Vagal influence on working memory and attention

https://doi.org/10.1016/S0167-8760(03)00073-4Get rights and content

Abstract

The aim of the present study was to investigate the effect of vagal tone on performance during executive and non-executive tasks, using a working memory and a sustained attention test. Reactivity to cognitive tasks was also investigated using heart rate (HR) and heart rate variability (HRV). Fifty-three male sailors from the Royal Norwegian Navy participated in this study. Inter-beat-intervals were recorded continuously for 5 min of baseline, followed by randomized presentation of a working memory test (WMT) based on Baddeley and Hitch's research (1974) and a continuous performance test (CPT). The session ended with a 5-min recovery period. High HRV and low HRV groups were formed based on a median split of the root mean squared successive differences during baseline. The results showed that the high HRV group showed more correct responses than the low HRV group on the WMT. Furthermore, the high HRV group showed faster mean reaction time (mRT), more correct responses and less error, than the low HRV group on the CPT. Follow-up analysis revealed that this was evident only for components of the CPT where executive functions were involved. The analyses of reactivity showed a suppression of HRV and an increase in HR during presentation of cognitive tasks compared to recovery. This was evident for both groups. The present results indicated that high HRV was associated with better performance on tasks involving executive function.

Introduction

Working memory has been assumed to involve moment to moment updating and rehearsal of information to prolong storage (Logie et al., 1990). Therefore, it can be viewed as a complex system used both for the storage of information and for the computational processing of that information. In that sense, working memory consists of a central control structure called the central executive, characterized by a flexible but limited capacity workspace (Baddeley and Hitch, 1974).

Baddeley (1986) suggested that the central executive is needed in planning future actions, decision-making and trouble-shooting. Executive components control response selection, such as the adoption of an overall strategy or plan, or the utilization of specific attentional inhibitory mechanisms during task performance (Robbins, 1996). Neural structures underlying executive function, concerned with selection and evaluation, have been assumed to be located in the prefrontal cortical areas of the brain (Luria, 1980). Executive functions are called into action when task demands are non-routine, and one way to test these functions are by tasks that involve aspects like planning, working memory, and selective and sustained attention (Robbins, 1996).

The process of sustained attention refers to an individual's ability to maintain their focus of attention and to remain alert to stimuli over prolonged periods of time (Johnsen et al., 2002). Tasks involving vigilance have come to be regarded as providing ‘the fundamental paradigm’ for defining sustained attention as a behavioral category (Jerison, 1977). The tasks require direction of attention to one or more sources of information in order to detect and respond to infrequent changes in the nature of the information being presented (Davies and Parasuraman, 1982).

One category of sustained attention tasks is the continuous performance test (CPT) originally developed by Rosvold et al. (1956). This type of test has played an increasing role in the assessment of attentional processes (Baker et al., 1995). One of the characteristics of a CPT is the involvement of higher mental workload levels such as memory search, choice reaction time (CRT), mental arithmetic, time estimation, simple tracking and grammatical reasoning (Warm, 1993). Parasuraman et al. (1987) suggested that the increase in working memory load might be the key factor in perceptual sensitivity or decrement during these tasks. Thus, working memory has been seen as a very general resource, which has played a role in a wide variety of cognitive tasks including sustained attention.

Since CPTs enable us to separate tasks into those involving and not involving executive function, the use of CPT makes it possible to compare the executive functions to the non-executive functions. Thus a generalized influence on performance can be separated from a specific effect related to prefrontal associated working memory performance.

One of the underlying mechanisms that has been associated with the understanding of attentional and memory processes is activity in the cardiovascular system. In the last decade increased attention has been directed to the fluctuation in the inter-beat-interval (IBI) between normal heartbeats. This has been referred to as heart rate variability (HRV). Porges and Raskin (1969) demonstrated that HRV was significantly reduced during sustained attention. The relation between attention and HRV has also been found in infants (Richards and Casey, 1991) and children (Porges, 1974).

