| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||||||||||
RESEARCH ARTICLE |
a Neuropsychology Unit, University of Liege, Belgium
Decision Editor: Toni C. Antonucci, PhD
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
SEVERAL researchers have suggested that cognitive decline in aging is linked to a deficit of executive functions, that is, of those control processes called into action in nonroutine or novel situations (Stuss and Benson 1986
, p. 244). For example, Daigneault, Braun, and Whitaker 1992 found an age-related decline in perseverative and nonperseverative performance on a battery of executive tests, including the Wisconsin Card Sorting Test (WCST), self-ordered pointing, Porteus Mazes, a word association test, verbal and design fluency, and Stroop interference. They concluded that their data support the notion of an age-related decrease in the ability to "regulate behavior on the basis of plans, abstract concepts, environmental feedback or one's own responses" (Daigneault et al. 1992, p. 99). Other recent studies have also sustained the link between cognitive aging and a deficit in executive functions (Brennan, Welsh, and Fisher 1997; Fristoe, Salthouse, & Woodward, 1997). Impaired executive performances have been shown in elderly people in the area of working memory as well (Daigneault and Braun 1993; Parkin and Walter 1991; Shimamura and Jurica 1994; Van der Linden, Beerten, and Pesenti 1998; Van der Linden, Bredart, and Beerten 1994). Finally, a disruption of inhibitory mechanisms has been fairly well documented in elderly people (see Zacks and Hasher 1994, for a review). Furthermore, executive processes have historically been linked to the frontal lobes (Baddeley 1986; Shallice 1988; Stuss and Benson 1986). This postulate has led a number of researchers to infer a frontal decline in elderly participants (Daigneault and Braun 1993; Parkin and Lawrence 1994; Parkin and Walter 1991, Parkin and Walter 1992; West 1996).
However, in a more global conception of age-related deficits, Salthouse, Fristoe, and Hyun Rhee 1996 have questioned the degree of independence of age-related influences on various cognitive measures considered to be sustained by different regions of the brain. They administered a battery of neuropsychological tests to 259 adults aged between 18 and 94 years. Moderate age-related declines were apparent in executive measures (postulated to be specifically sensitive to damage in the frontal lobes, e.g., WCST, trail making test, and verbal fluency), visuo-spatial measures (postulated to be sensitive to the parietal lobes, e.g., Block Design), and mnemonic measures (postulated to be sensitive to the temporal lobes, e.g., Rey Auditory Verbal Learning Test; RAVLT). As expected, most of the measures were found to have moderate negative correlations with age. Nevertheless, much of the age-related variance in a variety of different variables was shared and not all was independent and specific. Indeed, speed measures shared considerable age-related variance with the other measures. An important implication of the results of Salthouse and colleagues is that the age-related effects on the different measures were not exclusively determined by distinct and unique sets of influence. Salthouse and colleagues concluded that inferences about specific or discrete age-related influences should be interpreted with caution until there is confidence that these effects are truly distinct and independent of the general factors such as processing speed (see also Fisk and Warr 1996, for a similar proposal). However, the hypothetical links between cognitive tasks and brain localization, on which the study of Salthouse and colleagues is based, are not empirically well established (see, e.g., Janowsky, Shimamura, Kritchevsky, & Squire, 1989, for findings showing that Block Design and RAVLT performances, considered by Salthouse and colleagues to be sensitive to parietal and temporal lesions, respectively, are also impaired as a result of frontal lobe lesions).
Our aim in this study was to examine further the hypothesis of a link between cognitive aging and executive functions by using tasks designed to assess specific executive operations and that have been proved to be sensitive to frontal dysfunction. The theoretical framework we used is the control-to-action model developed by Norman and Shallice 1986. This model distinguishes two control-to-action mechanisms. The first one, called contention scheduling, is involved in the routine situations in which actions are automatically triggered. The second one, called the supervisory attentional system (SAS), is a separate mechanism at the highest level of control of action, coping with novelty. This mechanism, which is required in situations where the routine selection of actions is unsatisfactory, is involved in the genesis of plans and willed actions. Additionally, SAS is conceived as carrying out a variety of processes (Shallice 1988, Shallice 1994). Neuropsychological studies have shown that particular impairments within the general set of SAS operations can be isolated (e.g., Eslinger and Damasio 1985; Gurd 1995; Shallice and Burgess 1993, Shallice and Burgess 1996).
