Two-Way Communication Between the

Heart and the Brain

Significance of Time Within the Cardiac Cycle

BEATRICE C. LACEY Fels Research Institute

Wright State University School of Medicine
JOHN I. LACEY Pels Research Institute
Wright State University School oj Medicine

ABSTRACT: Results oj earlier and ongoing research process: Sensitive interoceptors feed back to the
dealing with time within the cardiac cycle as an ex- central nervous system information about the tim-
perimental variable are summarized. In a variety oj ing, force, volume, and pressure of each, heartbeat.
different experiments, meaningful sensorimotor events This feedback loop provides an oscillatory input
produced changes in heart rate that were systemati- to the central nervous system.
cally related to where in the cardiac cycle the events
We are concerned in this article with limited
occurred. This junction is proposed as a noninvasive
measure in intact humans oj cortically mediated ejects
aspects of the significance for behavioral science of
on vagal control oj the heart. Time within the cardiac these efferent and afferent connections. We will
cycle is also a dependent variable: Self-initiated re- deal with a relatively novel variable: time within
sponses are postponed to increasingly later times as the cardiac cycle. In our studies of efferent mech-
momentary heart rate increases. It is hypothesized anisms affecting the heart, time within the cardiac
that this may result from visceral afferent feedback to cycle is an independent variable, and we discuss,
the central nervous system via the baroreceptor nerves. in an elementary way, the biochemical processes
Preliminary results are presented from acute cat ex- at the sinoatrial node. In our studies of presuma-
periments showing that changes in frequency oj carotid bly afferent mechanisms, time is a dependent vari-
sinus stimulation, and differences in the direction oj able.
change, affect the temporal pattern of discharge of the
Our presentation is limited to bare essentials,
carotid sinus nerve.
since space does not permit an exposition of the
Since prehistoric times it has been known that the statistical and experimental controls we employed
heart has the characteristic of autorhythmicity. to provide assurance of the validity of our general
A heart removed from the chest of a slaughtered statements. Nor have we time to discuss other
animal, or from the chest of a human in prepara- than main effects.
tion for a transplant, continues to beat at its own
pace. The reason is that the origin of the heart- On the Efferent Path
beat lies within the heart itself. In the normal
heart, the beat is initiated at the sinoatrial node. We have spoken in the past of the "bradycardia
Here, a series of biochemical processes, of repeti- of attention," for one general formulation that
tive membrane depolarization and repolarization, emerged from our early work was that the inten-
results in repetitive transmission along the heart's tion to note and detect external stimuli results in
own conductive tissue system, which, in turn, re-
sults in cyclic contraction and relaxation of the This article was presented as a Distinguished Scientific
cardiac muscle. Contribution Award address at the meeting of the Ameri-
In situ, within the chest cavity, the heart's can Psychological Association, San Francisco, August 1Q77.
The research was supported by Grant MH 623 from the
autorhythmicity is modified by both neural and National Institute of Mental Health.
humoral factors. The heart is slowed and speeded Requests for reprints should be sent to either author,
Fels Research Institute, Wright State University School of
by the vagal and sympathetic cardiac efferents. Medicine, 800 Livermore Street, Yellow Springs, Ohio
But action at the effector organ does not end the 45387.

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 99

Copyright 1978 by the American Psychological Association, Inc.
All rights of reproduction in any form reserved.
slowing of the heart (Lacey, Kagan, Lacey, & cardiac cycles in which the ready and imperative
Moss, 1963). This formulation was quickly fol- stimuli occurred. The easily observed anticipa-
lowed by our observation of a then-startling fact, tory decelerations are typical. They often are
namely, that anticipatory cardiac deceleration sys- preceded by a short period of acceleration. The
tematically and reliably appeared in the foreperiod decelerations are time-locked, ending with the
of a simple visual reaction-time task, that is, in imperative stimulus and the immediately subse-
the period of time following a ready signal, during quent motor response. They range, in this figure,
which the subject is preparing to make a rapid from massive decreases of 40 beats per minute
motor response (Lacey & Lacey, 1970). (b/m) to smaller ones of 10 b/m. This basic
Figure 1 exemplifies the phenomenon in three phenomenon has by now been replicated many
subjects. The top panel for each subject is a times in other laboratories. More recently (Lacey
cardiotachometric recording of heart rate; each & Lacey, 1977), we discovered another at-first-
bottom panel shows the simultaneously recorded startling fact: One factor determining the ampli-
respiration. The letters R and S identify the tude of this anticipatory deceleration is the time
within the cardiac cycle at which the stimulus
happens to fall.
Figure 2 shows average curves for 66 male sub-
jects in a fixed 4-sec-foreperiod reaction-time ex-
periment. Heart period, rather than heart rate,
is the measure used. The duration of the cardiac
cycle was divided into tenths (or deciles). The
figure displays heart period as a function of the
decile in which each of three events happened to
occur. The events were (1) the imperative stimu-
lus, (2) the ready signal, and (3) a pseudo-event,
a control point 4 sec prior to the ready signal,
where no stimuli were presented and no motor re-
sponses made. The difference in height of the
three curves reflects the process of cardiac de-
celeration. Heart rate decelerated somewhat (that
is, heart period increased) from the control point
to the ready signal. A larger preparatory decelera-
tion occurred between the ready signal and the
imperative stimulus. But notice that the mag-
nitude of the deceleration was dependent on where
in the cycle the stimulus happened to fall. Stim-
uli occurring early in the cycle prolonged the cycle
jar more than did stimuli jailing late in the cycle.
Figure 1. Cardiotachometric recordings of heart By Ferguson's nonparametric trend test (Fer-
rate and thoracic pneumograph recordings of respira- guson, 196S), the linear decrease for the impera-
tion for three typical subjects (S30, S6S, and SOI) tive stimulus was highly significant (p < 10~ 8 ).
during a 30-sec period in which two reaction-time There was a significant quadratic trend as well
trials were presented. R identifies time of occurrence (p < .01). No significant trends were found for
of ready signal; S, the "imperative" or "go" stimulus either the ready signal or the control point.
requiring rapid release of a telegraph key that had Why did heart rate vary systematically with
been depressed at R. BPM means beats per minute. the time of presentation of the imperative stim-
Note that the anticipatory cardiac decelerations cannot
ulus? The answer, we think, lies in a temporal de-
be attributed to any consistent respiratory response.
pendency associated with vagal, but not sympa-
(From "Some Autonomic-Central Nervous System
Interrelationships" by J. I. Lacey and B. C. Lacey.
thetic, control of the heart. Such temporal de-
In P. Black (Ed.), Physiological Correlates of Emo- pendency was first reported in 1934 by Brown and
tion. New York, Academic Press, 1970, pp. 205-227. Eccles, and has been the subject of much recent
Copyright 1970 by Academic Press. Reprinted by work by Levy and his colleagues (e.g., Levy, lano,
permission.) & Zieske, 1972; Levy, Martin, lano, & Zieske,

