- Email: [email protected]
Impaired perceived timing of falls in the elderly
Gait & Posture 59 (2018) 40–45
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Full length article
Impaired perceived timing of falls in the elderly Julian Lupo, Michael Barnett-Cowan
⁎
MARK
Department of Kinesiology, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada
A R T I C L E I N F O
A B S T R A C T
Keywords: Aging Falls Postural perturbation Multisensory integration Temporal order Vestibular
Falls are the leading cause of injury-related deaths and hospitalizations, with older adults at an increased risk. As humans age, physical changes and health conditions make falls more likely. While we know how the body reflexively responds to prevent injury during a fall, we know little about how people perceive the fall itself. We previously found that young adults required a fall to precede a comparison sound stimulus by approximately 44 ms to perceive the two events as simultaneous. This may relate to common anecdotal reports suggesting that humans often describe distortions in their perception of time − time seems to slow down during a fall – with very little recollection of how and when the fall began. Here we examine whether fall perception changes with age. Young (19–25y) and older (61–72y) healthy adults made temporal order judgments identifying whether the onset of their fall or the onset of a comparison sound came first to measure the point of subjective simultaneity. Results show that fall perception is nearly twice as slow for older adults, where perturbation onset has to precede sound onset by ∼88 ms to appear coincident, compared to younger adults (∼44 ms). We suggest that such agerelated differences in fall perception may relate to increased fall rates in older adults. We conclude that a better understanding of how younger versus older adults perceive falls may identify important factors for innovative fall prevention strategies and rehabilitative training exercises to improve fall awareness.
1. Introduction Optimal perception of the world around us requires the central nervous system (CNS) to use multiple sources of sensory information from different sensory modalities. Doing so is important because sensory redundancy can improve the ability to extract meaningful signal from noise [1]. Integrating information from numerous sensory channels facilitates perception, cognition, and action of the self and of the environment by shaping an individual’s behaviour to perform better in an environment that is constantly changing [1]. One such system that is integral to maintaining a healthy and active lifestyle throughout the lifespan is balance control. To prevent us from falling down, balance control uses multiple sensory systems, including auditory, visual, somatosensory, and vestibular [2], but may be adversely affected by how the CNS integrates information from numerous sensory channels [3,4]. Sensory information that is redundant in time provides critical information to the CNS with which it can determine what stimuli should be bound together. Synchrony often characterizes whether stimuli from different modalities should be perceived as belonging together or originating from separate and independent events or objects [5]. However, maintaining a perception of simultaneity can be difficult due to the propagation of different stimulus energies [6]. For example, we see lightning before hearing thunder due to the differences in the physical
⁎
arrival time of the stimuli at the eye and ear [7], making it challenging for our perceptual systems to consistently and accurately perceive simultaneous events. Resultant transmission time for the information to reach the CNS [8], as well as different stimulus attributes characteristic of each event [9] can affect the perceived timing of sensory events. Furthermore, one’s attention can influence the processing speed of incoming stimuli [10], yielding delays between sensory events that do not correspond with the physical timing of phenomena. While not an apparent problem in younger adult populations, studies involving older populations have found that older adults are physically and perceptually slower than younger adults with respect to eye movements [4], response time [3], and temporal order judgments (TOJ) [11]. These results indicate that poor early perceptual processing may lead to less coherent multisensory perception, which may contribute to an increase in fall or sway-like behaviours [2]. Given that falls pose a threat to one's survival, it would seem reasonable to assume that we would perceive the onset of a fall with great accuracy and minimal delay. However, while the CNS is able to rapidly generate compensatory postural reactions, we have recently shown that the perceived onset of a postural perturbation is slow [12]. Here we found that young adults require the onset of a fall to occur 44 ms prior to the onset of a sound in order to perceive the two stimuli as simultaneous. Interestingly this may relate to other reports of vestibular
Corresponding author. E-mail addresses: [email protected] (J. Lupo), [email protected] (M. Barnett-Cowan).
http://dx.doi.org/10.1016/j.gaitpost.2017.09.037 Received 11 April 2017; Received in revised form 26 September 2017; Accepted 27 September 2017 0966-6362/ © 2017 Elsevier B.V. All rights reserved.
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
stimulation being perceptually slow [13–18], however, arguably a fall involves far more stimulation than vestibular input alone. It would seem then that during a fall the CNS may prioritize physiological responses such as postural reflexes over perceptual awareness by relying on other sensory modalities to confirm sensory onset, and thus delaying perception. As such, it may be far less important to contemplate the onset of a perturbation than to react and regain stability from an unexpected event. Despite physiological and non-physiological factors involved in modulating perception (e.g., stimulus intensity) and action, the CNS is remarkably good at being able to reconstruct the actual timing of a multisensory event [19]. It has been suggested that a neural mechanism called ‘simultaneity constancy’ resynchronizes incoming asynchronous multisensory signals representing an event by combining signals from different senses that have varying processing times [20,21]. Such a mechanism is at least partially responsible for the ability to accurately perceive simultaneity of vestibular events [13] despite sensory and perceptual delays. Recent work has found that relative to younger adults, older adults tend to have larger “temporal binding windows” [4,22,23] within which sensory stimuli are judged as occurring simultaneously. This allows for relevant and irrelevant stimuli to be processed and also reduces the probability of successfully integrating multiple stimuli [3]. Setti and colleagues [23] found that older adults were more susceptible to sound-induced flash illusions at longer stimulus onset asynchronies (SOAs), reflecting inefficient processing of irrelevant stimuli in the CNS. Additionally, when older participants are presented with pairs of visual and vibrotactile stimuli to either hand and asked to make TOJs, they required more time and were less accurate than younger observers to correctly perceive the temporal order of events [22]. These results indicate that poor early perceptual processing may lead to less coherent multisensory perception. Furthermore older adults are unable to ignore multiple ambiguous sensory cues as they fall or sway significantly more under sensory conflict situations, which may lead to a loss of balance and posture [2]. While some researchers claim that multisensory integration degrades as a function of age, with increased response times and a wider distribution or range of response times for example [4,24 for a review], others have not found this to be the case. Laurienti and colleagues [3] found that as an individual gets older, their unisensory perception starts to deteriorate and they tend to rely more on integration of the senses. Thus, it is possible that older adults can benefit from having larger temporal binding windows, which may allow them to exploit redundant cues to form a reliable and coherent perception. For example, Spence and colleagues [5] demonstrated that exploiting spatially redundant cues increased the precision with which observers made speeded audiovisual and visuo-tactile TOJs. To date, while there have been many studies that have explored the use of a lean-and-release perturbation system to indirectly elicit inertial cues in older adults, none have looked at the perceived timing of the postural perturbation. In the present study, we tested younger and older healthy adults hypothesizing that following unexpected perturbations, older adults will require the perturbation to occur earlier in time than younger adults to be perceived as simultaneous with a comparison sound stimulus.
