|Home | About | Journals | Submit | Contact Us | Français|
In older adults, maintaining balance and processing information typically interfere with each other, suggesting that executive functions may be engaged for both. We investigated associations between measures of inhibitory processes and standing postural control in healthy young and older adults. Perceptual and motor inhibition was measured using a protocol adapted from Nassauer and Halperin (2003, Dissociation of perceptual and motor inhibition processes through the use of novel computerized conflict tasks. Journal of the International Neuropsychological Society, 9, 25–30). These measures were then correlated to postural sway during standing conditions that required resolving various levels of sensory conflict, for example, world-fixed versus sway-referenced floor and visual scene. In the older adults, perceptual inhibition was positively correlated with sway amplitude on a sway-referenced floor and with a fixed visual scene (r = .68, p < .001). Motor inhibition was not correlated with sway on either group. Perceptual inhibition may be a component of the sensory integration process important for maintaining balance in older adults.
Certain cognitive processes, such as attention, have been shown to influence postural control, particularly in older adults. Interference between balance and cognitive task performance has been demonstrated in dual-task paradigms in which postural challenges are combined with a concurrent information processing task. The decline in performance of either balance or information processing relative to either task performed by itself is taken as evidence that the concurrent tasks compete for particular processing resources. Older adults appear to be particularly affected when balance is examined within dual-task paradigms (Brauer, Woollacott, & Shumway-Cook, 2001; Lundin-Olsson, Nyberg, & Gustafson, 1997; Maylor & Wing, 1996; Pellecchia, 2003; Rankin, Woollacott, Shumway-Cook, & Brown, 2000; Redfern, Jennings, Martin, & Furman, 2001). Importantly, falls in older adults have been associated with postural dual-task performance (Hauer et al., 2003; Lajoie & Gallagher, 2004; Verghese et al., 2002). However, not all studies have shown interference effects in standing postural dual-task paradigms (Andersson, Hagman, Talianzadeh, Svedberg, & Larsen, 2002; Dault, Frank, & Allard, 2001; Swan, Otani, Loubert, Sheffert, & Dunbar, 2004), even in older adults (Maylor, Allison, & Wing, 2001; Swan et al.). Thus, experiments have suggested that age-related changes may influence processes shared by balance and information processing, but the specific mechanisms contributing to this association are not understood.
Age-related decline in postural control and gait has been related to a decline in executive function (Hausdorff et al., 2006; Holtzer, Verghese, Xue, & Lipton, 2006; Redfern et al., 2001; Verghese, Wang, Lipton, Holtzer, & Xue, 2007; Yogev-Seligmann, Hausdorff, & Giladi, 2008). Executive functions, which include working memory, planning, and inhibitory processes (Braver & West, 2008; Hasher & Zacks, 1988), are known to decline with age (Wecker, Kramer, Wisniewski, Delis, & Kaplan, 2000), presumably due to selective deterioration of the dorsolateral prefrontal cortex (Phillips & Della Sala, 1988) and other cortical networks (Andres & Van der Linden, 2000). Working memory has been suggested as one component of executive function that affects postural control. However, experiments using working memory tasks concurrent with a postural task in older adults have found conflicting results, with some showing increased dual-task interference for older adults and others not (Swan et al., 2004; VanderVelde, Woollacott, & Shumway-Cook, 2005). Inhibitory function also has been implicated in aging’s impact on postural control. Redfern et al. employed reaction time (RT) tasks requiring inhibition concurrently with various standing postural tasks. A stop signal RT tasks was employed that was based upon Logan, Schachar, and Tannock (1997). The participant responded as fast as possible to a light emitting diode (the “go” signal) unless a “stop” tone was heard through headphones. The stop–delay time between the go signal and the tone was titrated based on inhibitory success of prior trials. The Go–Stop signal delay was initially set at 80 ms. If the participant successfully inhibited on a trial with this delay time, the delay time was increased by 10 ms. If the participant was unsuccessful in inhibiting, the delay time was decreased by 10 ms. The primary index of inhibition was the stop signal RT (i.e., time required to inhibit a response as derived from a “horse-race model”) (Logan et al.). Using this task design, the stop signal RT is defined as the delay time at which the participant successfully inhibited responses to the stop signals 50% of the time. Performing this task concurrent with standing resulted in an increase in the stop signal RT in healthy young and older participants only when there was a sensory conflict for postural control. Sensory conflict was created by performing sway referencing of the floor and visual scene. (Sway referencing is a method used to reduce proprioceptive cues from the ankles and/or visual cues for standing balance [Nashner, Black, & Wall, 1982]. This then produces a sensory conflict between the proprioceptive system and the visual and vestibular systems.) Based upon this result, Redfern et al. suggested that the sensory integration component of postural control shared a requirement for inhibitory processing with that required in the stop signal reaction task. They did not see a differentially greater effect in the older adults and therefore suggested that, perhaps, age-related changes in inhibitory function are specific to particular types of tasks and not general in nature, as suggested by Kramer, Humphrey, Larish, Logan, and Strayer (1994). Alternatively, older participants may have exercised greater inhibition than younger participants in order to obtain the observed equivalence between groups in stop signal RTs/inhibition times.
