In this study, we present an algorithm for the assessment of one's own perception of balance (equilibrioception). Upright standing position is maintained by continuous updating and integration of vestibular, visual and proprioceptive information, so that a compensatory reaction can be implemented
when perturbations occur. This ability to monitor and maintain balance can be considered as a physiological sense, so, as for the other senses, it is fair to assume that healthy people can perceive and evaluate differences between balance states. The aim of this study is to investigate how changes in stabilometric parametres are perceived by young, healthy adults. Participants were asked to stand still on a Wii Balance Board (WBB) with feet in a constrained position; 13 trials of 30 s each were performed by each subject, the order of Eyes Open (EO) and Eyes Closed (EC) trials being semi-randomized. At the end of each trial (except the first one), participants were asked to judge if their performance was better or worse than the one in the immediately preceding trial. SwayPath ratio data were used to calculate the Just Noticeable Difference (JND) between two consecutive trials, which was of 0.2 when participants improved their performance from one trial to the next, and of 0.4 when performance on a trial was worse than in the previous one. This “need” of a bigger difference for the worsening to be perceived seems to suggest a tendency towards overestimation of one’s own balance. Interestingly, participants’ judgement was more reliable when evaluating consecutive EC rather than EO trials, at least when performance was worsening.
Falls are a major public health problem with 30% of older people falling at least once a year.
Falling is associated with increased mortality, injuries,
loss of independence and adverse psychosocial
consequences [
In everyday life, standing balance is a rather simple
task regulated automatically by subcortical nervous
structures and spinal motor neuron pools [
When posture is disturbed, evaluation of afferent
multisensory information allows compensatory
reactions to be implemented [
The use of objective measures for the assessment of
standing balance has widely taken place over the last
years [
Material and Methods Participants 78 healthy individuals participated to the study, 29 males and 49 females (mean age = 23.1, SD = ±2.5). Participants were all normal weight and did not report any history of neurological diseases, orthopedic pathologies, use of medication or temporary problems that may have influenced the results of standing balance tests. The study was approved by the Institution’s Research Ethics Committee; all participants provided written informed consent.
Each subject performed 13 trials of 30 seconds each in a constrained foot position. Half of the participants performed an EO first trial and the other half performed an EC first trial. 6 out of the 12 remaining trials were performed with participants keeping their eyes open (EO) and 6 with their eyes closed (EC); the sequence of EO-EC trials was semi-randomized, so that 4 possible conditions could be investigated: EO trial followed by another EO trial (EO-EO); EC trial followed by another EC trial (EC-EC), EO trial followed by an EC trial (EOEC), EC trial followed by an EO trial (EC-EO). For all conditions, subjects were asked to place their feet according to the lines designed on the surface of the platform: heels had to be kept together where lines touched while toes were to be placed on top of the lines, so to form a 30° angle. Once in the correct position, participants were asked to maintain a relaxed upright standing position with their arms along the body. For EO trials, a fixation cross was placed at 3m distance and adjusted according to participants’ height, so that each subject was looking straight ahead during trials. For EC trials, participants were instructed to look at the fixation cross before closing their eyes, so that head position was the same across trials.
At the end of each trial, except the first one, subjects were asked to judge if their performance on the present trial was better or worse than their performance on the immediately preceding one. Subjects were given approximately 1 minute to relax between one trial and the following. The examiner always waited for participants to announce their intention to resume the test, stop talking and eventually close their eyes before proceeding with data acquisition. Trials were performed in a non-noisy environment.
Data were recorded with the Nintendo Wii Balance
Board (WBB) using Matlab [
A total of 1010 trials was recorded without any malfunction or technical problem leading to the test being invalidated.
In order to assess participants’ sense of balance, responses given by subjects to their own SwayPath parameter (SP) were compared. Since participants were asked to evaluate their own performance in subsequent trials, the ratio between parameters recorded in subsequent trials was considered according to the following formula: RSP=log2(SPi/SPi-1),where i=2,3…,13 is the considered trial.
Smaller values for SP indicate a better balance performance, while higher values indicate a worsened balance performance. Hence, when evaluating subsequent trials, we can say that if the RSP value is smaller than 0, performance in the latter trial is better than the one preceding it, while if RSP shows values bigger than 0, performance in the last trial was worse than performance in the immediately preceding one. In order to evaluate a just noticeable difference (JND) for the sense of balance we first identified RSP intervals where participants did not perceive differences, that is to say participants gave 50% of correct answers to RSP values falling within this interval (centered in 0). The range amplitude has been calculated iteratively, so that the percentage of correct answers of at least one of the two adjacent intervals to the one centered in 0 was statistically different from the chance level observed for the 0 interval. Following this procedure, we determined a range amplitude of 0,2 for RSP.
Paired t-tests were performed for SP in order to assess possible fatigue or practice effects (from the first to the thirteenth trial): no significant difference was found. In Table 1 are reported mean and standard deviation (SD) values for SP parameter in each tested condition (EO and EC) grouped by sequence order.
EO
EC
Furthermore a series of χ2 tests were performed in order to determine the JND. This has been calculated iteratively, for all trials changing the range interval until the percentage of correct answers of at least one of the two adjacent intervals to the one centered in 0 was significant. Following this procedure, we determined a JND of 0.2 for RSP (χ2 = 6.57 p<0.01) when participants improved their performance and a JND of 0.4 when participants’ performance worsened (χ2 = 5.69 p<0.05). In Figure 1 the overall distribution of balance self evaluation is reported.
Considering the responses given by participants in the same subsequent condition (EO-EO,EC-EC and EO-EC together) we found that : EO - EO χ2 = 7.56 p<0.01; EC - EC χ2 =2.42 ns; EO - EC χ2 = 6.38 p<0.05. The use of WBB and the algorithm used in this study permit to merge objective stabilometric data with self evaluation judgements, giving a measure of equilibrioception. A unimodal function has been obtained by RSP self evaluation distribution, demonstrating that RSP describes participants’ ability to perceive changings in their balance performance. This ability is not completely symmetric: JND is smaller for balance improvement and bigger for worsening. This “need” of a bigger difference for the worsening to be perceived seems to suggest a tendency towards overestimation of one’s own balance. Interestingly, participants’ judgement was more reliable when evaluating consecutive EC rather than EO trials, at least when performance was worsening.