Fig. 1. Lean-and-release cable set-up for both younger and older adults. Participants were blindfolded and instructed to lean forward such that their body weight was supported by the cable from neutral stance. Perturbation onset time was calculated by a mounted load cell attached to the mechanical lift. Participants wore noise-cancelling headphones (Sennheiser PXE 450; 70 dB) while standing in a standardized foot position (heel centers 0.17 m apart, 14° between the long axes of the feet [26]). For each of the 110 trials (10 practice), participants judged whether the onset of the fall or a comparison sound (500 Hz 250 ms square wave burst; 77 dB) came first.
through a University of Waterloo Research Ethics Committee, which complies with The Code of Ethics of the World Medical Association (Declaration of Helsinki). 2.2. Protocol Using a lean-and-release perturbation system [12,25,26] participants judged the temporal order of a fall and sound to determine the point of subjective simultaneity (PSS). Participants indicated whether they thought the fall or sound came first by button press. Participants were also required to cross their arms with each hand secured to the opposite shoulder. A restrictive Velcro strap was used to secure their arms from moving throughout the experiment (Fig. 1). Fig. 2 represents a schematic of the onset of the stimuli presented on each trial. We refer the reader to our previous study [12] for more details on the protocol and stimulus generation. 2.3. Data analysis TOJ data acquired at various SOAs are plotted as a function of the percentage of trials in which either response was chosen. Here, negative SOAs represent that a fall occurred prior to the auditory stimulus. A two-parameter logistic function (Eq. (1)) was fit to the TOJ data using Sigma Plot 12.0 (Fig. 3), where x0 refers to the PSS and b refers to the slope of the logistic curve that is proportional to the just noticeable difference (JND). The JND is indicative of the smallest change in interval that observers can reliably notice, thus it is a reflection of the participant’s precision in their decision-making.
2. Methods 2.1. Participants Twelve healthy younger adults (19–25y; M = 22.0y, SD = 1.71y; 7 female) and eleven healthy older adults (61–72y; M = 67.0y, SD = 4.38y; 9 female), free of musculoskeletal, auditory, visual, vestibular, or other neurological disorders, were recruited to participate in this study. Participants gave their informed written consent to participate in the study and all older adults were recruited through the Waterloo Research in Aging Participant Pool. The study was approved
(1) y =
100 1 + e−
(x − x 0) b
%
Recorded TOJs were analyzed for correct trials only. Trials were deemed as an error if there was a response prior to the end of the trial, or if no response was given. Trial errors occurred on fewer than 3% of all trials and these trials were not repeated later in the experiment. 41
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Fig. 2. Trial design schematic. The trial begins with the onset of a go signal (light grey) followed by the perturbation or sound. The onset of the sound (dark grey) and perturbation occurs anywhere from 0 to 650 ms after the offset of the go signal.
Fig. 3. Data from one participant’s raw TOJ data (black circles: 0s: “fall first” and 1s: “sound first”) acquired at different SOAs where negative SOAs represent when the fall occurred before sound onset. The grey line is the logistic function fit to the data, where the PSS (solid vertical line) is the point at which the function crosses 0.5 and participants cannot distinguish between the fall or sound as coming first (i.e., simultaneity). The dashed vertical line signifies the point of true simultaneity (SOA = 0 ms). Fall first and sound first on the x-axis are represented by two symbols (falling person and a speaker icon, respectively).
different from true simultaneity (one-sample t-test, t(11) = −3.501, p = 0.002; Fig. 5A). The average older adult PSS (-88.46 ms, s.e. 18.64) was also significantly different than true simultaneity (Wilcoxon signed ranks, Z = −2.93, p < 0.001; Fig. 5A). The average PSS of the older adults was significantly more delayed than that of younger adults (Mann-Whitney rank sum test, U = 32, p = 0.039; Fig. 5A). However, no significant difference of the JND was found between the younger (52.54 ms, s.e. 8.36) and older adults (57.73 ms, s.e. 12.40) (independent samples t-test, t(21) = 0.35, p = ns; Fig. 5B). Note that Fig. 4B clearly shows that two of the older adults tested responded with particularly negative PSSs and wider JNDs. While no JND values were identified as outliers, one PSS value (-239.6 ms) was a statistical outlier. As Bayesian estimation for two groups can handle outliers by describing the data as heavy tailed distributions instead of normal distributions [27], we performed a Bayesian independent t-test on the PSS values between younger and older adults using JASP v0.8.0.1. Here Bayes Factors (BF) provide a numerical value that quantifies how well a hypothesis (H1; older PSS significantly different from younger PSS) predicts the data relative to a competing null hypothesis (H0; no difference in PSS values across age groups), where a BF10 between 0 and 1, indicates support for the H0, and a BF10 greater than 1 indicates support for the H1. Our results show support for the alternative hypothesis (H1) that despite an outlier participant being included, the average PSS of the older adults was significantly more delayed than that of younger adults (BF10 = 1.502; default Cauchy prior width = 0.707).
2.4. Statistical analysis
3.1. SOA distribution
To assess the hypothesis that perceived simultaneity of the inertial and auditory stimuli will exhibit a similar perceptual lag during a sudden perturbation as found in previous multisensory processing experiments [12–15], a one-sample t-test was conducted comparing the average PSS value of the lean-and-release against true simultaneity (0 ms) for both the young and older adult groups. An independent t-test was conducted to compare the average PSS values of young and older adults. To assess differences in the certainty with which young and older adults made their judgements, another independent t-test between JND values of young and older adults was conducted. In instances where normality failed, Mann-Whitney U and Wilcoxon signed ranks tests were performed. A significance level of α = 0.05 was used for statistical analysis. In order to determine if the distribution of SOAs were influencing the resultant PSS, correlational analyses were performed between the PSS, JND and trial distribution descriptive statistics.
Table 1 lists descriptive statistics of the SOA distribution for both the younger and older adults. Note that because the timing of the fall was controlled by an experimenter, the distribution of SOAs is not tightly controlled within and between participants. Indeed, on average, the perturbation occurred first 66% of the time. To determine whether the distribution of SOAs could explain delayed fall perception, Spearman correlations were analyzed with respect to the PSS, JND and SOA distribution descriptive statistics (see Table 1). These correlational analyses found that there is no significant statistical relationship between the resultant PSS and the high percentage of fall-first trials. 4. Discussion Here we provide further evidence that the perceived timing of a fall is slow relative to a comparison auditory stimulus. Indeed, results from the younger adults closely replicated those from our previous study [12] with a different group of participants, where the perturbation has to precede an auditory stimulus by ∼44 ms in order for the stimulus pair to be perceived as simultaneous. Our results also clearly demonstrate that older adults require a fall to occur twice as early before an auditory stimulus (∼88 ms) than was required for younger adults. Importantly, this increased perceptual delay is not attributable to older
3. Results Fig. 4A and 4 B show the logistic fits for all participants as well as an average logistic curve derived from the average PSS and JND values. The average younger adult PSS (-44.12 ms, s.e. 12.60) was significantly 42
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Fig. 4. Perceived timing of unpredictable perturbations. Positive and negative SOA values represent which stimulus was perceived to be first. Negative SOAs represent the fall occurred first (load cell), where positive SOAs indicate the sound occurred first, as illustrated by the cartoons. The grey lines represent the individual participant data fits. A) The black curve represents the average data fit of all younger adult participants. The solid black vertical line indicates the average PSS value. B) The dark grey curve represents the average data fit of all older adult participants. The solid dark grey vertical line indicates the average PSS value. The dashed vertical line signifies the point of true simultaneity (SOA = 0 ms).