More recent studies have related HRV to memory performance, mental workload and attention (Vincent et al., 1996, Middelton et al., 1999, Redondo and Delvalleinclan, 1992, Veltman and Gaillard, 1998, Backs and Seljos, 1994, Ekberg et al., 1995, Schellekens et al., 2000). Using a continuous memory task, Backs and Seljos (1994) found that as memory load increased, good performers had a small heart rate (HR) period variability decrease (i.e. higher root mean squared successive differences, rMSSD), and poor performers had a large heart period variability decrease (i.e. lower rMSSD). Furthermore, Middelton et al. (1999) found that overall HR and blood pressure variability were influenced significantly by executive and attentional tasks. Regarding the blood pressure variability, there was a significant decrease during test conditions compared to resting values. They also found lower values of HRV during attentional tasks that involved small aspects of working memory compared to executive and planning tasks. However, executive functions require higher mental load capacity, thus in accordance with Backs and Seljos's (1994) results one should expect a suppression in HRV during both these types of tasks. Thus, these issues need further investigation.

The use of physiological measures such as rMSSD have been related to the underlying autonomic activity and individual differences found in cognitive tasks. Backs and Ryan (1992) found decreased rMSSD as attention demands increased from focused to divided attention. In that experiment they manipulated the difficulty of a continuous memory task by varying the memory load, and the temporal demand. By increasing the number of target items to be counted, for the memory load, and varying the inter-stimulus interval for the temporal demand, they manipulated the difficulty of the task. Thus, vagally mediated cardiac tone could be viewed as sensitive to cognitive tasks and acting as a measure of reactivity.

Another reported factor has been the correlation between measures of cardiac vagal tone and performance data, resulting in an indication of a relationship between HRV and attention or memory processes (Backs et al., 1994). Exceptions from this correlational approach are two studies on infants and children, where cardiovascular responses were used as a predictor variable (Richards, 1987, Suess et al., 1994). The Richards (1987) study reported that infants with high respiratory sinus arrhythmia (RSA) were less distractable than infants characterized by low RSA. The Suess et al. (1994) study supported the hypothesis that high resting cardiac vagal tone was associated with good attentional capacity, and vagal tone was an index of mental effort.

One common factor for most of the studies on HRV has been the use of HRV as a dependent variable. This involves HRV as a measure of reactivity to attentional tasks that resulted in a suppression in HRV. An open question is whether resting HRV can be used as an independent variable predicting performance on cognitive tasks in adult samples.

Studies on adult samples using cardiac vagal tone as an independent variable predicting cognitive performance are scarce. Johnsen et al. (2003) investigated attentional bias in odontophobic patients, using a modified Stroop paradigm. The results showed poor attentional performance for odontophobic patients characterized by low HRV compared to patients with high HRV. This finding was evident both for identifying the color of incongruent color words and for threat words related to the phobic stimuli. The results were interpreted as the low HRV group representing a low degree of neurovisceral integration in the organism and decreased ability to organize resources to meet demands such as in an attentional task, like the Stroop-test (see also Thayer and Lane, 2000, Porges, 1992). Because the subjects have to keep in mind the instructions, in order to select the correct responses the Stroop task can be regarded as a task that utilizes executive functions. Since the study was done on a clinical population there was a problem of pre- and co-morbidity. It still remains to establish a predictive relationship between HRV and cognitive processes in healthy normal subjects.

One of the key elements in the Johnsen et al. (2003) study was the ecological validity of the situation. In order to increase this type of validity the study was performed in an odontophobic context (i.e. dental unit). Ecological validity remains a concern in experimental research on cognitive phenomenon.

In order to increase ecological validity and reduce the effect of co- and pre-morbidity one could study military personnel. Since military personnel are screened for their physical and mental health, the problem of pre- and co-morbidity is less dominant. The natural setting for military personnel operating equipment with high demands on attentional and memory processes are in teams in combat training centers. Therefore, testing cognitive functions in such a military setting would increase the ecological validity of a study.

Thus, based on Johnsen et al.'s (2003) study, we hypothesized that subjects with high HRV would show faster mRT, more correct responses and fewer false positive responses than the subjects with low HRV on cognitive tasks that utilized executive functions. On the working memory test (WMT), more true positive responses were expected for the high compared to the low HRV group. This prediction was based on findings from Backs and Seljos (1994), since good performers showed higher rMSSD during exposure to cognitive tasks than poor performers. The same pattern was predicted concerning the CPT—faster mRT and better accuracy was expected for the high HRV group compared to the low HRV group. Furthermore, on the sub-tasks that involved executive functions, we expected that the high HRV group would show faster mRT, more true positive responses and less false positive responses than the low HRV group. Regarding reactivity, a suppression of HRV was hypothesized, resulting in a lower HRV and a higher HR during the cognitive tasks. These predictions were based on the Porges and Raskin, 1969, Porges, 1972, Backs and Seljos, 1994, Backs and Ryan, 1992 studies.

