Task coordination and aging: explorations of executive control processes in the task switching paradigm

https://doi.org/10.1016/S0001-6918(99)00011-6Get rights and content

Abstract

A number of models of cognitive aging suggest that older adults exhibit disproportionate performance decrements on tasks which require executive control processes. In a series of three studies we examined age-related differences in executive control processes and more specifically in the executive control processes which underlie performance in the task switching paradigm. Young and old adults were presented with rows of digits and were required to indicate whether the number of digits (element number task) or the value of the digits (digit value task) were greater than or less than five. Switch costs were assessed by subtracting the reaction times obtained on non-switch trials from trials following a task switch. Several theoretically interesting results were obtained. First, large age-related differences in switch costs were found early in practice. Second, and most surprising, after relatively modest amounts of practice old and young adults switch costs were equivalent. Older adults showed large practice effects on switch trials. Third, age-equivalent switch costs were maintained across a two month retention period. Finally, the main constraint on whether age equivalence was observed in task switching performance was memory load. Older adults were unable to capitalize on practice under high memory loads. These data are discussed in terms of their implications for both general and process specific cognitive aging models.

Introduction

Within the last decade or so, there has been a renewed interest among cognitive psychologists and cognitive aging researchers in executive control functions, that is those functions which are concerned with the selection, scheduling and coordination of the computational processes that are responsible for perception, memory, and action. The interest in the executive control of cognitive processes has been reflected in the development of models of cognition that reserve an important role for executive control functions Norman & Shallice, 1986, Meyer & Kieras, 1997, Meyer et al., 1995, Schneider & Oliver, 1991, Shallice, 1994 as well as in the detailed empirical examination of a subset of executive control processes of young and old adults Gopher, 1996, Kramer, Larish & Strayer, 1995, Kramer, Larish, Weber & Bardell, in press, Mayr & Kliegl, 1993, Mayr, Kliegl & Krampe, 1996, Rogers & Monsell, 1995, Verhaeghen, Kliegl & Mayr, in press.

In the studies which we report here we were particularly interested in examining the influence of aging as well as the potentially beneficial effects of practice on the efficacy of executive control processes of young and elderly adults. Given that executive control of cognition is most likely not unitary in nature but instead involves a number of different control processes Gopher, 1996, Lauber et al., 1996, Rogers & Monsell, 1995, Shallice, 1994, West, 1996 we decided to focus our investigation on a specific paradigm, referred to as task switching, which has been relatively well characterized in terms of executive control. In this paradigm, introduced by Jersild (1927), subjects perform two relatively simple tasks such as judging whether a letter is a vowel or a consonant, judging the number of elements present on a display, or comparing two multidimensional stimuli. On the non-switch trials subjects either perform these two tasks in separate blocks of trials Allport, Styles & Hsieh, 1994, Jersild, 1927 or perform one of the tasks several times before switching to the other task Gopher, 1996, Rogers & Monsell, 1995. The main point here is that subjects need only perform a single task on the non-switch trials. On the switch trials subjects either continuously alternate between two tasks (i.e. in a switch block) or instead switch to a new task after performing some number of trials on another task (i.e. in a block in which switch and non-switch trials can occur). The time required to complete the executive control processes necessary to switch from one task to the other task, such as the selection from long-term memory and configuration in working-memory of the appropriate processing algorithms and the inhibition of previously used but now inappropriate processing algorithms, is inferred from the increased response time observed when a task switch occurs compared with the response time for the same task performed separately or in a run of trials of the same task.

Several interesting results been observed in the task switching paradigm. First, several investigators have obtained data which suggests that executive control processes and task component processes (i.e. processes used to perform the separate tasks such as encoding, stimulus evaluation, response selection and response execution) are functionally independent (but see Jersild, 1927, Allport, Styles & Hsieh, 1994). For example, Gopher (1996) reported that instructions concerning the likelihood and nature of a switch influenced switching time but not component task time (i.e. non-switch time). Rogers and Monsell (1995) found that the time allotted to prepare for a task switch had a substantially larger influence on switch time than it did on component task time while the presence of a warning cue influenced component task time but not switching time. Rubinstein, Meyer and Evans (in press); (see also Lauber et al., 1996) reported a number of dissociations between non-switch and switch performance across 14 different task pairs in five separate studies. These dissociations included: (a) different component tasks with the same non-switch times but different switch times and (b) different component tasks with different non-switch times and the same switch times. In summary, the results obtained in a number of studies suggest that the processes which support switching performance are distinct, at least in part, from those processes that support performance in the separate tasks.