Fig. 5. PSS and JND results. (A) Average younger adult (black) and older adult (dark grey) PSS plotted relative to SOA for auditory stimulus. (B) Median younger adult (black) and older adult (dark grey) JND data for fall-sound pair. Error bars are ± 1 SEM. * p = 0.05, ** p = 0.001. Double asterisks also refer to one way t-tests relative to an SOA of 0 ms, represented by black arrows.
profound attentional bias towards the threatening stimulus. Their research suggests that through FOF adoption strategies, participants compromise their balance by neglecting important sensory information necessary to execute dynamic movements (e.g., a step). Given that older adults are generally more susceptible to falling than their younger counterparts, this may affect their outlook on the experiment which may bias the resultant PSS. Unfortunately, no FOF questionnaire was administered, making it difficult to determine whether anxiety of the fall influenced the results. However, recent research on prior entry has shown evidence for modulation of the PSS due to directing attentional resources towards a particular sensory modality with the resultant PSS being closer to true simultaneity (i.e., 0 ms; [10]). This suggests that if older participants were focusing primarily on the perturbation, a smaller PSS than the one found would be present. We suggest that future work on the perceived timing of falls should explicitly manipulate attention to assess this hypothesis. While the present paper and our previous work [12] present a relatively novel exploration of a minimally explored research area concerning fall perception, the interpretation of results and utility of the approach could benefit from additional insight provided using other established methods of predicting balance control. One such approach comes from systems control theory. Here, it has been proposed that in order for the CNS to trigger a compensatory response to a fall, a loss of balance is required [31,32]. By modeling balance control as a mechanical system, and defining a loss of balance as a loss of effective control of balance that can be detected from internal and external
adults being less precise (i.e., wider JNDs) than younger adults when judging the perceived timing of a fall. While previous work has assessed change in the PSS and JND for multisensory tasks between young and older adults [11], the difference in the perceived timing of fall onset in older adults provides an important insight into potential reasons why older adults are more likely to fall than younger adults. Given these findings, it would be interesting to see whether this misattribution of incoming stimuli is also further expressed in those who exhibit a history of fall behaviour. It is possible that this greater perceptual latency might be a bi-product of degrading vestibular or somatosensory axons as people age [28]. Strupp and colleagues [29] reported an increase in age-related proprioceptive sensory weighting to aid in maintaining balance and posture, suggesting that other sensory systems will compensate for those that are deteriorating. While vestibular, hearing, and intellectual deficits were generally controlled for through the selection process of the University WRAP pool, no additional vestibular, hearing or intellectual function tests were administered prior to beginning the experiment. Thus, future research should look to administer full clinical evaluations of auditory, vestibular, somatosensory and inhibitory executive intellectual function in order to assess the possible role of sensory and cognitive decline in this task. It may also be the case that the anxiety of the perturbation may have inadvertently focused the participant’s attention towards the fall rather than focusing on which stimulus came first. Young and Williams [30] suggest that those who report a higher fear-of-falling (FOF) have less balance confidence and a higher increase in autonomic activity with a 43
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Table 1 (continued)
Table 1 Descriptive SOA distribution statistics (left column) with corresponding Spearman correlations with the PSS (middle column) and JND (right column) for both young and older adults. YOUNGER ADULTS
PSS
JND
Mean (M = −132.88 ms, SD = 21.19 ms)
r = −0.06 p = 0.852 r = 0.23 p = 0.456 r = 0.18 p = 0.572 r = 0.18 p = 0.572 r = −0.01 p = 0.974 r = −0.13 p = 0.683 r = −0.45 p = 0.136 r = −0.18 p = 0.572 r = −0.22 p = 0.484 r = 0.11 p = 0.733 r = 0.01 p = 0.956 r = −0.22 p = 0.484 r = 0.07 p = 0.817 r = −0.18 p = 0.575 r = −0.19 p = 0.542 r = 0.02 p = 0.905 r = −0.10 p = 0.749 r = 0.25 p = 0.429
r = −0.06 p = 0.852 r = 0.13 p = 0.667 r = 0.14 p = 0.651 r = 0.14 p = 0.651 r = 0.02 p = 0.939 r = −0.11 p = 0.716 r = −0.15 p = 0.635 r = −0.04 p = 0.904 r = −0.20 p = 0.527 r = 0.03 p = 0.921 r = 0.13 p = 0.667 r = −0.15 p = 0.619 r = 0.49 p = 0.10 r = −0.37 p = 0.245 r = −0.56 p = 0.058 r = 0.63 p = 0.096 r = 0.01 p = 0.974 r = 0.22 p = 0.484
Standard deviation (M = 195.98 ms, SD = 11.22 ms) Standard error (M = 19.68 ms, SD = 1.08 ms) C.I. of mean (M = 39.06 ms, SD = 2.13 ms) Range (M = 785.0 ms, SD = 148.16 ms) Max (M = 329.63 ms, SD = 144.2 ms) Min (M = −455.38 ms, SD = 17.08 ms) Median (M = −136.38 ms, SD = 26.10 ms) 25% (M = −301.31 ms, SD = 24.39 ms) 75% (M = 36.94 ms, SD = 39.05 ms) Skewness (M = 0.17 ms, SD = 0.18 ms) Kurtosis (M = −0.89 ms, SD = 0.57 ms) K-S distribution (M = 0.09 ms, SD = 0.18 ms) K-S probability (M = 0.10 ms, SD = 0.12 ms) S Wilk (M = 0.95 ms, SD = 0.01 ms) S Wilk probability (M = 0.01 ms, SD = 0.004 ms) Sum (M = −13167.6 ms, SD = 2074.39 ms) Sum of squares (M = 5570005 ms, SD = 457724.7 ms) OLDER ADULTS
PSS
JND
Mean (M = −116.65 ms, SD = 24.67 ms)
r = 0.57 p = 0.06 r = −0.36 p = 0.270 r = −0.41 p = 0.199 r = −0.41 p = 0.199 r = −0.19 p = 0.557 r = −0.02 p = 0.946 r = 0.48 p = 0.124 r = 0.44 p .168=0.168 r = 0.49 p .116=0.116 r = 0.01 p = 0.968 r = −0.08 p = 0.797 r = 0.07 p = 0.818 r = −0.16 p = 0.633 r = 0.14 p = 0.653 r = −0.11 p = 0.734 r = −0.18 p = 0.575
r = 0.15 p = 0.653 r = 0.18 p = 0.575 r = 0.21 p = 0.520 r = 0.21 p = 0.520 r = 0.27 p = 0.40 r = 0.19 p = 0.557 r = −0.14 p = 0.673 r = 0.25 p = 0.450 r = 0.42 p .188=0.188 r = 0.19 p = 0.557 r = −0.26 p = 0.416 r = 0.20 p = 0.538 r = −0.37 p = 0.245 r = 0.37 p = 0.245 r = 0.79 p = 0.002 r = 0.77 p = 0.004
Standard deviation (M = 198.02 ms, SD = 13.71 ms) Standard error (M = 20.16 ms, SD = 1.70 ms) C.I. of mean (M = 40.03 ms, SD = 3.40 ms) Range (M = 913.14 ms, SD = 246.66 ms) Max (M = 398.29 ms, SD = 173.39 ms) Min (M = −514.86 ms, SD = 84.30 ms) Median (M = −103.86 ms, SD = 35.01 ms) 25% (M = −280.75 ms, SD = 31.64 ms) 75% (M = 34.96 ms, SD = 30.22 ms) Skewness (M = 0.07 ms, SD = 0.24 ms) Kurtosis (M = −0.56 ms, SD = 0.71 ms) K-S distribution (M = 0.08 ms, SD = 0.01 ms) K-S probability (M = 0.14 ms, SD = 0.09 ms) S Wilk (M = 0.97 ms, SD = 0.01 ms) S Wilk probability (M = 0.02 ms, SD = 0.03 ms)
OLDER ADULTS
PSS
JND
Sum (M = −11309.7 ms, SD = 2636.69 ms)
r = 0.45 p = 0.159 r = −0.50 p = 0.109
r = 0.24 p = 0.467 r = −0.16 p = 0.614
Sum of squares (M = 5131799 ms, SD = 710903.6 ms)
errors, it has been shown that older adults respond earlier than younger adults to a loss of balance [32]. One possible explanation is that older adults are more cautious than younger adults when maintaining balance, or alternatively increased sensory or motor noise could induce incorrect assumptions about system states and environmental stimuli [32]. While this example − from a seated whole-body balancing task − makes comparisons with the present study difficult, our results would seem to support the former hypothesis. If older adults are better able to detect a loss of balance than younger adults then one would expect the PSS to have been closer to true simultaneity than younger adults, as was found in the present paper. However, if the ratio of signal to noise decreased with age then not only would a delay in the PSS be expected, but also an increase in the JND which was not found here. As this experiment was not explicitly designed to consider fear of falling or caution during balance control, we suggest that future work using our paradigm for measuring fall perception should consider these factors as well as integrate this systems approach in combination with other clinical measures of complex and simple reaction time [33], which have also been shown to predict age-related differential