Some types of inhibition have been shown to be affected by age (Hasher & Zacks, 1979; Hasher & Zacks, 1988). Various paradigms, such as the Hayling task (Burgess & Shallice, 1996) and Stroop task (Golden, 1978; Spieler, Balota, & Faust, 1996), have shown decrements in inhibition with age. However, inhibition can be thought of as having subtypes that aging can affect differentially (Kramer et al., 1994). In a recent review, Lustig, Hasher, and Zacks (2007) identified three potential types of inhibition influenced by aging: (a) controlling access to attention’s focus, (b) deleting irrelevant information from attention and working memory, and (c) suppressing or restraining strong but inappropriate responses. This review of both behavioral and brain imaging results suggests that these types of inhibition overlap, sharing a core set of processes. Each, however, does have distinctive features that may vary for specific tasks and due to individual differences. These shared and unique processes constituting inhibitory function in different task settings could be differentially affected by aging.
One framework divides inhibition into perceptual and motor inhibition (Nassauer & Halperin, 2003) and introduces methods generally related to early work of Simon and the subsequent development of this task (Craft & Simon, 1970). Perceptual inhibition serves to maintain the focus of attention by reducing interference from stimuli irrelevant to the current primary task. Motor inhibition prevents execution of inappropriate motor responses. Relating to the view of Lustig and colleagues (2007), perceptual inhibition is related to controlling access to attention, whereas motor inhibition is related to restraining strong, but inappropriate, responses. Although it is difficult to separate perceptual and motor inhibitory processes, Nassauer and Halperin developed a RT test battery to differentially measure perceptual and motor inhibition by combining congruent and incongruent visual stimuli and varying degree of perceptual and motor conflict. Their results indicated that perceptual inhibition could be dissociated from motor inhibition with this type of analysis. Recently, we developed a Motor and Perceptual Inhibition Test (MAPIT) based upon the work of Nassauer and Halperin to examine these two types of inhibition in older adults (Jennings, Mendelson, Redfern, & Nebes, in press). Results showed that motor inhibition and perceptual inhibition measures differed between older and young participant groups, with greater differences between the groups in motor inhibition. Each of these measures demonstrated acceptable internal consistency.
The purpose of the current study was to determine if perceptual and/or motor inhibition as assessed by the MAPIT protocol is associated with postural control in healthy young and older adults. If our previous supposition was correct that inhibitory function was called upon by the sensory integration component of balance, then individual differences in inhibition should relate to the efficiency of maintaining balance. Postural control was assessed by measuring sway during different standing conditions, including conditions that lead to different levels of sensory conflict. We hypothesized that perceptual inhibition performance would be correlated with sway responses during conditions requiring sensory conflict resolution, that is, inhibition measures should relate to sway when visual and kinesthetic information conflicts but not when conflict is eliminated by stabilizing the balance platform or reducing visual information. Furthermore, we entertained the possibility that this association would be stronger in older adults compared with young adults due to possible reduced inhibitory capacity in older adults.
