition lasted 45 s, with the arms alongside the body and the feet together in the center of the platform and was followed by a 30 s rest; trials in which balance or posture was lost were repeated. A pilot test confirmed the feasibility of the protocol without modification; pilot participants were not included in the analysis. Signal processing CoP signals were processed in MATLAB R2023b following procedures previously applied to postural control data. Signals were downsampled by a factor of 20 to 50 Hz, a rate sufficient to retain the frequency content of postural signals (up to approximately 15 Hz) while attenuating high-frequency noise; the maximum frequency of the AP and ML signals was verified to exclude aliasing. Temporal variability was quantified with SampEn, which estimates the likelihood that similar sequences within a time series are repeated: lower values indicate more predictable and rigid signals, higher values more irregular and potentially more adaptable ones. Parameters followed Almeida et al.: pattern length m = 2, tolerance r = 0.2 and series length N = 2250 points (50 Hz × 45 s).