responses to postural perturbations. This work helps inform our understanding of how the CNS temporally processes sensory information during a fall. The observation that older adults require a fall to precede a sound by nearly twice the amount that younger adults require to be perceived as simultaneous represents an important discovery that could guide current and future fall preventative strategies in the older adult population. This research also has the potential to change the way in which researchers study falling behaviour. While this particular method of testing (i.e., fallsound comparison) is not an effective way of training older adults to prevent future fall occurrences in itself, future research can now integrate perceptual tasks such as the one used here to further assess the role of the perceived timing of a fall with balance control and potentially detect and monitor those more likely to fall. As such, this work may guide future research to identify effective methods of reducing falls in not only aging populations, but also those who suffer from balance impairments at any age. Conflict of interest There are no conflicts of interest. Acknowledgments This work was generously supported by an Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (#RGPIN-05435-2014) and a Network in Aging Research Emerging Scholar Grant to MB-C. We thank Adrienne Wise for helping to test participants and Jessy Parokaran Varghese and Aysha Basharat for comments on an earlier version of the manuscript. References [1] M.O. Ernst, H.H. Bülthoff, Merging the senses into a robust percept, Trends Cog. Sci. 8 (4) (2004) 162–169. [2] M. Hu, M. Woollacott, Multisensory training of standing balance in older adults: i. Postural stability and one-leg stance balance, J. Gerontol. 49 (2) (1994) M52–M61. [3] P.J. Laurienti, J.H. Burdette, J.A. Maldjian, M.T. Wallace, Enhanced multisensory integration in older adults, Neurobiol. Aging 27 (8) (2006) 1155–1163. [4] A. Diederich, H. Colonius, A. Schomburg, Assessing age-related multisensory
44
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
[5]
[6] [7] [8] [9]
[10] [11] [12] [13] [14] [15]
[16] [17]
[18]
[19]
Atten. Percept. Psychophys. 72 (4) (2010) 871–884. [20] G. Engel, W. Dougherty, Visual-auditory distance constancy, Nature 234 (1971) 308. [21] A. Kopinska, L.R. Harris, Simultaneity constancy, Perception 33 (9) (2004) 1049–1060. [22] E. Poliakoff, D.I. Shore, C. Lowe, C. Spence, Visuotactile temporal order judgments in ageing, Neurosci. Lett. 396 (3) (2006) 207–211. [23] A. Setti, K.E. Burke, R.A. Kenny, F.N. Newell, Is inefficient multisensory processing associated with falls in older people? Exp. Brain Res. 209 (3) (2011) 375–384. [24] J. Freiherr, J.N. Lundström, U. Habel, K. Reetz, Multisensory integration mechanisms during aging, Front. Hum. Neurosci. (2013) 72013. [25] B. Lakhani, A. Mansfield, E.L. Inness, W.E. McIlroy, Characterizing the determinants of limb preference for compensatory stepping in healthy young adults, Gait Posture 33 (2) (2011) 200–204. [26] W.E. McIlroy, B.E. Maki, Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing, Clin. Biomech. 12 (1) (1997) 66–70. [27] J.K. Kruschke, Bayesian estimation supersedes the t-test, J. Exp. Psychol. Gen. 142 (2) (2013) 573. [28] S.W. Shaffer, A.L. Harrison, Aging of the somatosensory system: a translational perspective, Phys. Ther. 87 (2) (2007) 193–207. [29] M. Strupp, V. Arbusow, C.B. Pereira, M. Dieterich, T. Brandt, Subjective straightahead during neck muscle vibration: effects of ageing, Neuroreport 10 (15) (1999) 3191–3194. [30] W.R. Young, A.M. Williams, How fear of falling can increase fall-risk in older adults: applying psychological theory to practical observations, Gait Posture 41 (1) (2015) 7–12. [31] A.A. Ahmed, J.A. Ashton-Miller, Is a loss of balance a control error signal anomaly? Evidence for three-sigma failure detection in young adults, Gait Posture 193 (2004) 252–262. [32] A.A. Ahmed, J.A. Ashton-Miller, Effect of age on detecting a loss of balance in a seated whole-body balancing task, Clin. Biomech. 20 (2005) 767–775. [33] J.K. Richardson, J.T. Eckner, L. Allet, H. Kim, J.A. Ashton-Miller, Complex and simple clinical reaction times are associated with gait: balance, and major fall injury in older subjects with diabetic peripheral neuropathy, Am. J. Phys. Med. Rehab. 96 (1) (2017) 8–16.
enhancement with the time-window-of-integration model, Neuropsychologia 46 (10) (2008) 2556–2562. C. Spence, R. Baddeley, M. Zampini, R. James, D.I. Shore, Multisensory temporal order judgments: when two locations are better than one, Percept. Psychophys. 65 (2) (2003) 318–328. A. King, A. Palmer, Integration of visual and auditory information in bimodal neurones in the guinea-pig superior colliculus, Exp. Brain Res. 60 (1985) 492–500. C. Spence, S. Squire, Multisensory integration maintaining the perception of synchrony, Curr. Biol. 13 (13) (2003) R519–R521. G.V. Békésy, Interaction of paired sensory stimuli and conduction in peripheral nerves, J. Appl. Physiol. 18 (6) (1963) 1276–1284. L.R. Harris, V. Harrar, P. Jaekl, A. Kopinska, Mechanisms of Simultaneity Constancy: Space and Time in Perception and Action, Oxford University Press, 2010. C. Spence, D.I. Shore, R.M. Klein, Multisensory prior entry, J. Exp. Psychol. Gen. 130 (4) (2001) 799–832. G. Bedard, M. Barnett-Cowan, Impaired timing of audiovisual events in the elderly, Exp. Brain Res. 234 (1) (2016) 331–340. J. Lupo, M. Barnett-Cowan, Perceived timing of a postural perturbation, Neurosci. Lett. 639 (2017) 167–172. M. Barnett-Cowan, L.R. Harris, Perceived timing of vestibular stimulation relative to touch, light and sound, Exp. Brain Res. 198 (2–3) (2009) 221–231. M. Barnett-Cowan, L.R. Harris, Temporal processing of active and passive head movement, Exp. Brain Res. 214 (1) (2011) 27–35. M. Barnett-Cowan, S.M. Raeder, H.H. Bülthoff, Persistent perceptual delay for head movement onset relative to auditory stimuli of different durations and rise times, Exp. Brain Res. 220 (1) (2012) 41–50. M. Barnett-Cowan, Vestibular perception is slow: a review, Multisensory Res 26 (2013) 1–17. M.C. Sanders, N.Y.N. Chang, M.M. Hiss, R.M. Uchanski, T.E. Hullar, Temporal binding of auditory and rotational stimuli, Exp. Brain Res. 210 (3–4) (2011) 539–547. F. Soyka, H.H. Bülthoff, M. Barnett-Cowan, Temporal processing of self-motion: modeling reaction times for rotations and translations, Exp. Brain Res. 228 (1) (2013) 51–62. J. Vroomen, M. Keetels, Perception of intersensory synchrony: a tutorial review,
45
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Full length article
Impaired perceived timing of falls in the elderly Julian Lupo, Michael Barnett-Cowan
⁎
MARK
Department of Kinesiology, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada
A R T I C L E I N F O
A B S T R A C T
Keywords: Aging Falls Postural perturbation Multisensory integration Temporal order Vestibular
Falls are the leading cause of injury-related deaths and hospitalizations, with older adults at an increased risk. As humans age, physical changes and health conditions make falls more likely. While we know how the body reflexively responds to prevent injury during a fall, we know little about how people perceive the fall itself. We previously found that young adults required a fall to precede a comparison sound stimulus by approximately 44 ms to perceive the two events as simultaneous. This may relate to common anecdotal reports suggesting that humans often describe distortions in their perception of time − time seems to slow down during a fall – with very little recollection of how and when the fall began. Here we examine whether fall perception changes with age. Young (19–25y) and older (61–72y) healthy adults made temporal order judgments identifying whether the onset of their fall or the onset of a comparison sound came first to measure the point of subjective simultaneity. Results show that fall perception is nearly twice as slow for older adults, where perturbation onset has to precede sound onset by ∼88 ms to appear coincident, compared to younger adults (∼44 ms). We suggest that such agerelated differences in fall perception may relate to increased fall rates in older adults. We conclude that a better understanding of how younger versus older adults perceive falls may identify important factors for innovative fall prevention strategies and rehabilitative training exercises to improve fall awareness.