Twenty-four young healthy adults (mean age 25.7 years, SD = 3.8 years, range 21–34 years; 13 women and 11 men) and 24 healthy older adults (mean age 74.2 years, SD = 4.4 years, range 70–82 years; 12 women and 12 men) with no history of vestibular or neurological disorders participated in this institutional review board approved study. Informed consent was obtained prior to any participation. Postural data from 2 of the older participants were not included in the analysis due to technical problems with data collection, resulting in 22 older participants being analyzed.
Participants were screened for sensory, cognitive, and musculoskeletal health. Participants were excluded for abnormalities in vestibular function, proprioceptive function, vision, or motor function. Vestibular function was evaluated through caloric and rotational vestibular testing by a neurologist using established clinical criteria (Furman & Cass, 1996). Ankle joint proprioceptive sensation was examined using an ankle joint position sense protocol (Lord, Menz, & Tiedemann, 2003). Plantar foot cutaneous pressure threshold was determined using Semmes–Weinstein monofilaments, with an exclusionary cutoff of 5.07 (Kumar et al., 1991). A neurologist also performed an examination for basic neurological health. Other exclusionary criteria was binocular visual acuity (with corrective lenses) of worse than 20/40 and hearing loss. Cognitive competency was evaluated with a Mini-Mental State Examination, with individuals scoring 23 or less being excluded.
A dynamic posturography platform (Equitest; Neurocom, Inc., Clackamas, OR) was used to conduct postural sway measurements. The platform provides rotations of the support surface and/or the visual scene about the ankle to allow sway referencing of vision and/or proprioception. Sway referencing was accomplished through rotation in direct proportion to the individual’s sway in the anteroposterior (AP) direction (Nashner, Black, & Wall, 1982). Sway referencing uses a low-pass filtered center of pressure (COP) captured from force sensors in the platform to estimate the center of gravity in the AP direction in an attempt to keep the ankle angle constant. This movement of the support surface reduces the reliability of proprioceptive information from the ankle for balance. The same principle is used for sway referencing the visual scene in an attempt to move the scene with the individual during sway. Visual sway referencing is performed to provide erroneously stable vision even though the person is swaying.
The MAPIT inhibitory function test (Jennings et al., in press), adapted from Nassauer and Halperin (2003), was performed. The test measures RTs to visual stimuli on a computer screen. The stimulus conditions use an arrow 3.7 cm long pointing either to the right or to the left. Participant response is a key press from either the right or the left index finger. Two types of RT tasks were presented in order to create a perceptual conflict and a motor conflict. Figure 1 shows a diagram representing the conditions for the perceptual and motor tasks.
For the perceptual tasks, a right- or left-pointing arrow appeared 8.5 cm to the right or left of a central fixation point on the screen. The participant was to press the button on the side toward which the arrow pointed. There were two conditions: congruous and incongruous. In the congruous condition, the spatial location of the arrow (i.e., the side of the screen on which it appeared) was the same as the direction the arrow pointed (e.g., a left-pointing arrow appeared to the left of fixation). In the incongruous condition, the location of the arrow conflicted with the direction it pointed (e.g., a right-pointing arrow appeared to the left side of the fixation point). In the incongruous condition, the participant was required to inhibit processing the arrow’s spatial location, focusing only on the direction it pointed. There were 40 congruous and 40 incongruous trials randomly intermixed. The participant’s mean RT was determined for each condition. Difficulty in inhibiting the processing of the irrelevant spatial information (i.e., arrow location) results in the incongruous condition being typically slower than responses in the congruous condition. As in Nassauer and Halperin (2003), the spatial response tendency was reinforced just prior to the congruous or incongruous block by presenting 40 unscored trials in which the participant responded on the congruous key to a rectangle presented in either the right or the left of the computer screen. In addition, two separate blocks of 40 choice RT baseline trials were presented just prior to and after these blocks. In these blocks, arrows were presented in the center of the screen, and participants responded on the key corresponding to the direction of the arrow.
The motor tasks consisted of two conditions presented in separate blocks of trials. In the first condition, an arrow appeared in the center of the computer screen with right- and left-pointing arrows being randomly intermixed. Participants were asked to press the key on the side toward which the arrow pointed. In the second condition, participants had to press the button on the side “opposite” the direction the arrow pointed (e.g., if the arrow pointed right, they were to press the left button). Thus, in this condition, the participant must inhibit a spatially compatible response in order to make a response that is spatially incompatible with the presented stimulus. Participants performed 40 trials of each condition. Difficulty inhibiting the overlearned response of pressing the button on the side toward which the arrow pointed results in slower responses.