We investigated age-related differences in three SAS functions: planning, inhibition, and abstraction of logical rules. These SAS functions have already been studied in frontal patients, especially by Shallice and coworkers by means of specific tasks (Burgess and Shallice 1996a, Burgess and Shallice 1996b; Shallice 1982). Planning, defined as the capacity to analyze and elaborate possible solutions to a new problem, has been explored with the Tower of London (TOL) test (Owen, Downes, Sahakian, Polkey, and Robbins 1990; Shallice 1982), which requires the participant to move an initial arrangement of colored beads presented in a device to match a goal arrangement presented by the experimenter, while making the minimum number of moves. To solve the TOL, that is, to make the minimum number of moves, the participant has to analyze the problem and plan the sequence of moves before starting. This task revealed a planning impairment in patients with frontal damage (Owen et al. 1990; Shallice 1982). More specifically, frontal patients showed longer initiation and execution times than controls; that is, they were slower to make the first move and to solve the problem. Additionally, they took more moves to solve the problem (see, however, Owen et al. 1990, for results showing normal initiation times in frontal patients).
Inhibition has been recently evaluated in frontal patients by means of the Hayling test (Burgess and Shallice 1996a). In this task, the participant is presented with sentences in which the last word is missing. What this last word should be is strongly cued by the rest of the sentence. Two sections are presented. In the initiation section of the test (Section A), participants have to complete the sentence by adding the missing word. In the inhibition section (Section B), a word that makes no sense in the sentence context must be produced by the participant. Therefore, the participant has to inhibit the automatic response before generating the new one. In this test, frontal patients were slower in the two sections, suggesting that they were impaired in both processes: initiating an automatic response and inhibiting a dominant response (Burgess and Shallice 1996a). Moreover, frontal patients made more errors in Section B than control participants, confirming that they are less accurate in the resistance to well-learned responses.
Burgess and Shallice 1996b conceived the Brixton test to measure the ability to abstract logical rules. This task consists of a series of plates on which 10 circles are presented. One of the 10 circles is filled, and its position changes with each trial. The changes in positions are governed by a series of simple rules that vary without warning, and the participants must predict the filled position on each subsequent trial. In other words, participants answer correctly only when they perform the cognitive integration of the logical rule by abstracting the relation between successive cards. Burgess and Shallice 1996b administered the Brixton test to frontal patients. The results showed that these patients were impaired in the abstraction of logical rules. Moreover, frontal patients showed an abnormally high incidence of bizarre responses (responses for which no apparent rationale could be discovered).
To summarize, it has been shown that frontal patients are impaired on the TOL, Hayling, and Brixton tests, which are specifically designed to assess, respectively, planning, inhibition, and abstraction of logical rules. Additionally, Shallice and Burgess 1993 found no correlation between the performance of frontal patients on the Brixton and Hayling tests after control of IQ and age. This suggests that both tests involve different executive processes and confirms the view that the SAS can be fractionated into several independent functions (Shallice 1988, Shallice 1994).
Our aim in the present study was to reexamine the effect of age on executive functions. More specifically, we explored three types of executive processes (planning, inhibition, and abstraction of logical rules) using the same tasks that Shallice 1982 and Shallice and Burgess (Burgess and Shallice 1996a, Burgess and Shallice 1996b) used in frontal patients: the TOL, Hayling, and Brixton tests. Additionally, we investigated how the three executive tasks are related (i.e., are they related to each other or can they be dissociated?) and to what extent a slowing down of the processing speed can account for the age-related differences in executive abilities.
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
) was 35.3 (SD = 3.6). The elderly participants (24 men and 24 women) were healthy people living in the community aged between 60 and 70 years (M = 65, SD = 3.9). Their mean score on the Mill Hill Vocabulary test was 38.1 (SD = 2.7). The vocabulary scores of the elderly participants were significantly higher than those of the young participants, t(93) = -4.32, p < .0001. All participants had completed 12 or more years of schooling. The mean number of years was 14.5 (SD = 1.3) for young participants and 15.1 (SD = 2.2) for elderly participants. The two groups of participants did not differ with regard to this measure, t(93) = -1.47, p = .146.