1. Introduction Optimal perception of the world around us requires the central nervous system (CNS) to use multiple sources of sensory information from different sensory modalities. Doing so is important because sensory redundancy can improve the ability to extract meaningful signal from noise [1]. Integrating information from numerous sensory channels facilitates perception, cognition, and action of the self and of the environment by shaping an individual’s behaviour to perform better in an environment that is constantly changing [1]. One such system that is integral to maintaining a healthy and active lifestyle throughout the lifespan is balance control. To prevent us from falling down, balance control uses multiple sensory systems, including auditory, visual, somatosensory, and vestibular [2], but may be adversely affected by how the CNS integrates information from numerous sensory channels [3,4]. Sensory information that is redundant in time provides critical information to the CNS with which it can determine what stimuli should be bound together. Synchrony often characterizes whether stimuli from different modalities should be perceived as belonging together or originating from separate and independent events or objects [5]. However, maintaining a perception of simultaneity can be difficult due to the propagation of different stimulus energies [6]. For example, we see lightning before hearing thunder due to the differences in the physical
⁎
arrival time of the stimuli at the eye and ear [7], making it challenging for our perceptual systems to consistently and accurately perceive simultaneous events. Resultant transmission time for the information to reach the CNS [8], as well as different stimulus attributes characteristic of each event [9] can affect the perceived timing of sensory events. Furthermore, one’s attention can influence the processing speed of incoming stimuli [10], yielding delays between sensory events that do not correspond with the physical timing of phenomena. While not an apparent problem in younger adult populations, studies involving older populations have found that older adults are physically and perceptually slower than younger adults with respect to eye movements [4], response time [3], and temporal order judgments (TOJ) [11]. These results indicate that poor early perceptual processing may lead to less coherent multisensory perception, which may contribute to an increase in fall or sway-like behaviours [2]. Given that falls pose a threat to one's survival, it would seem reasonable to assume that we would perceive the onset of a fall with great accuracy and minimal delay. However, while the CNS is able to rapidly generate compensatory postural reactions, we have recently shown that the perceived onset of a postural perturbation is slow [12]. Here we found that young adults require the onset of a fall to occur 44 ms prior to the onset of a sound in order to perceive the two stimuli as simultaneous. Interestingly this may relate to other reports of vestibular
Corresponding author. E-mail addresses: [email protected] (J. Lupo), [email protected] (M. Barnett-Cowan).
http://dx.doi.org/10.1016/j.gaitpost.2017.09.037 Received 11 April 2017; Received in revised form 26 September 2017; Accepted 27 September 2017 0966-6362/ © 2017 Elsevier B.V. All rights reserved.
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
stimulation being perceptually slow [13–18], however, arguably a fall involves far more stimulation than vestibular input alone. It would seem then that during a fall the CNS may prioritize physiological responses such as postural reflexes over perceptual awareness by relying on other sensory modalities to confirm sensory onset, and thus delaying perception. As such, it may be far less important to contemplate the onset of a perturbation than to react and regain stability from an unexpected event. Despite physiological and non-physiological factors involved in modulating perception (e.g., stimulus intensity) and action, the CNS is remarkably good at being able to reconstruct the actual timing of a multisensory event [19]. It has been suggested that a neural mechanism called ‘simultaneity constancy’ resynchronizes incoming asynchronous multisensory signals representing an event by combining signals from different senses that have varying processing times [20,21]. Such a mechanism is at least partially responsible for the ability to accurately perceive simultaneity of vestibular events [13] despite sensory and perceptual delays. Recent work has found that relative to younger adults, older adults tend to have larger “temporal binding windows” [4,22,23] within which sensory stimuli are judged as occurring simultaneously. This allows for relevant and irrelevant stimuli to be processed and also reduces the probability of successfully integrating multiple stimuli [3]. Setti and colleagues [23] found that older adults were more susceptible to sound-induced flash illusions at longer stimulus onset asynchronies (SOAs), reflecting inefficient processing of irrelevant stimuli in the CNS. Additionally, when older participants are presented with pairs of visual and vibrotactile stimuli to either hand and asked to make TOJs, they required more time and were less accurate than younger observers to correctly perceive the temporal order of events [22]. These results indicate that poor early perceptual processing may lead to less coherent multisensory perception. Furthermore older adults are unable to ignore multiple ambiguous sensory cues as they fall or sway significantly more under sensory conflict situations, which may lead to a loss of balance and posture [2]. While some researchers claim that multisensory integration degrades as a function of age, with increased response times and a wider distribution or range of response times for example [4,24 for a review], others have not found this to be the case. Laurienti and colleagues [3] found that as an individual gets older, their unisensory perception starts to deteriorate and they tend to rely more on integration of the senses. Thus, it is possible that older adults can benefit from having larger temporal binding windows, which may allow them to exploit redundant cues to form a reliable and coherent perception. For example, Spence and colleagues [5] demonstrated that exploiting spatially redundant cues increased the precision with which observers made speeded audiovisual and visuo-tactile TOJs. To date, while there have been many studies that have explored the use of a lean-and-release perturbation system to indirectly elicit inertial cues in older adults, none have looked at the perceived timing of the postural perturbation. In the present study, we tested younger and older healthy adults hypothesizing that following unexpected perturbations, older adults will require the perturbation to occur earlier in time than younger adults to be perceived as simultaneous with a comparison sound stimulus.