The inhibitory test protocol was conducted on a day prior to that the postural conditions were performed. During the day postural sway was collected, six postural conditions were included by providing changes in the support surface (i.e., floor) and visual scene. The three visual conditions were (a) eyes open in the light (EO), (b) eyes open in the dark (DARK), and (c) a sway-referenced visual scene (VSR). The platform conditions were (a) fixed support surface (FIXED) and (b) sway-referenced floor (SRF). These six postural conditions provide different levels of sensory information for maintaining balance. (Note that the conditions are similar to those in clinical dynamic posturography protocols, except that our participants were in the dark instead of just performing with eyes closed.)
Each of the six visual–floor combinations was conducted three times, with each trial lasting for 180 s. Postural sway was recorded as the change in the COP under the feet in the AP direction at a sampling rate of 100 Hz. All trials were randomly distributed.
The measure of postural sway calculated from the COP recordings was the root mean square (RMS).
where N is the total number of points in the series for a trial. The RMS estimates the overall amount of movement of the COP (Prieto, Myklebust, Hoffmann, Lovett, & Myklebust, 1996).
Median RTs for each condition within participant were computed. The data from any participant performing at less than 75% accuracy were excluded from the analyses. These criteria resulted in data from four older participants not being used in calculating the motor inhibition scores. The perceptual inhibition score was calculated as the difference between the perceptual incongruous RT and the perceptual congruous RT. The incongruous and congruous components of the perceptual inhibition score were also investigated by subtracting the baseline choice RTs from each component; the median of perceptually congruous RTs was less than that of the baseline RTs and the median of the perceptually incongruous RTs was less than that of the baseline RTs. Motor inhibition scores were computed as the difference between the motor incongruous RT and the motor congruous RT. In a larger sample of participants participating in balance-related studies, the reliability (Cronbach’s alpha) of these difference scores was shown to be r = .70 or greater except for an r = .65 for the perceptual incongruent minus perceptual congruent score (Jennings et al., in press). The greater the inhibition score, the greater the slowing associated with the incongruous displays. Thus, the interpretation is that greater inhibition scores reflect increased difficulty in inhibition.
An analysis of variance (ANOVA) approach was used to compare mean values throughout, typically with a between-group comparison for age and a repeated measure factor for task or score. Pearson product moment correlations were used to relate individual differences in our inhibition and balance measures. Statistical analyses were implemented within Statistica (StatSoft, Tulsa, OK) using a general linear model approach for mixed model analyses of variance. A significance level of α = .05 was used throughout the analyses. When data were unavailable for a participant, such as when the participant took a step off the force plate, data for that particular trial were eliminated.
The RTs collected during the inhibitory test protocol were analyzed using ANOVA to determine the characteristics for the two groups. A between-groups ANOVA of the baseline RTs demonstrated a significant difference between groups (F(1, 43) = 40.4, p < .001); older adult RTs were 431 (SD = 53) ms, and younger adult RTs were 344 (SD = 37) ms. A similar analysis found the inhibition measures longer in older participants compared with the young for both perceptual inhibition (F(1, 43) = 6.35, p = .02) and motor inhibition (F(1, 43) = 24.1, p < .001; Figure 2). The motor and perceptual inhibition measures were positively correlated for the older participants (r = .45, p = .02) but not for the young participants (r = .02, p = .67). There was not, however, a significant difference between these correlations for the young and older participants (p = .14).
Sway RMS varied across conditions and between groups (Table 1; for further detail on these balance results, see Redfern, Muller, & Jennings, in press). An ANOVA with a between participant effect of Group and repeated measures of Floor and Scene was applied to the RMS measure. RMS was found to be influenced by the main effects of Group (F(1, 44) = 6.83, p = .01), Floor (F(1, 44) = 135.6, p < .0001), Scene (F(2, 88) = 52.4, p < .0001), Group × Scene (F(2, 88) = 6.45, p = 0.002), and Floor × Scene (F(2, 88) = 33.4, p < .0001). As expected, the older participants had greater sway than young participants and sway referencing the floor increased postural sway. Reducing correct visual information (i.e., DARK or VSR) increased sway. This effect was stronger in the older adults. Vision (eyes open vs. in the dark) has a particularly strong effect during the SRF condition.