Procedure
The tests were administered following a Latin square procedure.
TOL test.—To evaluate planning capacities, we used Coyette and Van der Linden 1992 adapted version of Shallice's TOL test (Shallice 1982). Twelve problems were proposed to each participant. In each problem (see Fig. 1) three beads, one red, one green, and one blue, had to be moved from a starting configuration on three sticks of unequal length to a target configuration in a minimum number of moves.
|
The measures derived from the TOL test offered information on both the rapidity of the plan elaboration and its efficacy. Rapidity of the plan elaboration was estimated by the time taken by the participant to start the first move (initiation time). This can also be taken as a measure of impulsiveness. Efficacy of the elaborated plan was measured by the number of moves needed to solve the problem and by the time taken from the first move to the solution of the problem (subsequent time).
Hayling test.—To evaluate participants' capacity to inhibit a prominent response, we created a French adaptation of the Hayling test used by Burgess and Shallice 1996a. It consisted of two sections (A and B), each containing 15 sentences in which the final word was omitted. What this last word should be was strongly cued by the rest of the sentence. The examiner read aloud each sentence. In Section A (initiation/automatic) participants were asked to give as quickly as possible the word they thought fitted the end of the sentence. For example, for the sentence Most sharks attack very close to the the correct response should be coast. In Section B (inhibition), participants were asked to give a word that made no sense at all in the context of the sentence. For example, for the sentence Most cats see well at participants could give the word talk.
Time of response latency was measured in both Sections A and B. Time latency in Section A assessed the participants' rapidity in initiating an automatic answer, and time latency in Section B provided information about the time they took to inhibit the dominant response and find an anomalous one. These latencies were measured with a stopwatch, which was started as soon as the last word of the sentence had been read by the examiner and stopped when the participants began responding. In Section B, we also scored errors to evaluate the efficacy of the strategy elaborated by the participant to give an anomalous response.
Brixton test.—This test evaluated participants' capacity to discover and shift logical rules. To administer the test we used the same material that Burgess and Shallice 1996b used. The test consisted of series 56 A4-sized pages that were presented 1 at a time to participants. Each page had the same basic template: a 2 x 5 array of circles numbered 1–10; the only difference between pages was the position of the filled circle. The task of the participant was to predict which circle would be filled on the next page. The correct position could be determined from that on the current page by a simple rule (see Fig. 2 for an example), which changed after between 3 and 8 pages. The experiment involved eight rule changes and six different rules. Typical rules were to move to the next lower number or to alternate between Circles 4 and 10. The number of errors made by participants was scored. We examined the nature of the errors to find out whether elderly and young participants differed qualitatively or quantitatively on this task.
|
; Van der Linden et al. 1999). In this task, participants saw 100 colored small rectangles (red, yellow, green, and blue) on a sheet of paper and had to name all the rectangles' color as quickly as possible. The overall color naming time was used as an index of processing speed. |
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
|
For problems requiring five moves (Table 2 ), we performed a two-way 2 (Group) x 3 (Type of Problem: neutral, facilitating, or misleading) ANOVA on the number of moves. This ANOVA revealed one main effect of group F(1,93) = 5.329, p < .05, showing that elderly participants took significantly more moves to solve the problems. There was also a significant main effect of type of problem, F(2,186) = 20.932, p < .0001. Newman-Keuls post hoc comparisons revealed that the number of moves was equivalent between neutral and misleading problems but was greater for neutral and misleading problems than for facilitating problems. There was no significant Group x Type of Problem interaction, F(2,186) = 0.662, p = .517.