Fig. 1. Lean-and-release cable set-up for both younger and older adults. Participants were blindfolded and instructed to lean forward such that their body weight was supported by the cable from neutral stance. Perturbation onset time was calculated by a mounted load cell attached to the mechanical lift. Participants wore noise-cancelling headphones (Sennheiser PXE 450; 70 dB) while standing in a standardized foot position (heel centers 0.17 m apart, 14° between the long axes of the feet [26]). For each of the 110 trials (10 practice), participants judged whether the onset of the fall or a comparison sound (500 Hz 250 ms square wave burst; 77 dB) came first.
through a University of Waterloo Research Ethics Committee, which complies with The Code of Ethics of the World Medical Association (Declaration of Helsinki). 2.2. Protocol Using a lean-and-release perturbation system [12,25,26] participants judged the temporal order of a fall and sound to determine the point of subjective simultaneity (PSS). Participants indicated whether they thought the fall or sound came first by button press. Participants were also required to cross their arms with each hand secured to the opposite shoulder. A restrictive Velcro strap was used to secure their arms from moving throughout the experiment (Fig. 1). Fig. 2 represents a schematic of the onset of the stimuli presented on each trial. We refer the reader to our previous study [12] for more details on the protocol and stimulus generation. 2.3. Data analysis TOJ data acquired at various SOAs are plotted as a function of the percentage of trials in which either response was chosen. Here, negative SOAs represent that a fall occurred prior to the auditory stimulus. A two-parameter logistic function (Eq. (1)) was fit to the TOJ data using Sigma Plot 12.0 (Fig. 3), where x0 refers to the PSS and b refers to the slope of the logistic curve that is proportional to the just noticeable difference (JND). The JND is indicative of the smallest change in interval that observers can reliably notice, thus it is a reflection of the participant’s precision in their decision-making.
2. Methods 2.1. Participants Twelve healthy younger adults (19–25y; M = 22.0y, SD = 1.71y; 7 female) and eleven healthy older adults (61–72y; M = 67.0y, SD = 4.38y; 9 female), free of musculoskeletal, auditory, visual, vestibular, or other neurological disorders, were recruited to participate in this study. Participants gave their informed written consent to participate in the study and all older adults were recruited through the Waterloo Research in Aging Participant Pool. The study was approved
(1) y =
100 1 + e−
(x − x 0) b
%
Recorded TOJs were analyzed for correct trials only. Trials were deemed as an error if there was a response prior to the end of the trial, or if no response was given. Trial errors occurred on fewer than 3% of all trials and these trials were not repeated later in the experiment. 41
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Fig. 2. Trial design schematic. The trial begins with the onset of a go signal (light grey) followed by the perturbation or sound. The onset of the sound (dark grey) and perturbation occurs anywhere from 0 to 650 ms after the offset of the go signal.
Fig. 3. Data from one participant’s raw TOJ data (black circles: 0s: “fall first” and 1s: “sound first”) acquired at different SOAs where negative SOAs represent when the fall occurred before sound onset. The grey line is the logistic function fit to the data, where the PSS (solid vertical line) is the point at which the function crosses 0.5 and participants cannot distinguish between the fall or sound as coming first (i.e., simultaneity). The dashed vertical line signifies the point of true simultaneity (SOA = 0 ms). Fall first and sound first on the x-axis are represented by two symbols (falling person and a speaker icon, respectively).
different from true simultaneity (one-sample t-test, t(11) = −3.501, p = 0.002; Fig. 5A). The average older adult PSS (-88.46 ms, s.e. 18.64) was also significantly different than true simultaneity (Wilcoxon signed ranks, Z = −2.93, p < 0.001; Fig. 5A). The average PSS of the older adults was significantly more delayed than that of younger adults (Mann-Whitney rank sum test, U = 32, p = 0.039; Fig. 5A). However, no significant difference of the JND was found between the younger (52.54 ms, s.e. 8.36) and older adults (57.73 ms, s.e. 12.40) (independent samples t-test, t(21) = 0.35, p = ns; Fig. 5B). Note that Fig. 4B clearly shows that two of the older adults tested responded with particularly negative PSSs and wider JNDs. While no JND values were identified as outliers, one PSS value (-239.6 ms) was a statistical outlier. As Bayesian estimation for two groups can handle outliers by describing the data as heavy tailed distributions instead of normal distributions [27], we performed a Bayesian independent t-test on the PSS values between younger and older adults using JASP v0.8.0.1. Here Bayes Factors (BF) provide a numerical value that quantifies how well a hypothesis (H1; older PSS significantly different from younger PSS) predicts the data relative to a competing null hypothesis (H0; no difference in PSS values across age groups), where a BF10 between 0 and 1, indicates support for the H0, and a BF10 greater than 1 indicates support for the H1. Our results show support for the alternative hypothesis (H1) that despite an outlier participant being included, the average PSS of the older adults was significantly more delayed than that of younger adults (BF10 = 1.502; default Cauchy prior width = 0.707).
2.4. Statistical analysis
3.1. SOA distribution
To assess the hypothesis that perceived simultaneity of the inertial and auditory stimuli will exhibit a similar perceptual lag during a sudden perturbation as found in previous multisensory processing experiments [12–15], a one-sample t-test was conducted comparing the average PSS value of the lean-and-release against true simultaneity (0 ms) for both the young and older adult groups. An independent t-test was conducted to compare the average PSS values of young and older adults. To assess differences in the certainty with which young and older adults made their judgements, another independent t-test between JND values of young and older adults was conducted. In instances where normality failed, Mann-Whitney U and Wilcoxon signed ranks tests were performed. A significance level of α = 0.05 was used for statistical analysis. In order to determine if the distribution of SOAs were influencing the resultant PSS, correlational analyses were performed between the PSS, JND and trial distribution descriptive statistics.
Table 1 lists descriptive statistics of the SOA distribution for both the younger and older adults. Note that because the timing of the fall was controlled by an experimenter, the distribution of SOAs is not tightly controlled within and between participants. Indeed, on average, the perturbation occurred first 66% of the time. To determine whether the distribution of SOAs could explain delayed fall perception, Spearman correlations were analyzed with respect to the PSS, JND and SOA distribution descriptive statistics (see Table 1). These correlational analyses found that there is no significant statistical relationship between the resultant PSS and the high percentage of fall-first trials. 4. Discussion Here we provide further evidence that the perceived timing of a fall is slow relative to a comparison auditory stimulus. Indeed, results from the younger adults closely replicated those from our previous study [12] with a different group of participants, where the perturbation has to precede an auditory stimulus by ∼44 ms in order for the stimulus pair to be perceived as simultaneous. Our results also clearly demonstrate that older adults require a fall to occur twice as early before an auditory stimulus (∼88 ms) than was required for younger adults. Importantly, this increased perceptual delay is not attributable to older
3. Results Fig. 4A and 4 B show the logistic fits for all participants as well as an average logistic curve derived from the average PSS and JND values. The average younger adult PSS (-44.12 ms, s.e. 12.60) was significantly 42
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Fig. 4. Perceived timing of unpredictable perturbations. Positive and negative SOA values represent which stimulus was perceived to be first. Negative SOAs represent the fall occurred first (load cell), where positive SOAs indicate the sound occurred first, as illustrated by the cartoons. The grey lines represent the individual participant data fits. A) The black curve represents the average data fit of all younger adult participants. The solid black vertical line indicates the average PSS value. B) The dark grey curve represents the average data fit of all older adult participants. The solid dark grey vertical line indicates the average PSS value. The dashed vertical line signifies the point of true simultaneity (SOA = 0 ms).