Correlations between the postural sway and the inhibitory function measures were examined within each of the six postural conditions. This was done for young and older adults separately. For the young participants, there were no significant correlations between postural sway RMS and either motor inhibition or perceptual inhibition. For the older adults, there were no significant correlations between either motor inhibition or perceptual inhibition and sway RMS for the three conditions with a fixed floor (FIXED–EO, FIXED–DARK, and FIXED–VSR).
Sway correlations with perceptual inhibition emerged among the older adults in the sway-referenced floor conditions and did so whether or not sensorimotor speed was partialled out. Table 2 presents the correlations by condition and group with the mean sensorimotor speed derived from the participant’s simple RT partialled out. Perceptual inhibition correlated with sway RMS for the postural condition with a sway-referenced floor and eyes open (SRF–EO, r = .67, p < .001; correlation significantly different from r = −.13 in the young, p < .05). As the perceptual inhibition measure increased (i.e., their performance on the perceptual inhibition task worsened), sway RMS increased (Figure 3a). A significant, but weaker, correlation occurred between RMS and perceptual inhibition for postural condition with a sway-referenced floor and sway-referenced visual scene (SRF–VSR; r = .44, p = .046). There were no significant correlations between the motor inhibition measure and RMS for any of the conditions; however, the correlation between motor inhibition and RMS did approach significance for the SRF–EO condition (r = .39, p = .07; Figure 3b).
To further explore the finding that the perceptual inhibition measure correlates with sway during the SRF–EO condition, we analyzed the congruous and incongruous components of the measure. This was performed to determine if one component (congruous response or incongruous response) was having a dominant impact on the perceptual inhibition measures that were correlated with sway. In order to accomplish this analysis, the baseline RT responses to arrows presented in the center of the screen were subtracted from the congruous and incongruous RTs to normalize individual RTs (Figure 4). A general linear model was applied to the congruous RT differences and incongruous RT differences separately. A mixed model approach was used with the categorical age variable (young vs. older groups), the appropriate RT difference score for congruous or incongruous trials, and the interaction of RT with age group. For the incongruous RT analysis, RT difference score (F(1, 40) = 7.2, p = .01) and the interaction of age group and RT score (F(1, 40) = 5.4, p = .025) were significant. For the congruous RT analysis, RT difference score was marginally significant (F(1, 40) = 3.5, p = .07) and the interaction of age group and RT score (F(1, 40) = 8.6, p = .006) was significant. Figure 4 illustrates the significant interactions using a scatter diagram to show the data. The differences in the slopes of the fitted lines illustrate the significant interactions as well as the enhanced age-group difference for congruous relative to incongruous responses.
The main finding in this study was that perceptual inhibition was associated with postural sway in older adults under well-defined conditions. Specifically, sway was correlated with perceptual inhibition for the sway-referenced floor with eyes open condition (SRF–EO) in the older group. There was also a significant, but reduced, correlation between sway and perceptual inhibition for the sway-referenced floor and scene (SRF–VSR) condition. Motor inhibition was not found to be significantly correlated with sway under any conditions for either age-group.
The results suggest that perceptual inhibition is particularly involved in the sensory integration processes for postural condition in older adults. Sway referencing the floor makes proprioceptive information from the ankles unreliable, and therefore vision and vestibular inputs must be used for postural control. Thus, a “sensory reweighting” must increase reliance on vision and/or vestibular information while decreasing the influence of proprioception (Peterka, 2002; Peterka & Loughlin, 2004). Based on our results, sensory reweighting appears to involve a perceptual inhibition of information from the proprioceptive system under conditions when this information is not veridical. The highest correlation was found when visual information was correct, and there was no correlation in the dark. Increased reliance of older adults on vision to control balance compared with young adults (Borger, Whitney, Redfern, & Furman, 1999) may be a contributing factor to increased sway under conditions of proprioceptive conflict. However, inhibitory processes may be integrally involved. Our inhibitory measure would not be expected to relate to an enhancement of visual information, but the availability of such information likely requires a greater inhibition of the proprioceptive information to resolve conflict. Therefore, our measure of perceptual inhibition appears to be particularly sensitive to the ability to modulate the weighting of proprioceptive and visual sensory inputs.