100 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 Our results, we believe, constitute the first
demonstration in intact man, using an innocuous,
IMPMATIVC STIMULUS
noninvasive, behavioral technique, of the temporal
dependencies of the effects of vagal stimulation,
so clearly demonstrable in acute animal experi-
ments. But for direct electrical stimulation of the
vagus, we have substituted stimulation by punc-
tate sensorimotor events. The effects on the vagus
must be mediated, in our experiments, by cortical
events. Because these quantitative studies of
"cerebrovagal" competence are of interest to our
CONTROL cardiological colleagues, and because of their prom-
ise for basic psychophysiological theory, we are
intensively investigating the phenomenon.
790 Our data show clearly that not all centrally
mediated sensorimotor events yield time-dependent
1 2 3 4 5 6 7 8 91 0 cardiac slowing. The effects were absent in re-
DECILE OF CARDIAC sponse to the ready signal, although the ready
PERIOD IN WHICH signal, as well as the imperative stimulus, involved
EVENT OCCURRED both sensory input and motor response. The im-
perative stimulus, however, differed from the ready
Figure 2. Relationship of concurrent heart period
to relative cycle time of event occurrence. Ordinate
signal in several ways: (a) Intuitively, it was
is average of individual median heart periods. N = somehow more salient, (b) it required a speedy
66. (From "Change in Heart Period: A Function of motor response, and (c) it was shorter in duration,
Sensorimotor Event Timing Within the Cardiac Cycle" since both the ready and imperative stimuli were
by B. C. Lacey and J. I. Lacey, Physiological Psy- terminated simultaneously by the subject's re-
chology, 1977, 5, 383-393. Copyright 1977 by The sponse.
Psychonomic Society, Inc. Reprinted by permission.) A second experiment was designed to further
study these phenomena and to extend the paral-
1969, 1970) and by others (e.g., Dong & Reitz, lelisms with animal data. Each of 20 male sub-
1970, Reid, 1969). It is now amply documented jects self-initiated tachistoscopic exposures of a
that direct electrical stimulation of the vagus clock face outlined by light-emitting diodes placed
early in a cycle prolongs that cycle but has little in labeled clock positions 1 through 12. Each sub-
effect on the subsequent cycle. Stimulation late ject worked at his own pace, lifting his finger
in a cycle does not affect that cycle but prolongs gently from a touch plate whenever he felt ready
the next cycle, the more so, the later the stimula- to receive another brief exposure. The exposure
tion. These phenomena are attributable to the followed instantaneously. The subject's task was
time-dependent effectiveness of the action of acet- to report which light was off.
ylcholine on the sinoatrial node. Acetylcholine In Figure 3, heart period is plotted as a func-
(ACh) is the transmitter substance released upon tion of the fifth (quintile) of the cardiac cycle
vagal stimulation. It results in a decrease of the in which the self-initiated "volitional" act oc-
slope of diastolic depolarization and in an increase curred. These periods are shown by the solid line.
of the magnitude of the membrane potential dur- The dashed line is the average group heart period
ing diastole (Levy et al., 1970). Both effects for the preceding cardiac cycle. Once again, the
prolong the interbeat interval but in a time-de- heart is slowed more when significant events oc-
pendent manner. If the slope of diastolic de- cur early in the cardiac cycle. The linear decrease
polarization is diminished by the arrival of a quan- was significant (p < .01).
tity of ACh early in diastole, the cycle is prolonged Figure 4 displays a phenomenon also found in
more than if the same quantity arrives and alters the acute animal experiments. The dashed line
the slope late in diastole. If, however, ACh ar- again is the average group heart period for the
rives before diastole begins, then it is partially hy- preresponse cardiac cycle. The solid line shows
drolyzed before it can effectively begin to alter the the period of the cycle following the response.
rate of depolarization. When response occurred early in the previous cycle,

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 101

 935 db. (A) re 20 juN/m2. Eighty percent were of one
O
frequency ("standard" tones), and 20% were of
</> 930 another frequency ("rare" tones). Two rare tones
5
*m^
never appeared consecutively. The number of in-
tervening standard tones varied from two to six,
§925
each number occurring equally often. Six blocks
£ of trials were used. The number of rare tones
ol 920 varied from 20 to 25 in the different blocks, and
the number of intervening standard tones was
varied as well to maintain a 20% probability of
rare-tone occurrence. Subjects were rewarded
with 25$ for reporting correctly, at the end of
each block, the number of rare tones in the block.
We shall present first the results of two pilot
905 experiments in which the difficulty of discrimina-
tion was varied. In one experiment, half the sub-
jects (n — 10) received 1,000 Hz as the standard
1 2 3 4 5 tone and 1,200 Hz as the rare tone. For the re-
QUINTILE OF CARDIAC maining half of the subjects, the tonal frequencies
PERIOD IN WHICH
RESPONSE OCCURRED of rare and standard were reversed. Difficulty of
discrimination was increased in the second experi-
Figure 3. Relationship of concurrent heart period
to relative cycle time of subject response. Dashed
ment by using only a SO-Hz difference; the two
line indicates group average of median heart periods frequencies were thus 1,000 Hz and 1,050 Hz.
for the single preceding cardiac cycle. AT = 20. (From On the hypothesis that the relative frequency of
"Change in Heart Period: A Function of Sensorimotor occurrence of a stimulus might be one dimension
Event Timing Within the Cardiac Cycle" by B. C. defining "significance of a stimulus-event," we ex-
Lacey and J. I. Lacey, Physiological Psychology, 1977,
5, 383-393. Copyright 1977 by The Psychonomic 915
(••H,

Society, Inc. Reprinted by permission.)