Researchers have also argued that multiple executive control processes can be distinguished in the task switching paradigm. Rogers and Monsell (1995) had subjects perform a sequence of trials in which they alternated between sets of two trials in which they decided whether a letter was a vowel or a constant and then two trials in which they decided whether a digit was odd or even. Alternation between the two tasks continued over a block of 24 trials. Within this paradigm Rogers and Monsell manipulated the response stimulus interval (RSI) between tasks from 150 and 1200 ms and found substantial decreases in switch costs as the RSIs increased from 150 to 600 ms but relatively stable switch costs between RSIs of 600 and 1200 ms. On the basis of these results the authors concluded that two different executive control processes could be distinguished in the task switching paradigm. They proposed that the decrease in switch costs observed when the RSI was increased from 150 to 600 ms were compatible with an endogenous, stage-like process of task re-configuration which can be carried out in anticipation of the stimulus. This stage like process likely requires the loading of processing algorithms required for the new task into working memory and the inhibition of the processing algorithms that are no longer appropriate (i.e. the processing algorithms that had been used for the previously performed task). However, such an endogenous preparatory process cannot be the whole story since a substantial and stable switch cost remained as RSIs were increased from 600 to 1200 ms. To account for this effect the authors argued that a component of the task reconfiguration process cannot be executed in advance of the stimulus but instead is triggered only exogenously by the appearance of a stimulus associated with the task to be performed.

Rubinstein et al. (in press), Lauber et al. (1996) utilized Sternberg's (1966) additive factors logic within the context of the task switching paradigm to determine (a) if the processes that underlie task switching could be distinguished from those processes which underlie the performance of the individual or component tasks and (b) to determine whether multiple executive control processes support task switching. On the basis of the performance patterns obtained in a series of studies (i.e. the patterns of additivities and interactions among experimental variables) the authors concluded that two different executive control processes could be distinguished in the task switch paradigm. One process, which the authors refer to as the rule activation stage, was argued to be responsible for the activation of task rules in preparation for the new tasks as well as the deactivation and inhibition of rules needed for the performance of the previous task. The second executive control process, which the authors refer to as the goal shifting stage, involves keeping track of the current appropriate task. Within the context of Rogers and Monsell's distinction between endogenous and exogenous executive control processes, Rubinstein, Meyer and Evans's rule activation and goal-shifting stages can be thought of as sub-components of the endogenous process. In any event, it now seems relatively clear that not only can executive control processes be distinguished from component task processes but that different executive control processes but can also be distinguished from each other.

Finally, a number of results suggest that executive control processes which are utilized in the task switching paradigm can benefit from practice. For example, Jersild (1927) reported that task switching time was reduced relative to component task time with practice for a large number of different task combinations. Rogers and Monsell (1995) also observed reduced switch costs with practice in their predictable task switching paradigm. However, even after fairly extensive practice the switch costs were still considerable (186 ms in Experiment 1; see also Rogers, 1993). Finally, Gopher (1996) reported that information concerning the possible occurrence of a switch becomes increasingly more useful with practice. Thus, it would appear that subjects require some experience with the paradigm before they can effectively employ pre-cue information to prepare for a task switch.

In his recent critical review of the literature on the neuroanatomy, neurophysiology and neuropsychology of aging West (1996) concluded that relatively strong evidence exists for the frontal lobe hypothesis of cognitive aging (see also Dempster, 1992). The frontal lobe hypothesis suggests that older adults are disproportionately disadvantaged on tasks that rely heavily on cognitive processes that are supported by the frontal and prefrontal lobes of the brain. Indeed, there is a good deal of evidence to suggest that morphological and functional changes in brain activity do not occur uniformly during the process of normal aging. Researchers have reported substantially larger reductions in gray matter volume in association areas of cortex, and in particular in the prefrontal and frontal regions, than in sensory cortical regions Coffey et al., 1992, Pfefferbaum et al., 1992, Raz et al., 1994. Studies of functional brain activity employing Positron Emission Tomography (PET) have reported similar trends, with prefrontal regions showing substantially larger decreases in metabolic activity than sensory areas of cortex Azari et al., 1992, Salmon et al., 1991.