Fig. 5. PSS and JND results. (A) Average younger adult (black) and older adult (dark grey) PSS plotted relative to SOA for auditory stimulus. (B) Median younger adult (black) and older adult (dark grey) JND data for fall-sound pair. Error bars are ± 1 SEM. * p = 0.05, ** p = 0.001. Double asterisks also refer to one way t-tests relative to an SOA of 0 ms, represented by black arrows.
profound attentional bias towards the threatening stimulus. Their research suggests that through FOF adoption strategies, participants compromise their balance by neglecting important sensory information necessary to execute dynamic movements (e.g., a step). Given that older adults are generally more susceptible to falling than their younger counterparts, this may affect their outlook on the experiment which may bias the resultant PSS. Unfortunately, no FOF questionnaire was administered, making it difficult to determine whether anxiety of the fall influenced the results. However, recent research on prior entry has shown evidence for modulation of the PSS due to directing attentional resources towards a particular sensory modality with the resultant PSS being closer to true simultaneity (i.e., 0 ms; [10]). This suggests that if older participants were focusing primarily on the perturbation, a smaller PSS than the one found would be present. We suggest that future work on the perceived timing of falls should explicitly manipulate attention to assess this hypothesis. While the present paper and our previous work [12] present a relatively novel exploration of a minimally explored research area concerning fall perception, the interpretation of results and utility of the approach could benefit from additional insight provided using other established methods of predicting balance control. One such approach comes from systems control theory. Here, it has been proposed that in order for the CNS to trigger a compensatory response to a fall, a loss of balance is required [31,32]. By modeling balance control as a mechanical system, and defining a loss of balance as a loss of effective control of balance that can be detected from internal and external
adults being less precise (i.e., wider JNDs) than younger adults when judging the perceived timing of a fall. While previous work has assessed change in the PSS and JND for multisensory tasks between young and older adults [11], the difference in the perceived timing of fall onset in older adults provides an important insight into potential reasons why older adults are more likely to fall than younger adults. Given these findings, it would be interesting to see whether this misattribution of incoming stimuli is also further expressed in those who exhibit a history of fall behaviour. It is possible that this greater perceptual latency might be a bi-product of degrading vestibular or somatosensory axons as people age [28]. Strupp and colleagues [29] reported an increase in age-related proprioceptive sensory weighting to aid in maintaining balance and posture, suggesting that other sensory systems will compensate for those that are deteriorating. While vestibular, hearing, and intellectual deficits were generally controlled for through the selection process of the University WRAP pool, no additional vestibular, hearing or intellectual function tests were administered prior to beginning the experiment. Thus, future research should look to administer full clinical evaluations of auditory, vestibular, somatosensory and inhibitory executive intellectual function in order to assess the possible role of sensory and cognitive decline in this task. It may also be the case that the anxiety of the perturbation may have inadvertently focused the participant’s attention towards the fall rather than focusing on which stimulus came first. Young and Williams [30] suggest that those who report a higher fear-of-falling (FOF) have less balance confidence and a higher increase in autonomic activity with a 43
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
Table 1 (continued)
Table 1 Descriptive SOA distribution statistics (left column) with corresponding Spearman correlations with the PSS (middle column) and JND (right column) for both young and older adults. YOUNGER ADULTS
PSS
JND
Mean (M = −132.88 ms, SD = 21.19 ms)
r = −0.06 p = 0.852 r = 0.23 p = 0.456 r = 0.18 p = 0.572 r = 0.18 p = 0.572 r = −0.01 p = 0.974 r = −0.13 p = 0.683 r = −0.45 p = 0.136 r = −0.18 p = 0.572 r = −0.22 p = 0.484 r = 0.11 p = 0.733 r = 0.01 p = 0.956 r = −0.22 p = 0.484 r = 0.07 p = 0.817 r = −0.18 p = 0.575 r = −0.19 p = 0.542 r = 0.02 p = 0.905 r = −0.10 p = 0.749 r = 0.25 p = 0.429
r = −0.06 p = 0.852 r = 0.13 p = 0.667 r = 0.14 p = 0.651 r = 0.14 p = 0.651 r = 0.02 p = 0.939 r = −0.11 p = 0.716 r = −0.15 p = 0.635 r = −0.04 p = 0.904 r = −0.20 p = 0.527 r = 0.03 p = 0.921 r = 0.13 p = 0.667 r = −0.15 p = 0.619 r = 0.49 p = 0.10 r = −0.37 p = 0.245 r = −0.56 p = 0.058 r = 0.63 p = 0.096 r = 0.01 p = 0.974 r = 0.22 p = 0.484
Standard deviation (M = 195.98 ms, SD = 11.22 ms) Standard error (M = 19.68 ms, SD = 1.08 ms) C.I. of mean (M = 39.06 ms, SD = 2.13 ms) Range (M = 785.0 ms, SD = 148.16 ms) Max (M = 329.63 ms, SD = 144.2 ms) Min (M = −455.38 ms, SD = 17.08 ms) Median (M = −136.38 ms, SD = 26.10 ms) 25% (M = −301.31 ms, SD = 24.39 ms) 75% (M = 36.94 ms, SD = 39.05 ms) Skewness (M = 0.17 ms, SD = 0.18 ms) Kurtosis (M = −0.89 ms, SD = 0.57 ms) K-S distribution (M = 0.09 ms, SD = 0.18 ms) K-S probability (M = 0.10 ms, SD = 0.12 ms) S Wilk (M = 0.95 ms, SD = 0.01 ms) S Wilk probability (M = 0.01 ms, SD = 0.004 ms) Sum (M = −13167.6 ms, SD = 2074.39 ms) Sum of squares (M = 5570005 ms, SD = 457724.7 ms) OLDER ADULTS
PSS
JND
Mean (M = −116.65 ms, SD = 24.67 ms)
r = 0.57 p = 0.06 r = −0.36 p = 0.270 r = −0.41 p = 0.199 r = −0.41 p = 0.199 r = −0.19 p = 0.557 r = −0.02 p = 0.946 r = 0.48 p = 0.124 r = 0.44 p .168=0.168 r = 0.49 p .116=0.116 r = 0.01 p = 0.968 r = −0.08 p = 0.797 r = 0.07 p = 0.818 r = −0.16 p = 0.633 r = 0.14 p = 0.653 r = −0.11 p = 0.734 r = −0.18 p = 0.575
r = 0.15 p = 0.653 r = 0.18 p = 0.575 r = 0.21 p = 0.520 r = 0.21 p = 0.520 r = 0.27 p = 0.40 r = 0.19 p = 0.557 r = −0.14 p = 0.673 r = 0.25 p = 0.450 r = 0.42 p .188=0.188 r = 0.19 p = 0.557 r = −0.26 p = 0.416 r = 0.20 p = 0.538 r = −0.37 p = 0.245 r = 0.37 p = 0.245 r = 0.79 p = 0.002 r = 0.77 p = 0.004
Standard deviation (M = 198.02 ms, SD = 13.71 ms) Standard error (M = 20.16 ms, SD = 1.70 ms) C.I. of mean (M = 40.03 ms, SD = 3.40 ms) Range (M = 913.14 ms, SD = 246.66 ms) Max (M = 398.29 ms, SD = 173.39 ms) Min (M = −514.86 ms, SD = 84.30 ms) Median (M = −103.86 ms, SD = 35.01 ms) 25% (M = −280.75 ms, SD = 31.64 ms) 75% (M = 34.96 ms, SD = 30.22 ms) Skewness (M = 0.07 ms, SD = 0.24 ms) Kurtosis (M = −0.56 ms, SD = 0.71 ms) K-S distribution (M = 0.08 ms, SD = 0.01 ms) K-S probability (M = 0.14 ms, SD = 0.09 ms) S Wilk (M = 0.97 ms, SD = 0.01 ms) S Wilk probability (M = 0.02 ms, SD = 0.03 ms)
OLDER ADULTS
PSS
JND
Sum (M = −11309.7 ms, SD = 2636.69 ms)
r = 0.45 p = 0.159 r = −0.50 p = 0.109
r = 0.24 p = 0.467 r = −0.16 p = 0.614
Sum of squares (M = 5131799 ms, SD = 710903.6 ms)