Our measures of perceptual and motor inhibition are linked to visuomotor tasks. Either a visual dimension (spatial location) or an overlearned perceptual–motor response (movement in pointer direction) must be inhibited. In a sway-referenced balance setting, proprioceptive information is rendered nonveridical requiring an inhibition of this information and attention to the visual scene. The correlation of measures in the elderly participants raises the possibility of a shared inhibitory process not specifically tied to one sensory system or perceptual dimension. It may be, however, that because our tests only assessed visuomotor inhibition, we are seeing correlations with sensory integration that involved visual inputs. Other inhibitory processes may be involved more with vestibular or proprioceptive signals that we did not measure. The fact that there was no correlation within the dark conditions could reflect the presence of other inhibitory processes and/or the limitation of the process measured by our perceptual inhibition task to interactions with visual inputs.
The two components of the perceptual measure (the congruous response and the incongruous response) contributed to the association between sway and the inhibitory measure in the SRF–EO condition. Sway was greater with slower incongruous responses and with faster congruous responses for the older participants. For younger participants, both the interference due to the mixing of perceptually congruous and incongruous trials and the perceptual incongruity itself elicited interference—slowing their RTs. The resolution of the interference appeared to require attentional processes that were shared with the maintenance of sway under perceptually challenging conditions, that is, sway increased with perceptual interference. This interpretation also fits the RT and sway results for the older group for perceptually incongruous RTs. The older participants, however, who responded more quickly to the perceptually congruous stimuli had greater sway in the perceptually challenging conditions. Loss of capacity to cope with interference (as indicated by overall greater interference in the old and enhanced correlation between types of inhibition) may have elicited a strategy in the older participants that explains this result. Focus on the RT task to the exclusion of the balance challenge would yield relatively faster RTs to the congruous stimuli (independent of delays when incongruity occurred) but at a cost to the maintenance of balance. Presumably, this cost was not evident in the young as inhibition could be exercised without the enhanced focus on the RT task. Further research will, however, be required to test this speculation.
Motor inhibition was not correlated with sway in any of the postural conditions, even though the motor inhibition measures for the older adults were significantly larger than the young. Thus, in this healthy population, it appears that motor inhibition does not influence quiet stance under varying conditions that require sensory conflict resolution. However, motor inhibition may still be important in postural control. One would anticipate that motor inhibition may play a role in postural conditions where active motor response is required, such as during recovery from a perturbation of the support surface. Future studies need to explore potential correlates of motor inhibition with different aspects of postural control beyond quiet stance.
The lack of any significant correlation between sway and inhibition in the young adults suggests that the effect in the older adults is due to aging. These aging effects could be due to some changes in inhibitory functioning. Note that inhibitory scores in the young participants were much better than those in the older participants. Thus, the inhibitory test battery may be tapping into processes that reflect declines in frontal and prefrontal functions that have been associated with age (Andres & Van der Linden, 2000; Phillips & Della Sala, 1988). Also, note that the older participants in this study were very healthy. They were screened for sensory health (visual, somatosensory, and vestibular), neurological health, and musculoskeletal health. Thus, the findings are probably not due to central compensation for some major reduction in peripheral sensory or musculoskeletal function. We believe that the findings are directly related to the central capability to deploy inhibition in the service of maintaining attentional focus.
The results have implications for understanding balance problems in older adults. Although it has been suggested that declines in cognitive function are associated with falling in older adults, this study finds an independent measure of cognitive function that relates to balance. Future studies investigating measures of inhibitory function in older adults with balance disorders will help define its role. Interactions between peripheral sensorimotor deficits and central inhibitory function also may be a key to understanding balance disorders in the elderly participants.
This study was supported by National Institutes of Health through grants AG14116 and AG024827.