These data on the structure and function of the aging brain are consistent with numerous reports of large and robust age-related deficits in the performance of tasks that are largely supported by the frontal and prefrontal regions of the cortex as compared to relatively small age-related deficits on non-frontal lobe tasks Ardila & Rosselli, 1989, Daigneault, Braun & Whitaker, 1992, Shimamura & Jurica, 1994. Indeed, many of the tasks subserved, in large part, by the frontal lobes involve processes associated with executive control functions such as the selection, control and coordination of computational processes that are responsible for perception, memory and action. For example, large age-related deficits have generally been reported when adults are required to perform two or more tasks at the same time or to rapidly shift emphasis among tasks Kramer, Larish & Strayer, 1995, Kramer, Larish, Weber & Bardell, in press, Rogers, Bertus & Gilbert, 1994. Functional magnetic resonance imaging and positron emission tomography studies have shown enhanced activation of regions of the prefrontal and frontal cortices when two tasks are performed together but not when they are performed separately Corbetta et al., 1991, D'Esposito et al., 1995. Kliegl and colleagues Mayr & Kliegl, 1993, Verhaeghen, Kliegl & Mayr, in press have also reported that reliably larger age-related performance decrements are observed in tasks which require coordinative operations than for tasks which require sequential operations. Therefore, in general, there would appear to be strong support for the hypothesis that older adults are more disadvantaged by tasks that rely heavily on executive control processes (and the frontal and prefrontal cortices) than tasks which rely on component task processes (i.e. perceptual and action-related processes).

Interestingly, the few instances in which the task switching paradigm has been employed to examine age-related differences in executive control processing have produced mixed results. For example, Botwinick, Brinley and Robbin (1958) had young and old adults perform subtraction and addition problems in either separate blocks of trials or in a mixed block of trials in which the problem type alternated from trial to trial. Although the absolute difference between switch and non-switch trials was larger for the old than for the young subjects, the proportional difference (switch time- non-switch time/non-switch time) although larger for the old than for the young (1.43 versus 1.08, respectively) did not attain statistical significance. Brinley (1965) had young and elderly subjects perform a variety of verbal, arithmetical, and perceptual tasks separately and in pairwise alternation. Brinley found that the older adults were proportionally slower than the young adults on the switch than on the non-switch trial blocks for eight out of nine tasks. Although such a finding could be taken to suggest disproportionate age-related slowing on switch and non-switch processes, Brinley also found that when he regressed performance on the switch and non-switch blocks of the elderly on those of the young subjects a single linear function with a slope of 1.68 provided an excellent account of the data. This function, latter to be know as a Brinley plot, suggested to Brinley that age differences in performance could be attributed to general rather than process specific slowing. Indeed, this thesis has been adopted by modern data theorists to suggest that the great majority of mental processes are slowed at uniform rate during aging Cerella, 1985, Salthouse, 1996a, Salthouse, 1996b. Finally, Hartley et al. (1990) found that older adults were slower to switch between two simple tasks than younger adults but that the proportional switch cost was equivalent for young and older adults.

In summary, there has been little consistent support obtained in the task switching paradigm, a paradigm which clearly requires executive processes subserved by the frontal lobes, for the frontal lobe hypothesis of cognitive aging West, 1996, Dempster, 1992. Why might this be the case? One possibility concerns the way in which the task switching paradigm has been implemented in the aging studies described above. In each case, switch trials were presented in separate blocks from non-switch trials. With such a design it is conceivable that factors other than the presence or absence of executive control processes, such as arousal and motivation, also varied between the switch and non-switch trials. If so, such factors might have served to mask any difference in the efficacy of the executive processes between young and old adults. In the present studies switch and non-switch trials were presented in the same trial block to preclude the influence of these potentially confounding factors on performance.

A potentially important factor which has not yet been examined during the study of age differences in task switching performance is the influence of practice on switch and non-switch processes. Thus, it is conceivable that age groups may differ in their ability to capitalize on practice to reduce task switching costs. Furthermore, the magnitude of age-related practice effects may differ on switch and non-switch trials. Such differences, if observed, would be important in indicating a dissociation, as a function of practice, between executive control and component task processes. In an effort to examine this issue both practice and retention effects are examined in the present studies.

Section snippets

Experiment 1

In this study we re-examine age-related differences in the task switching paradigm. More specifically, we were interested in (a) whether old adults would show disproportionate performance decrements on switch trials as compared to trials on which the same task is continued (i.e. non-switch trials) and (b) whether practice on the switching task would reduce age-related switch costs. Indeed, several studies have found that age-related differences in dual-task performance can be reduced through

Experiment 2

Several interesting and theoretically important results were obtained in Experiment 1. First, consistent with the frontal lobe hypothesis of cognitive aging West, 1996, Dempster, 1992 older adults showed a disproportionate decrement in the performance of the switch as compared to non-switch trials early in practice. Older adults displayed switch costs which were four times larger than those observed for younger adults during the first session of the study.