Two-way 2 (Group) x 3 (Type of Problem) ANOVAs were carried out on the initiation and subsequent times. For the initiation time, the analysis showed one main effect of age, F(1,93) = 10.42, p < .01, but no significant effect of type of problem, F(2,186) = 0.38, p = .682, or of interaction, F(2,186) = 0.4, p = .674. For the subsequent time, significant effects of age, F(1,93) = 18.25, p < .0001, and of type of problem, F(2,186) = 11.37, p < .0001, were observed. Post hoc comparisons (Newman-Keuls) on the type of problem showed that neutral and misleading problems were similar but differed from facilitating problems. However, no significant interaction between these two factors was observed, F(2,186) = 0.82, p = .442. In sum, the analysis on the measures of time revealed that elderly participants needed significantly more time to initiate and complete the sequence of moves to solve the problems. Moreover, this difference was equivalent for the three types of problems. Insofar as elderly participants took more moves to solve the problems, the differences in the subsequent time could be due to the higher number of moves taken by elderly participants. Therefore, a two-way ANOVA was also carried out on subsequent time divided by the number of moves. This analysis revealed that a significant age effect existed, F(1,93) = 15.36, p < .0001. However, the significant effect of type of problem disappeared, F(2,186) = 2.19, p = .128. The interaction between these factors was not significant, F(2,186) = 1.51, p = .224. These results suggest that elderly participants took more time solving the problem, independent of the higher number of moves taken.
Hayling Test
The average response latencies (sum of latencies across 15 trials in seconds, as in Burgess and Shallice 1996a) in Section A of the Hayling test were 10.37 (SD = 3.7) and 11.91 (SD = 5.8) for young and elderly participants, respectively. The same measure in Section B was 39.03 (SD = 19.6) and 58.91 (SD = 32.4). The 2 (Group) x 2 (Section) ANOVA revealed a significant age effect, F(1,93) = 14.1393, p < .001, a significant effect of section, F(1,93) = 190.2077, p < .0001, and a significant interaction between these two factors, F(1,93) = 11.1896, p < .01. Post hoc comparisons (Newman-Keuls) on the interaction revealed that elderly participants were significantly slower than the younger participants to give the answer in Section B (p < .001) but not in Section A (p = .6938).
The 1,425 responses (95 participants x 15 sentences) were classified by two raters who were blind to the purpose of the study. They agreed on 76.5% of these responses. The results presented are the average of the two ratings. The same results were observed when each rating was analyzed individually.
The errors were classified by the two raters into one of three possible categories: responses that were actual completion-response received an error score of 3; responses that were semantically connected to the sentence in some way received an error score of 1; and responses unrelated to the sentence, as required by the task instructions, received an error score of 0 (Table 3 ). The proportions of the different types of response were analyzed. The analyses revealed that elderly participants committed more Type 3 errors, t(93) = -10.76, p < .001, and gave fewer Type 0 responses, t(93) = 2.283, p < .05, than young participants. The groups did not differ significantly on Type 1 responses, t(93) = -1.29, p = .199. In addition, we calculated an error score by adding the error points. Under these conditions, the average error score was 6.8 (SD = 3.4) for elderly participants and 4.8 (SD = 2.6) for young participants. This difference between performances was statistically significant, t(93) = -3.28, p < .01, showing that elderly participants globally made more errors.
|
, we considered three possible types of errors. The first type (I) was the perseverations at the stimulus-response or the set level. The second (II) was the application of the other rules that were previously relevant. The third (III) was bizarre responses and guesses, that is, responses that could not be included in first or second type because no apparent rationale could be discovered. In the original study, Burgess and Shallice 1996b considered that control participants could not make Type III errors. Indeed, they considered that all responses given by these participants were rationally driven and that only patients could produce bizarre responses and guesses. In the present study, in which no patients were involved, we took the position that both young and elderly participants can produce responses for which no underlying logical reasoning or strategy can be identified.
|
Role of Processing Speed
To examine the role of age and processing speed on the different measures of executive control (Salthouse 1996), we first analyzed performance on the color-naming task. Elderly participants were slower than young participants, 66.9 s (SD = 13.7) versus 54.2 (SD = 9.2), respectively, t(93) = -5.29, p < .0001. We then included processing speed in a correlational analysis (Salthouse 1992) to examine the relative role of age and slowing down on the different executive measures. This analysis (see Table 5 ) confirmed that executive tasks shared a significant proportion of their total variance with age. This proportion appeared to be more important for some tasks, particularly for the Brixton test.
|
Fractionation of the SAS
Finally, we examined the question of the interrelations between the three executive tasks to explore to what extent the results obtained in elderly people converge with the idea of fractionating the SAS (Shallice and Burgess 1993). In other words, our purpose in this analysis was to evaluate the hypothesis according to which the SAS fulfills different independent functions (Shallice 1988, Shallice 1994; Shallice and Burgess 1993, Shallice and Burgess 1996).