errors, it has been shown that older adults respond earlier than younger adults to a loss of balance [32]. One possible explanation is that older adults are more cautious than younger adults when maintaining balance, or alternatively increased sensory or motor noise could induce incorrect assumptions about system states and environmental stimuli [32]. While this example − from a seated whole-body balancing task − makes comparisons with the present study difficult, our results would seem to support the former hypothesis. If older adults are better able to detect a loss of balance than younger adults then one would expect the PSS to have been closer to true simultaneity than younger adults, as was found in the present paper. However, if the ratio of signal to noise decreased with age then not only would a delay in the PSS be expected, but also an increase in the JND which was not found here. As this experiment was not explicitly designed to consider fear of falling or caution during balance control, we suggest that future work using our paradigm for measuring fall perception should consider these factors as well as integrate this systems approach in combination with other clinical measures of complex and simple reaction time [33], which have also been shown to predict age-related differential responses to postural perturbations. This work helps inform our understanding of how the CNS temporally processes sensory information during a fall. The observation that older adults require a fall to precede a sound by nearly twice the amount that younger adults require to be perceived as simultaneous represents an important discovery that could guide current and future fall preventative strategies in the older adult population. This research also has the potential to change the way in which researchers study falling behaviour. While this particular method of testing (i.e., fallsound comparison) is not an effective way of training older adults to prevent future fall occurrences in itself, future research can now integrate perceptual tasks such as the one used here to further assess the role of the perceived timing of a fall with balance control and potentially detect and monitor those more likely to fall. As such, this work may guide future research to identify effective methods of reducing falls in not only aging populations, but also those who suffer from balance impairments at any age. Conflict of interest There are no conflicts of interest. Acknowledgments This work was generously supported by an Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (#RGPIN-05435-2014) and a Network in Aging Research Emerging Scholar Grant to MB-C. We thank Adrienne Wise for helping to test participants and Jessy Parokaran Varghese and Aysha Basharat for comments on an earlier version of the manuscript. References [1] M.O. Ernst, H.H. Bülthoff, Merging the senses into a robust percept, Trends Cog. Sci. 8 (4) (2004) 162–169. [2] M. Hu, M. Woollacott, Multisensory training of standing balance in older adults: i. Postural stability and one-leg stance balance, J. Gerontol. 49 (2) (1994) M52–M61. [3] P.J. Laurienti, J.H. Burdette, J.A. Maldjian, M.T. Wallace, Enhanced multisensory integration in older adults, Neurobiol. Aging 27 (8) (2006) 1155–1163. [4] A. Diederich, H. Colonius, A. Schomburg, Assessing age-related multisensory
44
Gait & Posture 59 (2018) 40–45
J. Lupo, M. Barnett-Cowan
[5]
[6] [7] [8] [9]
[10] [11] [12] [13] [14] [15]
[16] [17]
[18]
[19]
Atten. Percept. Psychophys. 72 (4) (2010) 871–884. [20] G. Engel, W. Dougherty, Visual-auditory distance constancy, Nature 234 (1971) 308. [21] A. Kopinska, L.R. Harris, Simultaneity constancy, Perception 33 (9) (2004) 1049–1060. [22] E. Poliakoff, D.I. Shore, C. Lowe, C. Spence, Visuotactile temporal order judgments in ageing, Neurosci. Lett. 396 (3) (2006) 207–211. [23] A. Setti, K.E. Burke, R.A. Kenny, F.N. Newell, Is inefficient multisensory processing associated with falls in older people? Exp. Brain Res. 209 (3) (2011) 375–384. [24] J. Freiherr, J.N. Lundström, U. Habel, K. Reetz, Multisensory integration mechanisms during aging, Front. Hum. Neurosci. (2013) 72013. [25] B. Lakhani, A. Mansfield, E.L. Inness, W.E. McIlroy, Characterizing the determinants of limb preference for compensatory stepping in healthy young adults, Gait Posture 33 (2) (2011) 200–204. [26] W.E. McIlroy, B.E. Maki, Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing, Clin. Biomech. 12 (1) (1997) 66–70. [27] J.K. Kruschke, Bayesian estimation supersedes the t-test, J. Exp. Psychol. Gen. 142 (2) (2013) 573. [28] S.W. Shaffer, A.L. Harrison, Aging of the somatosensory system: a translational perspective, Phys. Ther. 87 (2) (2007) 193–207. [29] M. Strupp, V. Arbusow, C.B. Pereira, M. Dieterich, T. Brandt, Subjective straightahead during neck muscle vibration: effects of ageing, Neuroreport 10 (15) (1999) 3191–3194. [30] W.R. Young, A.M. Williams, How fear of falling can increase fall-risk in older adults: applying psychological theory to practical observations, Gait Posture 41 (1) (2015) 7–12. [31] A.A. Ahmed, J.A. Ashton-Miller, Is a loss of balance a control error signal anomaly? Evidence for three-sigma failure detection in young adults, Gait Posture 193 (2004) 252–262. [32] A.A. Ahmed, J.A. Ashton-Miller, Effect of age on detecting a loss of balance in a seated whole-body balancing task, Clin. Biomech. 20 (2005) 767–775. [33] J.K. Richardson, J.T. Eckner, L. Allet, H. Kim, J.A. Ashton-Miller, Complex and simple clinical reaction times are associated with gait: balance, and major fall injury in older subjects with diabetic peripheral neuropathy, Am. J. Phys. Med. Rehab. 96 (1) (2017) 8–16.
enhancement with the time-window-of-integration model, Neuropsychologia 46 (10) (2008) 2556–2562. C. Spence, R. Baddeley, M. Zampini, R. James, D.I. Shore, Multisensory temporal order judgments: when two locations are better than one, Percept. Psychophys. 65 (2) (2003) 318–328. A. King, A. Palmer, Integration of visual and auditory information in bimodal neurones in the guinea-pig superior colliculus, Exp. Brain Res. 60 (1985) 492–500. C. Spence, S. Squire, Multisensory integration maintaining the perception of synchrony, Curr. Biol. 13 (13) (2003) R519–R521. G.V. Békésy, Interaction of paired sensory stimuli and conduction in peripheral nerves, J. Appl. Physiol. 18 (6) (1963) 1276–1284. L.R. Harris, V. Harrar, P. Jaekl, A. Kopinska, Mechanisms of Simultaneity Constancy: Space and Time in Perception and Action, Oxford University Press, 2010. C. Spence, D.I. Shore, R.M. Klein, Multisensory prior entry, J. Exp. Psychol. Gen. 130 (4) (2001) 799–832. G. Bedard, M. Barnett-Cowan, Impaired timing of audiovisual events in the elderly, Exp. Brain Res. 234 (1) (2016) 331–340. J. Lupo, M. Barnett-Cowan, Perceived timing of a postural perturbation, Neurosci. Lett. 639 (2017) 167–172. M. Barnett-Cowan, L.R. Harris, Perceived timing of vestibular stimulation relative to touch, light and sound, Exp. Brain Res. 198 (2–3) (2009) 221–231. M. Barnett-Cowan, L.R. Harris, Temporal processing of active and passive head movement, Exp. Brain Res. 214 (1) (2011) 27–35. M. Barnett-Cowan, S.M. Raeder, H.H. Bülthoff, Persistent perceptual delay for head movement onset relative to auditory stimuli of different durations and rise times, Exp. Brain Res. 220 (1) (2012) 41–50. M. Barnett-Cowan, Vestibular perception is slow: a review, Multisensory Res 26 (2013) 1–17. M.C. Sanders, N.Y.N. Chang, M.M. Hiss, R.M. Uchanski, T.E. Hullar, Temporal binding of auditory and rotational stimuli, Exp. Brain Res. 210 (3–4) (2011) 539–547. F. Soyka, H.H. Bülthoff, M. Barnett-Cowan, Temporal processing of self-motion: modeling reaction times for rotations and translations, Exp. Brain Res. 228 (1) (2013) 51–62. J. Vroomen, M. Keetels, Perception of intersensory synchrony: a tutorial review,
45