Second, and perhaps most surprising was

Experiment 3

Experiment 2 was conducted to determine whether the age-equivalence in practiced task switching performance obtained with fairly long preparatory times and low working memory load (i.e. unpredictable switches) in Experiment 1 would fail to be observed when the time to prepare for a task switch was reduced and the working memory load was increased (i.e. with predictable task switches). That is, we were interested in whether older adults ability to rapidly and accurately switch between two tasks

General discussion

The present study was conducted to examine two main hypotheses; whether large and robust age-related performance decrements would be obtained when subjects were required to switch between tasks and whether practice would serve to reduce age-related decrements in switching performance. The results obtained early in practice in each of the three studies were consistent with the hypothesis that older adults display substantial performance costs on tasks which, in large part, require executive

Acknowledgements

This research was supported by grants from the National Institute on Aging (R01 AG122203 and AG14966). We gratefully acknowledge the assistance of Lynn Bardell, Heidi Proulx and Angela Glass in running subjects.

References (53)

  • Allport, A., Styles, E., & Hsieh, S. (1994). Shifting intentional set: exploring the dynamic control of tasks. In C....
  • Ardila, A., & Rosselli, M. (1989). Neuropsychological characteristics of normal aging. Developmental Neuropsychology,...
  • Azari, N. P., Rapport, S. I., Salerno, J. A., Grady, C. L., Gonzales-Aviles, A., Schapiro, M. B., & Horwitz, B. (1992)....
  • Baddeley, A. (1992). Working memory. Science, 255,...
  • Baron, A., & Mattila, W. (1989). Response slowing of older adults: effects of time limit contingencies on single and...
  • Birren, J. E. (1956). The significance of age changes in speed of perception and psychomotor skills. In J. E. Anderson,...
  • Botwinick, J., Brinley, J. F., & Robbin, J.S. (1958). Task alternation in relation to problem difficulty and age....
  • Brinley, J. F. (1965). Cognitive sets, speed and accuracy of performance in the elderly. In A. T. Welford, J. E....
  • Cerella, J. (1985). Information processing rates in the elderly. Psychological Bulletin, 98,...
  • Coffey, C. E., Wilkinson, W. F., Parashos, I. A., Soady, A. A. R., Sullivan, R. J., Patterson, L. J., Figiel, G. S.,...
  • Cohen, J. (1988). Statistical power analysis for the behavioral Sciences (2nd ed.). Hillsdale, NJ: Lawrence...
  • Cohen, J. D., & Servan-Schreiber, D. (1992). Context, cortex and dopamine: a connectionist approach to behavior and...
  • Corbetta, M., Miezin, F. M., Dobmeyer, S., Shulman, G. L., & Petersen, S. E. (1991). Selective and divided attention...
  • Craik, F. I. M., & Jacoby, L. J. (1996). Aging and memory: Implications for skilled performance. In W. A. Rogers, A. D....
  • D'Esposito, M., Detre, J. A., Alsop, D. C., Shin, R. K., Atlas, S., & Grossman, M. (1995). The neural basis of the...
  • Daigneault, S., Braun, C. M. J., & Whitaker, H. A. (1992). Early effects of normal aging on perseverative and...
  • Dempster, F. N. (1992). The rise and fall of the inhibitory mechanism: Toward a unified theory of cognitive development...
  • Einstein, G. O., & McDaniel, M. A. (1990). Normal aging and prospective memory. Journal of Experimental Psychology:...
  • Fuster, J. M. (1989). The prefrontal cortex (2nd ed). New York: Raven...
  • Goldman-Rakic, P. S. (1987). Circuitry of primate prefrontal cortex and regulation of behavior by representational...
  • Gopher, D. (1996). Attentional control: explorations of the work of an executive controller. Cognitive Brain Research,...
  • Hahn, S., & Kramer, A. F. (1995). Attentional flexibility and aging: you do not need to be 20 years of age to split the...
  • Hasher, L., & Zacks, R. (1988). Working memory, comprehension and aging: A review and a new view. In G. K. Bower, The...
  • Jersild, A. (1927). Mental set and shift. Archives of Psychology,...
  • Kramer, A. F., Humphrey, D. G., Larish, J. F., Logan, G. D., & Strayer, D. L. (1994). Aging and inhibition: beyond a...
  • Kramer, A. F., Larish, J. F., & Strayer, D. L. (1995). Training for attentional control in dual task settings: a...
  • Cited by (357)

    View all citing articles on Scopus
    View full text