For the TOL test, the neutral-type problems were selected because no interaction between age and type of problem was observed when ANOVAs were carried out. As in Shallice and Burgess 1993, the error scores of the Hayling and Brixton tests were taken into account. Finally, the latency in Section B of the Hayling test was also considered. In sum, five measures of performance representing planning, inhibition, and rule anticipation functions were taken into account:
A unique significant correlation was observed, first of all between the Hayling error score and the initiation time of the TOL test (r = .40, p < .001) in the young adults, and then in the elderly group between the subsequent time of the TOL test and the Brixton error score ( r = .33, p < .01). Considering the total sample, there were significant correlations between initiation and subsequent time of the TOL test (r = .27, p < .01), between subsequent time of the TOL test and the Brixton error score (r= .31, p < .01) and between the Hayling and the Brixton error scores (r= .31, p < .01). When we performed partial correlation analysis by statistically controlling the general effect of age (Salthouse 1992), we found a reduction of the correlations between the initiation and subsequent time of the TOL test (r =.23, p = .01) and between the subsequent time of the TOL test and the Brixton error score (r =.23, p = .01). Moreover, the correlation between the Hayling and Brixton error scores was no longer significant.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
. Additionally, the extent to which processing speed could explain age-related differences in executive abilities was examined.
The results show that elderly people present impaired performance on the executive tasks specifically designed by Shallice and coworkers (Burgess and Shallice 1996a, Burgess and Shallice 1996b; Shallice 1982) to assess some of the SAS functions, especially planning, inhibition, and abstraction of logical rules.
First, the results on the TOL test reveal that, like frontal patients (Shallice 1982), elderly participants took more moves than young participants to solve the problems and were significantly slower on the two time measures examined, that is, initiation and subsequent times. However, partial correlations showed that controlling for processing speed significantly attenuated age-related differences in the TOL test. This result suggests that at least one part of the age-related difficulties in planning is related to a more general factor, namely the speed with which elementary operations are realized.
Second, on the Hayling test, elderly participants were slower compared with young participants on Section B but not on Section A (unlike frontal patients, who were slower also in Section A; Burgess and Shallice 1996a). Moreover, elderly participants made more errors than young participants and particularly more inhibition errors (actual completion response or error score of 3). The observed dissociation between Sections A and B tackles directly the contention scheduling/SAS (or automatic/controlled) distinction proposed by Normal and Shallice (1986): An age difference is observed in the controlled process allowing to inhibit a dominant response but not in the production of an automatic response. The results are consistent with Zacks and Hasher 1994 view that some of the age-related differences in cognition may be explained by a deficiency in inhibitory mechanisms. However, a puzzling result is that although elderly participants showed an inhibition deficit on the Hayling test, they were not particularly disrupted by misleading problems of the TOL. This suggests that the inhibitory mechanisms involved in the analysis and elaboration of a plan are not affected by age and that the inhibitory processes involved in the TOL and Hayling tests are independent. These results might indicate that the suppression of a mental representation stored in long-term memory and strongly triggered by the context of the sentence (in the Hayling test) demands a greater allocation of attentional resources than the suppression of a move that has no long-term memory representation. These findings suggest that the attentional graduation of inhibitory resources is important, as suggested by Arbuthnott 1995, and confirm the view that there exist multiple inhibitory mechanisms that are not affected in the same way by age (see Connelly and Hasher 1993). A last important result concerning the executive function of inhibition was that the differences between young and elderly participants in the error score remained after processing speed had been statistically controlled for.
Finally, our results on the Brixton test reveal a significant global influence of aging on the ability to discover and apply logical rules. Contrary to frontal lesions (Burgess and Shallice 1996b), aging increases the level of error for each error type similarly, not specifically for responses not driven by logical rules (bizarre responses). Therefore, although they made more errors, elderly participants applied logical rules to their responses. It seems, then, that the difference between frontal patients and elderly participants is that performance of elderly participants is more directed by logical rules. This result suggests that the process of rules identification is affected by age in a weaker and less specific way than by frontal lobe lesions. Moreover, the age effect persisted even after statistical control of the processing speed.