$910
recovery toward faster heart-rate levels was well on
its way by this next cycle, but this next cycle was l 90S
progressively longer when response had been later
> 900
in the previous cycle. The linearly increasing
trend was significant (p < 10"5). I89S
The two experiments demonstrate that the time-
dependency phenomenon can be found under very I 890
different circumstances: The eliciting event can be
under either experimenter control or subject con- 1883
trol, and the phenomenon is seen whether or not
the motor response must be executed speedily.
1 2 3 4 5
In exploring the reasons for the failure of the QUINTILE OF CARDIAC
ready signal in the reaction-time experiment to PERIOD IN WHICH
RESPONSE OCCURRED
produce time-dependent effects, we decided to focus
on the problem of the significance or saliency of Figure 4. Relationship of subsequent heart period
the evoking event. This is the topic of our cur- to relative cycle time of subject response. Dashed
line indicates group average of median heart periods
rent research in this area. We use a procedure
for the single cardiac cycle preceding the response.
often used in studying the late positive component N = 20. (From "Change in Heart Period: A Func-
(P300) that appears in averaged scalp-recorded tion of Sensorimotor Event Timing Within the Car-
evoked potentials, since P300 itself is related to diac Cycle" by B. C. Lacey and J. I. Lacey, Physio-
some aspects of "event-significance." Tones were logical Psychology, 1977, 5, 383-393. Copyright 1977
presented, through stereo earphones, at a constant by The Psychonomic Sodety, Inc. Reprinted by per-
3-sec interval. They lasted for 100 msec, at 72 mission.)

102 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 PREC. STAND. PREC. STAND.
RARE • RARE
SUBS. STAND. • SUBS. STAND.

QUINTILE OF CARDIAC PERIOD QUINTILE OF CARDIAC PERIOD

IN WHICH STIMULUS OCCURRED IN WHICH STIMULUS OCCURRED

Figure S. Relationship of concurrent heart period to relative time of occurrence of

standard tones immediately preceding rare tones, rare tones, and standard tones immedi-
ately subsequent to rare tones. Results of an "easy" discrimination experiment in which
standard and rare tones differed by 200 Hz are shown in the right panel; results of a
more difficult discrimination experiment in which standard and rare tones differed by SO
Hz are shown in the left panel. N — 10 in both experiments.

pected to find greater time-dependency of cardiac for -both rare and preceding standard stimuli were
slowing for rare than for standard stimuli. To larger and both trends were significant: p < .01
control for time within the experiment, we com- for rare tones and p < .02 for preceding tones.
pared the results for rare tones with those for Again, cycle effects for the subsequent standard
standard tones immediately preceding and imme- tones were not significant.
diately subsequent to the rare tones. Because of this replicated difference in the ef-
Figure S shows the results. Those for the first, fects of standard tones preceding and subsequent
easy discrimination experiment are on the right. This to rare tones, we hypothesized that expectancy, not
experiment is very different from the experiments frequency of occurrence, must be the critical vari-
with reaction time and self-initiated tachistoscopic able. We reasoned that subjects learn to apprei
exposures. In particular, no overt motor response ciate the sequencing and probabilities of the tone
was required. Nevertheless, the basic effect of series, and that their expectancy of a rare tone
time-dependent cardiac slowing was again found. grows as some function of the number of inter-
Contrary to our initial hypothesis, however, the vening standard tones. Thus, both tones tended
slowing was not dependent upon frequency of event to produce the effect. However, since two rare
occurrence. Standard stimuli immediately preced- tones never occur consecutively, the standard tones
ing rare stimuli were as effective or perhaps even subsequent to rare tones would be of no conse-
more effective than rare stimuli. These standard quence for the task at hand and thus would not
tones yielded a significant time-dependent effect yield temporal dependency of cardiac slowing.
on heart period (p < .01) for the linear decreasing To study the specific role of expectancy and a
trend, while rare stimuli showed a slightly weaker variety of subsidiary issues, the more difficult dis-
effect, significant in the first half of the experi- crimination procedure is being used in a full-scale
ment only. But not all standard stimuli were experiment. The results for the first 38 subjects
effective. Standard tones subsequent to the rare (not including the pilot subjects) are shown in
tones resulted in no cycle effect at all. Figure 6. The data continue to show remarkable
In the experiment with increased difficulty of consistency. Cycle effects were significant for rare
discrimination, shown on the left in Figure S, we (p < .01) and for preceding standard tones (p
hoped both to enhance the effect and to verify the < 10-°) but not for subsequent standard tones.
differential effects for the various stimuli. We The overall effect, again, was slightly larger for
were successful in both aims. The cycle effects preceding standard tones than for rare tones.

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 103

 PREC. STAND. aged cortical evoked potentials, from vertex to
RARE
854
SUBS. STAND.
linked ears, and have begun analyses to compare
the effects of the various stimuli on the two
physiological measures. To date, we have ana-
848 lyzed the averaged evoked potentials of 10 sub-
jects. For 9 of the 10 we have found, not unex-
844
pectedly of course, that rare stimuli evoke large,
840 clear P300s. P300s to preceding standard trials
are infrequent (3 out of 10 subjects) and, when
836
they occur, are much smaller than those for rare
832 tones. P300 was not found in response to subse-
quent standard trials, a not-unexpected result in
828 view of its well-documented, selective association
5 824 with infrequent stimuli. This result contrasts with
the fact that cardiac-cycle effects on heart period
820 were so very similar when evoked by either rare
or immediately preceding standard stimuli. P300
1 2 3 4 5 and the effect of cardiac cycle time of stimulation
QUINTILE OF CARDIAC PERIOD on heart period obviously are responsive to dif-
IN WHICH STIMULUS OCCURRED ferent aspects of event significance.
Figure 6. Relationship of concurrent heart period There are divergent views of the psychological
to relative time of occurrence of three tones in a processes accompanying P300. One popular one
group of 38 subjects. Conditions are as in the more holds that P300 reflects a process of poststimulus
difficult discrimination experiment shown on the left
judgment or decision-making, arising after and in
in Figure 5.
response to significant stimuli (e.g., see Rohr-
baugh, Donchin, & Eriksen, 1974). Our results
That the effect does not habituate can be seen
are compatible with such an interpretation, al-
in Figure 7, which compares effects in the first
though they do not exclude other processes speci-
and second halves of the experiment. Although
heart period is longer in the second half of the fically associated with the infrequent stimulus. In
experiment, the cycle effects for both rare and any event, it follows that the cardiac-cycle effect
preceding standard tones are undiminished and, does not involve such a poststimulus or other
indeed, may be larger. relevant process specific to the infrequent target
In this experiment we are also recording aver- stimulus but rather is elicited nonselectively by