Section snippets

Subjects

Fifty-three male sailors, with a mean age of 23 years, (range 18–34 years), from the Royal Norwegian Navy participated in this study. Four subjects were excluded from the data analysis because of technical problems. In addition to this one outlier was excluded because of extreme reaction time.

Experimental tasks

Two cognitive tests were presented using Micro Experimental Laboratory (version 2; Schneider, 1988) installed on a Fujitsu Life Book with 10×7.5 inch2 screen. The tests were computerized versions of a WMT

Heart rate variability and task performance

When looking at specific contrasts a group difference on the WMT was found. The high HRV group showed more true positive responses than the low HRV group, t(46)=2.24, P<0.02 (r=0.31; see Fig. 1).

The results from the CPT (all sub-tests pooled together) revealed faster mRT for the high HRV group compared to the low HRV group, t(45)=1.73, P<0.04 (r=0.25; see Fig. 2).

True positive data for all the sub-tests pooled together showed that the high HRV group had more true positive responses than the low

Discussion

The results from the present study showed that the high HRV group performed better on both the WMT and the CPT compared to the low HRV group. This was evident on accuracy measures on the WMT as well as mRT measures and accuracy measures on the CPT task, taxing executive functions. The correlation analysis from this study showed that there was a relationship between resting HRV and cognitive performance. Furthermore, the results showed an increase in HR during baseline and task conditions

Acknowledgments

The present research was supported by grants from the Norwegian Ministry of Defense and the Meltzer Foundation University of Bergen, Norway.

References (57)

  • A. Vincent et al.

    Relations among memory performance, mental workload and cardiovascular responses

    Int. J. Psychophysiol.

    (1996)
  • Backs, R.W., Ryan, A.M., 1992. Multimodal measures of mental workload during dual-task performance: energetic demands...
  • R.W. Backs et al.

    Psychophysiological measures of workload during continuous manual performance

    Hum. Factors

    (1994)
  • A.D. Baddeley

    Working Memory

    (1986)
  • D.B. Baker et al.

    Continuous performance tests: a comparison of modalities

    J. Clin. Psychol.

    (1995)
  • E.E. Benarroch

    The central autonomic network

  • D.M. Caretti et al.

    Cognitive performance during respirator wear in the absence of other stressors

    Am. Ind. Hyg. Assoc. J.

    (1995)
  • J. Cohen

    A power primer

    Psychol. Bull.

    (1992)
  • R. Colombo et al.

    A comprehensive PC solution to heart rate variability analysis in mental stress

    1989 IEEE Comput. Cardiol.

    (1990)
  • A. Compte et al.

    Synaptic mechanisms underlying spatial working memory in a cortical network model

    Cereb. Cortex

    (2000)
  • D.R. Davies et al.

    The Psychology of Vigilance

    (1982)
  • J.E. Driskell et al.

    Stress and Human Performance

    (1996)
  • K. Ekberg et al.

    Psychological stress and muscle activity during data entry at visual displayunits

    Work Stress

    (1995)
  • G.A. Ferguson et al.

    Statistical Analysis in Psychology and Education

    (1989)
  • B.H. Friedman et al.

    Validity concerns of common heart rate variability indices. Addressing quantification issues in time- and frequency-domain measures of HRV

    IEEE Eng. Med. Biol.

    (2002)
  • P.S. Goldman-Rakic

    Neurobiology. Space and time in the mental universe

    Nature

    (1997)
  • S. Gronich

    An investigation on the social components of anxiety

    Dissert.—Abstr. Int. B:—Sci. Eng.

    (1995)
  • K. Hugdahl et al.

    Separating mental arithmetic from working memory: a fMRI-study

    NeoroImage

    (2000)
  • Cited by (569)

    View all citing articles on Scopus
    View full text