Our present findings confirm and extend the results reported in previous studies of cognitive aging (Daigneault and Braun 1993; Daigneault et al. 1992; Kramer, Hahn, and Gopher 1999; Parkin and Walter 1991, Parkin and Walter 1992; Shimamura and Jurica 1994): Elderly persons are impaired on tasks assessing executive functions. Our findings also suggest that processing speed, measured by a color-naming time, constitutes an important mediator between age and cognitive performance. However, processing speed explains only partially the performance on the executive measures, because it does not contribute to the age-related differences on the Hayling and Brixton error scores. This suggests that some age-related differences in executive tasks are linked to more specific factors than processing speed (see also Kramer et al. 1999; Spieler, Balota, and Faust 1996). More generally, a productive approach of cognitive aging taking into account both common (or shared) and specific (or unique) age-related influences is needed.
A final point of this article concerned the fractionation of the SAS. Partial correlation analysis in which we controlled the general effect of age gave rise to a disappearance of correlations between Hayling and Brixton tests. This confirms that the SAS can be fractionated into different independent executive functions, as already suggested by Shallice 1988, Shallice 1994 and Shallice and Burgess 1993, Shallice and Burgess 1996, on the basis of dissociations in frontal patients.
Globally, these age-related differences in executive functions are consistent with the studies showing a frontal decline in elderly persons. For example, in the study of Raz and colleagues 1997, age-related structural differences in the human cerebral cortex were examined in 148 healthy volunteers (aged 18–77 years) using in vivo evidence from magnetic resonance imaging. The most substantial age-related decline was observed in the volume of the prefrontal gray matter (4.9% per decade). Smaller or no age-related differences were found in the other regions, including hippocampal formation and prefrontal white matter. Additionally, Raz, Gunning-Dixon, Head, Dupuis, and Acker 1998 observed that age-related shrinkage of the prefrontal cortex might specifically mediate age-related differences in executive functions such as flexibility (evaluated by the WCST). Researchers should conduct studies to relate the age-related differences on TOL, Hayling, and Brixton tests and frontal dysfunction.
A limitation of the present study concerns the general issue of the relation between executive measures and underlying constructs such as inhibition or planning (Burgess 1997). First, executive tasks are less pure measures than non-executive ones. Because they are assessed by complex and multidetermined tasks, nonexecutive factors such as sensory deficits can affect age-related differences in executive functions to some extent. For example, it could be argued that sensory deficits contributed to the age-related differences (Baltes and Lindenberger 1997) observed in the current study. However, age-related differences in peripheral processes (e.g., vision) are shown in some but not all aspects of vision. Thus, HaegerstromPortnoy, Scheck, and Brabyn 1999 observed that, whereas under conditions of reduced contrast or luminance spatial vision measures can show some age-related differences, high-contrast acuity is well maintained with aging. It would be difficult to see how this specific deficit could explain the age-related differences observed in the tasks administered in the present study, which required clearcut perceptual discrimination only (e.g., only primary colors were used in the TOL test).
Second, the relative lack of correspondence between behavior and putative cognitive processes in executive functions (e.g., Phillips 1997) can sometimes lead to difficult interpretations of age-related differences in executive tasks. In this vein, if we consider, for example, the age-related differences in latency times of Section B of the Hayling test taken alone (without considering the result that elderly participants make more inhibition errors), it is difficult to distinguish between a diminution of the capacity to generate nonstereotypical responses and a difficulty in inhibiting strong stereotypical responses. Researchers should conduct studies to break down the complex executive tasks into their different cognitive components and to understand which particular cognitive processes (e.g., working memory) could contribute to the age-related differences observed in executive functions.
|
Acknowledgments |
|---|
We thank Professor Tim Shallice and Dr. Paul Burgess, University College of London, for offering us the material for the Hayling and Brixton tests before it was published. We also thank Fabrice Parmentier and two anonymous reviewers for many invaluable comments.
Address correspondence to Pilar Andrés, School of Psychology, Cardiff University, P.O. Box 901, Cardiff CF10 3YG, United Kingdom. E-mail:
Received for publication May 20, 1998. Accepted for publication April 13, 2000.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. S. Redfern, J. R. Jennings, D. Mendelson, and R. D. Nebes Perceptual Inhibition is Associated with Sensory Integration in Standing Postural Control Among Older Adults J Gerontol B Psychol Sci Soc Sci, September 1, 2009; 64B(5): 569 - 576. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