PREC STAND.
•RARE
• SUBS. STAND.
865 FIRST HALF SECOND HALF
860
a
« 855

85
§ °
| 845
u.
£ 840
2
1
83S
u
<830

>8«

820

1 2 3 4 5 1 "I
2 3-
3 4 5
QUINTILE OF CARDIAC PERIOD IN WHICH STIMULUS OCCURRED

Figure 7. Relationship of concurrent heart period to relative time of occurrence of

three tones in the first and second halves of the experiment shown in Figure 6. N = 38.

104 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 both the target and at least some other stimuli, TRIALS 1 AND 2
when expectation of the target stimulus is high. TRIALS 3 AND 4
The "growth" of expectant anticipation was TRIALS 5 AND 6
evaluated in this experiment by studying the ef-
fects of standard tones that had been grouped
according to their position in a series. Figure 8
shows the results. In this graph, standard trial
positions have been combined in pairs for sim-
plicity of presentation. Note first that a tendency
for heart-rate level to be slower with increasing
expectation is revealed by the differences in height
of the three curves. Note, second, that time-de-
pendent cardiac slowing also is a function of ex-
pectation, being clearly larger for Trials 3 and 4
than for Trials 1 and 2; there is another small
increase in effect for Trials 5 and 6. The function
relating time-dependent cardiac slowing to the
number of intervening standard tones, therefore,
is nonlinear. Individual trend analyses for each 1 2 3 4 5
trial position show that the cycle effect is signifi- QUINTILE OF CARDIAC PERIOD
cant by the second standard trial and for each IN WHICH STIMULUS OCCURRED
later position in the series. The only standard Figure 8. Effect of expectancy on the relationship
tone which showed no cycle effect was the one between concurrent heart period and relative times
immediately subsequent to a rare tone. Since two of occurrence of standard tones. Standard-tone trials
rare tones never occurred together, there was zero have been grouped according to their position between
probability of the target stimulus at this time. rare-tone occurrences. N = 38.
In summary: We have found in human subjects
that sensorimotor events can differentially modify results are expressed in global and sometimes un-
the duration of the cardiac cycle within which they clear concepts; very often, the outcome is only a
occur and the duration of the subsequent cycle, qualitative formulation of the obvious—that the
depending upon the time of event occurrence brain does indeed control visceral functions. Per-
within the cycle. The events which elicit such haps simple quantitative functions such as those
phenomena can be stimuli presented by the ex- just presented will prove to be more sharply cut-
perimenter through either visual or auditory chan- ting tools for analyses of basic mechanisms.
nels. Or they can be events initiated by the sub-
jects. They can, but need not, require an overt On the Afferent Path
motor response. We now know also that rare stim-
uli are sufficient, but not necessary, to produce Like the stimulus that releases the prepared-for
cardiac-cycle effects, since the cardiac effect is also response in a reaction-time experiment, the stimuli
found with standard stimuli. By contrast, rarity in the experiments just discussed have a com-
seems to be necessary to produce P300 in the manding, urgent, imperative quality. They pro-
current design. The standard stimulus is not of duce immediate cardiac slowing, the slowing being
itself a sufficient condition, since the effect is delayed at most to the next cardiac cycle if the
present only when there is some expectation of stimulus happens to occur late in diastole. The
the occurrence of the rare, target stimulus. Ex- temporal dependencies we have described seem
pectation seems to increase nonlinearly with suc- readily explained by the temporal history of the
cessive standard tones. release and hydrolysis of ACh at the sinoatrial
We have much to learn about the significance node.
and implications of these effects for the theory and What is not yet explained is why the pathways
practice of human psychophysiology. The study from brain to vagus are so promptly opened to
of cerebral-visceral systems, whether in a neuro- these significant stimuli; what is also not yet ex-
physiological, psychophysiological, or clinical set- plained is why the brain does not call for accelera-
ting, is beset with complexities. Very often, the tion, as it does, for example, in mental arithmetic

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 105

 SUBJECT calculated by computer. It can be seen that the
heart starts to slow some 7 sec before the overt
motor response and reaches a low point at the
time of response. In this subject, the average
deceleration was approximately 7 b/m. The bot-
tom half of this graph is the computer-averaged
representation of respiration. It preserves the
amplitudes and timing of inspiration (upward) and
expiration (downward). It is obvious that the
changes in heart rate are not artifacts of respira-
tory changes.
We wish to emphasize two facts: (a) that the
course of cardiac deceleration is describable as
a monotonic progression of decelerative transient
changes to successively lower heart-rate levels,
and (b), that the transients—the abrupt decreases
from one heartbeat to the next—tend to become
increasingly large as the time of voluntary motor
response approaches.
\
This cardiac preparation for self-initiated acts
-9 -8 -7 -8 -S -4 -3 -2 -1 0
SECONDS is exemplified again in Figure 10. This figure
Figure 9. Computer-averaged curves for one typi- shows the computer-averaged cardiotachometric
cal subject showing deceleration in advance of self- curve for a human subject who, working for mone-
initiated motor responses, which resulted in a brief tary reinforcement, pressed a telegraph key on an
tachistoscopic stimulus, and lack of concomitant respi- operant schedule (DRL IS sec, LH 4 sec). The
ratory changes. R is the time of the subject's re- cycle within which the press was made is shown
sponse. Respiratory curve shows times of onset of at the low point of the curve by the word press.
inspiration (upward deflection) and expiration (down- Averaging was accomplished backward and forward
ward deflection) and relative amplitude of chest move- from this cycle. We see again a series of decelera-
ment. tive transients, beginning, this time, some 6 sec
before the overt motor act. The average decelera-
or preparation for even modest muscular effort. tion is about 6 b/m. It is followed by a rebound
The simplest answer is: That's the way the or- overshooting acceleration. The tendency already
ganism is wired. illustrated for the successive transients to increase
But we wish to propose a different answer. in size is very clear in this subject.
Before doing so, we wish to add, as another phe- The question we now raise is whether these
nomenon to be explained, the fact with which we decelerations—whether preparatory or in immedi-
began this article: Given prior knowledge of the ate response to imperative stimuli—have any reg-
onset of a significant stimulus, an anticipatory de- ulatory function in behavior. Many years ago
celeration of the heart is found. Examples of this (Lacey, 19S9, 1967; Lacey & Lacey, 1970; Lacey
were presented for single reaction-time trials. Such et al., 1963), we proposed that the answer to such
anticipatory slowings, however, are not limited to a question was yes, that decelerative and hypo-
the reaction-time experiment. tensive episodes facilitate sensory and motor func-
The same anticipatory slowing was seen in the tions, because they result in a decrease of baro-
experiment with self-paced tachistoscopic exposure. receptor-mediated feedback.
Figure 9 shows the results for a typical subject Scattered up and down the arterial tree are the
in that experiment. The decreasing step-wise baroreceptors, exquisitely responsive to blood pres-
function in the top part of the figure is an average, sure. The major concentrations of these baro-
over all trials, of successive heart rates. The time receptors are in the aortic arch and in the carotid
of the response is indicated by the long vertical sinus. For reasons of convenience, the carotid
line at the right-hand extreme of the graph, topped sinus baroreceptors are studied more often than
by the letter R. From this time, beat-by-beat the aortic receptors. The carotid sinus itself is a
averages of successively earlier heart rates were thin evagination of the arterial wall at the division

106 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 of the common carotid artery into its external and
internal branches. This thin wall is distended,
stretched, and distorted by the pressure wave pro-
g 80
DRL15 LH4
SUBJ. 64E704
55 TRIALS
duced at each heartbeat. The resulting strain on
the sinus is the adequate stimulus for the baro-
receptors. At each heartbeat, then, a burst of im-
pulses is produced in the baroreceptor afferents, £76
which have their first synapse in the brain stem
in nucleus tractus solitarius, close to reticular S•7475
structures involved in cardiovascular control. 573
But activity of these baroreceptor afferents has
widespread effects on other than cardiovascular I72
activity. Of particular interest to us were the
early reports of inhibitory effects on skeletal mus- fc£
cle tone and on electroencephalographic activity
!ji " 0 1 2 3 4 5 6 7 8 9 10 11 12 13
(for an early summary, see Heymans & Neil,
TIME IN SECONDS
1958). Very recent experiments demonstrate such
inhibitory effects in single-cell recordings in areas Figure 10. Computer-averaged curve, showing for
that are functionally and anatomically very re- one typical subject anticipatory cardiac deceleration in
advance of subject-initiated key presses and postpress
mote from cardiovascular areas. Coleridge, Cole-
rebound acceleration. The subject was required to
ridge, and Rosenthal (1976) showed that inflation press the key no earlier than 15 sec, and no later than
of the carotid sinus resulted in prolonged inhibi- 19 sec, after the previous press. (From "Some Auto-
tion of pyramidal cell activity in the motor cortex! nomic-Central Nervous System Interrelationships" by
Careful controls eliminated changes in cerebral J. I. Lacey and B. C. Lacey. In P. Black (Ed.),
blood flow as the cause of the inhibition. The ef- Physiological Correlates oj Emotion. New York, Aca-
fect, then, is directly neurally implemented. Gahery demic Press, 1970, pp. 205-227. Copyright 1970 by
and Vigier (1974) showed that stimulation of Academic Press. Reprinted by permission.)
vago-aortic afferents—a significant portion of
which are baroreceptor afferents—profoundly de- the anticipatory deceleration. This has now been
pressed the responses of single cells in nucleus widely, although not universally, confirmed. But
cuneatus to skin stimulation. Gahery and Vigier the intrasubject correlations are typically low, and
(1974, p. 246) concluded cautiously that their alternative hypotheses to explain the data arose
results "suggest that vago-aortic afferents are not quickly. To date, no conclusive evidence to either
limited to a vegetative role but may also be in- confirm or refute the hypothesis is available. In
volved in a variety of physiological regulations." the resulting controversy, the favored alternative
Coleridge et al. (1976, p. 645) concluded more hypothesis seems to be that the cardiac slowing is
forcefully, "Evidence has been accumulating for the result of, or part of, a centrally prepro-
more than 40 years to support the notion that grammed, overall "somatic quieting," including
stimulation of carotid sinus baroreceptors exerts heart rate, respiration, and muscular activity, and
widespread and prolonged inhibitory effects on the that its relation to behavior—for example, reac-
central nervous system . . . It is now clear that tion time—is only an indirect consequence of this
these effects involve single neurons of the motor pattern.
cortex." To the mounting number of instances of be-
These two studies are perhaps the most con- haviorally related cardiac slowing which cannot be
clusive to date, showing that sensory and motor accounted for by this viewpoint (see particularly
functions can be inhibited by increases in barore- Bohlin & Graham, 1977; Lawler, Obrist, & Lawler,
ceptor activity. By inference, sensorimotor func- 1976), we must now add time-dependent slowing
tions can be facilitated by decreases in barorecep- evoked by "significant" events. Recent data from
tor activitiy. And this was and is the essence of our laboratory, moreover, further weaken the posi-
our hypothesis. tion that cardiac slowing itself is not somehow
We offered, as the first suggestive evidence in instrumental in the modification of behavior. The
intact man in support of the hypothesis, the fact essence of the new evidence is provided in Figure
that simple reaction time was faster, the greater 11.

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 107

 600 cated because of a statistical constraint (the dis-
SSO cussion of which here would consume excessive
500 time) that restricted analyses to trials in which
450
elapsed times were less than 600 msec. The pre-
press median heart rates ranged from 60.8 to 69.2
400
b/m, a difference of only 8.4 b/m. A straight line
350
through the endpoints yields a rough estimate of
300 the regression. There is a change of 44.3 msec
250 per unit change in heart rate. Small increases in
200 heart rate were thus associated with significant
150 postponements of the motor act to later and
100
later portions of the cardiac cycle.
This is an extreme case. The average postpone-
50
ment is about 5 or 6 msec per unit change in heart
0
1 2 3 4 5 rate. By all tests, the results were significant for
QUINTILE OF the entire group of 52 subjects. Precisely the
PRE-PRESS HEART RATE
^^^^^^^^^^^^^^^^J^^^^jl
same result was obtained in the experiment with
60.8 62.8 63.5 67.0 69.2 self-paced tachistoscopic exposure. We, at least,
CORRESPONDING are satisfied that the functional association is
MEDIAN HEART RATE (B/M)
real and generalizable across at least these two
Figure 11. For one subject, the postponement of very different experiments.
subject-initiated responses to increasingly later times These are data in search of a theory. No
within the cardiac cycle as immediately prevailing
current theory in psychophysiology can explain
heart rate increased. The curve shows the empirical
regression of central locations (medians) of distribu-
them, so far as we can see, including our own in
tions of elapsed times from the R-wave on central its present incomplete state. Our first interpre-
locations (medians) of successive heart-rate "bands," tation rested on a hazardous extrapolation from a
that is, successive fifths of the distribution of heart study by Spickler and Kezdi (1967) showing that
rates for the cardiac cycle immediately preceding the increases in the frequency of carotid sinus pulsa-
cycle in which the response occurred. tion resulted in greater baroreceptor discharge
that, moreover, was shifted forward in time rela-
These data are from 1 of the 52 subjects in tive to the phase of the input. Spickler and Kezdi,
the DRL IS sec, LH 4 sec study. There were however, used a sine-wave forcing function, a fa-
60 successful responses that fell in the time-win- vorite tool in baroreceptor physiology. But when
dow imposed by the schedule. Because the heart sine waves of pressure are used, there are perfectly
was showing anticipatory deceleration, the heart correlated changes in the rates of rise and fall of
rate for the single cardiac cycle just preceding the pressure. Since such changes are among the most
key press was used as an estimate of the heart- effective determinants of baroreceptor activity, it
rate level at which the response was emitted, an is impossible to assert that the differences in baro-
estimate uncontaminated by the response itself. receptor discharge were attributable only to fre-
The distribution of 60 prepress heart rates was quency changes. We need to know specifically if
divided into fifths. For each quintile of the heart- frequency alone has independent status as a de-
rate distributions, there was a distribution of the terminant of baroreceptor discharge, and if fre-
times, to the nearest msec, when the motor response quency changes alone cause prolongation of baro-
occurred. These were expressed as elapsed times receptor discharge.
from the onset of the R-wave, which signifies the For this purpose we have developed a pump that
expulsion of blood into the arterial tree. The me- provides a quasi-triangular wave form of pressure
dians of each distribution of elapsed times were to the carotid sinus, as illustrated in Figure 12.
plotted as a function of the heart-rate quintiles. The bottom half of each trace is the pressure
What the figure shows is that as heart rate in- within the carotid sinus, monitored by a catheter
creased, the central locations of key-press distri- in the lingual artery. In the top trace, the fre-
butions occurred later and later in the cardiac quency of stimulation is 100 pulses per minute;
cycle. The elapsed times from the R-wave ranged in the bottom trace it is 200 pulses per minute.
from 100 msec t6 472.S msec. This range is trun- One can see that the wave forms are identical at

108 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 mm Hg

/\

Figure 12. The top of each panel is a whole-nerve electroneurogram of the right caro-
tid sinus nerve (of a cat under chloralose-urethane anesthesia). The bottom of each panel
is pressure within the carotid sinus produced by a servo-controlled pump, pressure being
monitored by a catheter in the lingual artery and a strain-gauge transducer. In the top
panel, the frequency is 100 pulses/min; in the bottom panel, 200 pulses/min.

the two frequencies. The only change, then, is teristic series of dynamic transients. Indeed, the
the change in frequency, accomplished by varying data for which we seek an explanation are based
the duration of the steady-state diastolic pressure on heart-rate levels arrived at by just such a
between pulses. process. It seems important, then, to determine
The top part of each trace is the electroneuro- the effect of dynamic transients in frequency on
gram of the whole carotid sinus nerve. Each pulse baroreceptor discharge.
produces a crescendo of dense, high-amplitude Our investigations in this area are still in an
spiking activity starting at the very onset of the early stage, but the results are provocative, and we
pressure rise and diminishing shortly after the cannot resist the opportunity to present at least
pressure starts to fall. This temporal pattern of the nature of the experiment and some of the pre-
discharge is identical to that seen in response to liminary results, with the caveat that our control
normal, in vivo, carotid sinus pulsation. of the multiple variables in the experiment is as
The figure does not support the expectation that yet probably insufficient to guarantee high re-
the baroreceptor discharge is prolonged as fre- producibility of the results from one preparation
quency increases. Nor have we found this despite to another.
wide variation of pressure levels and pumping fre- Our first experiments are designed only to de-
quencies. What the figure does show is more termine whether transient changes have any spe-
high-amplitude spiking per unit of time at the cial effect. The changes we use are large ones.
higher pumping frequency than at the lower. We leave for later investigation the effects of pat-
While it may be useful to have pictured this only- terned but smaller changes in frequency.
to-be-expected fact for use in interpreting some Figure 13 illustrates the experimental approach.
of the data concerning the behavioral correlates of The bottom half of each strip shows the intra-
heart-rate level, the fact itself does not explain the sinusal pressure; the top half, the electroneuro-
postponement of motor acts to increasingly later gram. The pumping wave forms in this prepara-
parts of the cardiac cycle as heart rate increases. tion are imperfect because of a slight overshoot
That phenomenon is still in search of a theory. on the falling edge of the stimulus. This over-
Studies are now needed designed specifically to shoot, however, is identical at both frequencies
investigate questions posed by the human data. used. Under the letter A on the top trace is the
We have emphasized the fact that in the human, electroneurogram at the end of a train of 10 pulses
the anticipatory deceleration consists of a charac- at 80 b/m. Under the letter B is.the first dis-

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 109

 charge after the increase in frequency to 160 b/m. The two topmost histograms show the results for
This accelerative transient had only a slight effect decelerative transients. The lower of these two
on baroreceptor discharge: The discharge at B is graphs shows the discharge pattern obtained for
denser and has one spike of larger amplitude than 21 pulsations, each at the end of a stimulus train
at A. In the bottom ttace, C labels the discharge at 160 b/m, just before the shift to 80 b/m.
at the end of a train of 10 pulses at the high fre- There is a broadly dispersed concentration at the
quency, and D, the first discharge after the de- very short intervals, and then a fairly flat dis-
crease in frequency. This decelerative transient tribution of interspike intervals up to about in-
had a greater effect: The discharge at D is less tervals of 12 msec, with a trivial number of spikes
dense, and the spike heights are considerably at longer intervals. The count at the point la-
smaller than at C. This implies participation of belled X is the count of spike intervals greater
different subpopulations of afferent fibers at the than 31.25 msec, which was the largest interval
transient change. These are illustrations of prob- we set out to measure. The average number of
abilistic, rather than of completely deterministic, spike intervals per pulse at the high frequency of
effects. To avoid biased selection of data, a more pumping was 17.2.
rigorous form of analysis is needed, such as is The effect of the decelerative transient can be
shown in Figure 14. seen by comparing this histogram with the one
This figure shows, for another cat, the interval immediately above, which shows the results for the
histograms obtained for 21 decelerative transients 21 immediately subsequent stimuli, those follow-
and 21 accelerative transients, using all spikes ing the shift to 80 b/m. The average spike in-
above the noise level. These data were extracted terval count drops to 12.0, a decrease of 30%.
from a continuous run in which two frequencies There obviously are fewer short interspike inter-
were used in a completely counterbalanced order: vals, and this continues to be true up to about in-
abba, baab, and so on, where a stands for a train tervals of 12 msec. There is a minor flurry of
of 5 pulses at 160 b/m, and b for a train of 5 activity at about 12-msec intervals, the significance
pulses at 80 b/m. There were 64 stimulus-trains of which we do not know.
in a pumping run. The low- and high-pressure The two histograms at the bottom display the
levels were 90 mm Hg and ISO mm Hg. The results for accelerative transients. The observable
rise time was 72 msec, yielding a rate of pressure differences are minor.
rise of 833 mm Hg/sec, a value within the physio- The differences between accelerative and de-
logically normal range. celerative transients are clearly shown in Figure

mm Hg

180

CAT 491

Figure 13. The top of each panel is a whole-nerve electroneurogram of the right carotid
sinus nerve (of a cat under chloralose-urethane anesthesia). The bottom of each panel is
pressure within the carotid sinus. The top panel illustrates the small effect of accelerative
transients: At A is the discharge after a sequence of pulsations at 80 b/m, and at B is
the discharge at the shift to a frequency of 160 b/m. Bottom panel shows a greater effect
of decelerative transients: At C is the discharge after a sequence of pulsations at 160 b/m,
and at D, the discharge at the shift to a frequency of 80 b/m.

110 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

 PUMP B'M SB
RVE.RR&E. N 12.

20 30 X

PUMp B/M 168

RVERflGt N 17.2

0 10 20 30 X
INTERVRL BETWEEN SPIKES IN MILLISECONDS
717N9 21 D E C E L E R R T I V E TRflNSIENTS

20-,
PUMD R'M en
WRRGE N \2.k

-T^
10 20 30

PUMD fi/M 163

RVERfl&t N 13.IJ

j.iuL.u
10 20 30 X
INTERVRL BETWEEN SPIKES IN MILLISECONDS
717N10 21 RCCELERRTIVE TRRNSIENTS
Figure 14. Interval histograms from whole-nerve electroneurograms of right carotid
sinus nerve (of a cat under chloralose-urethane anesthesia) for decelerative transients from
160 pulses/min to 80 pulses/min, and for accelerative transients from 80 pulses/min to
160 pulses/min. Data gathered for duration of carotid sinus forcing pulse, as in Figures
12 and 13.

15, which displays cumulative percentage curves Figure 16 shows cumulative curves for steady-
for the data just presented. On the left are the state pumping, for the same pumping run from
curves for decelerative transients. The curve for which the transient data were extracted. The
low-frequency pumping is identified by a down- number of spike intervals is slightly greater at
ward-going tick; that for high-frequency pumping high- than at low-frequency pumping, averaging
by an upward-going tick. The computer has placed 13.3 and 11.7 per pulse, respectively. The mag-
the ticks at the 50% point of the curves. One nitude of the difference in location of the 50%
can see that the proportion of spikes at very short points lies between that found for decelerative and
intervals is identical for only the first two inter- accelerative transients. The differences obtained
vals. The curves then diverge sharply. As the for steady states, then, give no hint of the differ-
slopes of these curves show, at high-frequency ential effects of changing from one frequency to
pumping, short ihterspike intervals accumulate in another.
increasingly larger proportion than at low-fre- There was, therefore, under the conditions in
quency pumping, up to intervals of about 10 msec, this preparation, a dynamic asymmetry of response
and the 50% points are widely separated. at this first input phase of the feedback process.
The curves on the right are for accelerative Decelerative transients resulted in larger decreases
transients. They are close neighbors throughout, in presumably inhibitory input to the central ner-
and the 50% points are rather close together. vous system than would be expected from steady-

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 111

 100 -i lOO-i

0 10 0 10 20 30
INTERVflL IN MILLISECONDS INTERVAL IN MILLISECONDS
717N9 DECEL. TRRNSIENTS 717N10 RCCEL. TRflNSIENTS
Figure IS. Cumulative percentage curves for the data in Figure 14. Upward-going
tick marks identify the 50% points for high-frequency pumping; downward-going ticks,
the 50% points for low-frequency pumping.

state data. Accelerative transients resulted in these amplitude differences are caused, in unknown
smaller increases in that input than would be ex- proportion, both by differences in fiber diameter
pected from steady-state data. and in distance from the recording electrodes, sys-
In our most recent experiments, we pass the tematic variation with frequency in the behavior
electroneurogram through a circuit in which out- of spikes of different amplitudes could be attribu-
put voltage is approximately the square of input table only to differences in the functional proper-
voltage, thus dramatically increasing the differ- ties of fiber subpopulations.
ences among spikes of different amplitudes, for In our most recent preparation, we seem to have
easy selection by a window-discriminator. While found a subpopulation of fibers with surprising
properties. While steady-state data again showed
100 -i more spikes with high-frequency pumping, and also
a proportionately greater number at short inter-
spike intervals, transients in either direction in-
LD
or creased the number of discharges. Decelerative
LjJ
transients caused a greater increase than accelera-
LJ
or tive transients! For decelerative transients, how-
ever, this increased incidence was accompanied by
•50 - a shift to longer interspike intervals. For accelera-
tive transients, there were no differences in the
CE distributions across intervals.
A long series of experiments lies ahead. The
v < an B/M N = 1873
1 y 169 8/M N = 2122 data we have so far gathered seem to us to clearly
support only the general principle that the effects
1 1 of carotid sinus pulsation on baroreceptor dis-
D 10 20 30 charge depend on the immediate past history of
INTERVRL IN MILLISECONDS stimulation. The results presented above should
7 .7N8 STEflDY STflTES be taken only as suggestive of some of the detailed
ways in which this general principle may be ex-
Figure 16. Cumulative percentage curves of in- pressed. One of our most important tasks will be
terval histograms of steady-state discharges of the
to determine which aspects of these differences are
right carotid sinus nerve in the same run in which
the data for accelerative and decelerative transients
effective in modifying central nervous system func-
shown in Figure 14 were obtained. Results for high- tion in behaviorally significant ways. Perhaps
frequency and low-frequency pumping, again, are iden- these studies will make it clearer why heart rate
tified by upward-going and downward-going tick marks, responds as it does, and how the cardiovascular
respectively, placed at the 50% points of the curve. system—either fortuitously or through evolution-

112 • FEBRUARY 1978 • AMERICAN PSYCHOLOGIST

ary selection—has come to exert, or to be asso- psychotherapy, Washington, D.C.: National Publishing,
1959.
ciated with, modulating effects on sensorimotor
Lacey, J. I. Somatic response patterning and stress: Some
behavior. revisions of activation theory. In M. H. Appley & R.
One important function of theory is to guide Trumbull (Eds.), Psychological stress: Issues in research.
research. Another is to explain empirical data. New York: Appleton-Century-Crofts, 1967.
Lacey, J. I., Kagan, J., Lacey, B. C., & Moss, H. A. The
The concept of behaviorally relevant two-way com- visceral level: Situational determinants and behavioral
munication between heart and brain, it seems to correlates of autonomic response patterns. In P. H.
us, is viable on both counts. Knapp (Ed.), Expression of the emotions in man. New
York: International Universities Press, 1963.
Lacey, J. I., & Lacey, B. C. Some autonomic-central
REFERENCES nervous system interrelationships. In P. Black (Ed.),
Physiological correlates of emotion. New York: Aca-
Bohlin, G., & Graham, F. K. Cardiac deceleration and
demic Press, 1970.
reflex blink facilitation. Psychophysiology, 1977, 14,
Lawler, K. A., Obrist, P. A., & Lawler, J. E. Cardiac
423-430.
and somatic response patterns during a reaction time
Brown, G. L., & Eccles, J. C. The action of a single
task in children and adults. Psychophysiology, 1976,
vagal volley on the rhythm of the heart beat. Journal
13, 448-455.
of Physiology (London), 1934, 82, 211-240.
Levy, M. N., lano, T., & Zieske, H. Effects of repetitive
Coleridge, H. M., Coleridge, J. C. G., & Rosenthal, F.
bursts of vagal activity on'heart rate. Circulation Re-
Prolonged inactivation of cortical pyramidal tract neu-
search, 1972, 30, 186-195.
rones in cats by distension of the carotid sinus. Jour-
Levy, M. N., Martin, P. J., lano, T., & Zieske, H.
nal of Physiology (London), 1976, 256, 635-649.
Paradoxical effect of vagus nerve stimulation on heart
Dong, E., Jr., & Reitz, B. A. Effect of timing of vagal
rate in dogs. Circulation Research, 1969, 25, 303-314.
stimulation on heart rate in the dog. Circulation Re-
search, 1970, 27, 635-646. Levy, M. N., Martin, P. J., lano, T., & Zieske, H.
Ferguson, G. A. Nonparametric trend analyses. Mon- Effects of single vagal stimuli on heart rate and atrio-
treal, Canada: McGill University Press, 1965. ventricular conduction. American Journal of Physi-
Gahery, Y., & Vigier, D. Inhibitory effects in the cuneate ology, 1970, 218, 1256-1262.
nucleus produced by vago-aortic afferent fibers. Brain Reid, J. V. O. The cardiac pacemaker: Effects of regu-
Research, 1974, 75, 241-246. larly spaced nervous input. American Heart Journal,
Heymans, C., & Neil, E. Reflexogenic areas of the cardio- 1969, 78, 58-64.
vascular system, Boston: Little, Brown, 1958. Rohrbaugh, J. W., Donchin, E., & Eriksen, C. W. De-
Lacey, B. C., & Lacey, J. I. Change in heart period: A cision making and the P300 component of the cortical
function of sensorimotor event timing within the cardiac evoked response. Perception &• Psychophysics, 1974, 15,
cycle. Physiological Psychology, 1977, 5, 383-393. 368-374.
Lacey, J. I. Psychophysiological approaches to the evalu- Spickler, W. J., & Kezdi, P. Dynamic response charac-
ation of psychotherapeutic process and outcome. In teristics of the carotid sinus baroreceptors. American
E. A. Rubinstein & M. B. Padoff (Eds.), Research in Journal of Physiology, 1967, 212, 472-476.

Gold Medal Award: 1978

The Trustees of the American Psychological Foundation invite the members of the
American Psychological Association to nominate candidates for the 1978 Gold Medal
Award. The Award is given to an American psychologist in recognition of a distin-
guished and long-continued record of scientific and scholarly accomplishments. The
Award is limited to psychologists 65 years of age or older and to those residing in
North America. Nominations should be accompanied by a statement highlighting the
accomplishments of the nominee. The nominees on each year's ballot will auto-
matically be reconsidered for the Gold Medal Award in all subsequent years. Nomi-
nations should be addressed to: Secretary, American Psychological Foundation, 1200
Seventeenth Street, N.W., Washington, B.C. 20036, to be received no later than
March 15, 1978. The decision will be made by the Trustees of the American Psy-
chological Foundation.

AMERICAN PSYCHOLOGIST • FEBRUARY 1978 • 113