Biofeedback and Self-Regulation, VoL 2, No. 2, 1977
Baroreflex Sensitivity during Operant Blood Pressure Conditioning D. S. Goldstein, A. H. Harris, and J. V. Brady The Johns Hopkins University School o f Medicine
Baroreflex sensitivity was measured in baboons operantly conditioned to increase their diastolic blood pressure in daily, 12-hr sessions, by using the extent o f increases in interpulse interval per unit o f increase in systolic pressure after intravenous phenylephrine injection as an index o f baroreflex sensitivity. Following training, baroreflex sensitivity increases averaging 32% were observed before and after the 12-hr conditioning sessions. During the conditioning sessions, however, consistent diastolic blood pressure elevations averaging 1707o (14 mmHg) were accompanied by significant decreases in baroreflex sensitivity averaging 44% relative to the increased "'before" and "after'" sensitivity levels. The results suggest that changes in baroreflex sensitivity participate in operantly conditioned blood pressure changes. Several reports have described operantly conditioned blood pressure changes using biofeedback on both laboratory animals (Harris, Gilliam, Findley, & Brady, 1973; Miller, 1969) and humans (Benson, Shapiro, Tursky & Schwartz, 1971; Shapiro, Schwartz, & Tursky, 1972). Adequate understanding of the physiologic processes underlying these conditioned changes requires consideration of the complex system of homeostatic mechanisms controlling blood pressure (Korner, 1971; Sagawa, Kumada, & Schramm, 1974). Changes in arterial baroreceptor activity are of importance in this regard, because they constitute the major source of afferent information to the brain about changes in the systemic circulation (Korner, 1972) and because the baroreceptor reflex (baroreflex) is a particularly powerful and rapidly acting circulatory homeostatic mechanism (Forsyth, 1974). The arterial baroreceptors are nerve endings located in the adventitia of the walls of large blood vessels--particularly in the carotid sinus and aortic arch--which respond to stretching of the vessel walls by generating 127 3"his journal is c o p y r i g h t e d by Plenum. Each article is available for $7.50 f r o m Plenum Publishing C o r p o r a t i o n , 227 West 17th Street, New Y o r k , N.Y, 10011.
128
Goidstein et al.
impulses along afferent neural pathways to the brainstem (Sagawa et al., 1974). Barosensory signals impinge on suprabulbar sites including the central gray of the midbrain (Hockman, Livingston, & Talesnik, 1972), the anterior hypothalamus (Spyer, 1972), and the amygdala (Hockman et al., 1972). The final efferent pathway of the reflex includes both sympathetic adrenergic (Heymans & Neil, 1958) and cardiac vagal (Jewett, 1964) nerves. In response to phasic increases in blood pressure, the baroreflex produces decreases in heart rate (Kezdi, 1967), cardiac contractility (DeGeest, Levy, & Zieske, 1964), cardiac output (Kumada & Iriuchijima, 1965), and arterial resistance (Heymans & Neil, 1958). All these effects tend to oppose the initial perturbations in blood pressure, so that the baroreflex acts to buffer phasic blood pressure changes in a homeostatic fashion. In 1969, Smyth, Sleight, and Pickering reported a simple method for measuring the sensitivity of the cardiac limb of the baroreflex which would not disturb the conscious, behaving animal. The index of sensitivity was the extent of increase in interpulse interval per unit of increase in systolic blood pressure after intravenous injection of a peripheral vasoconstrictor having no direct effect on the heart. We used this technique to assess baroreflex sensitivity in baboons trained using operant conditioning and biofeedback to increase their diastolic blood pressures in daily, 12-hr sessions. Decreases in baroreflex sensitivity accompany some experimental models of hypertension in laboratory animals (Aars, 1968; Alexander & DeCuir, 1966; McCubbin, Green, & Page, 1956), and clinical hypertension in humans (Bristow, Honour, Pickering, Sleight, & Smyth, 1969; Gribbin, Picketing, Sleight, & Peto, 1971; Takeshita, Tanaka, Kuroiwa, & Nakamura, 1975). Our working hypothesis therefore was that baroreflex sensitivity would decrease during periods of operantly conditioned hypertension.
METHODS
Six adult male baboons (Papio anubis), each weighing approximately 20 kg, served as subjects. Each animal was maintained in a primate restraining chair (Findley, Robinson, & Gilliam, 1971), and housed in a sound-reducing experimental chamber provided with stimulus lights and an automatic food dispenser, as described in a previous report (Harris et al., 1973). Each animal was surgically prepared (Werdegar, Johnson, & Mason, 1964) with two silicone-coated polyvinyl catheters, the arterial catheter implanted either into a femoral or carotid artery and the venous catheter into the corresponding vein. The distal end of each catheter was tunnelled under the skin and exited in the interscapular area, fitted with an 18-gauge Luer stud adapter, and the arterial catheters connected to a Statham transducer (P23De) mounted on the outside-top of the experimental chamber. The proximal end of each femoral catheter was advanced to just above the iliac
Baroreflex Sensitivity and Blood Pressure Conditioning
129
bifurcation, while that of each neck catheter was advanced until intrathoracic. Patency of the arterial catheter was maintained by continuous infusion of lightly heparinized saline (5,000 USP units per liter) at a constant rate of about 4 ml/hr and by a more rapid " f l u s h " once each day. Daily calibration of the system was accomplished without dismantling the components by integration of a mercury manometer through a series of three-way valves (Findley, Brady, & Robinson, 1971). Pressure signals from the transducer were amplified and displayed on an Offner polygraph (Type R) which provided continuous heart rate and beat-by-beat blood pressure recordings. In addition, the pressure and rate signals were analyzed by an electronic averager (Swinnen, 1968) which provided on-line printout of heart rate and systolic and diastolic blood pressures over consecutive 40-rain intervals. Throughout the extended course of the experiment, blood pressure and heart rate were measured continuously, 24 hours each day, and adjustable meter relays integrated with the physiologic recording system provided for selection of criterion pressure levels and automatic programming of contingent food and shock events. Two "feedback" lights, one white and one red, were mounted in front of the animal and signalled when diastolic blood pressure was above or below the prescribed criterion, respectively. After recovery from the surgery, a baseline period lasting approximately 6 weeks provided for continuous 12-hr presentations of the white signal light beginning at noon each day during which five 1-g food pellets were delivered every 10 min. The instrumental conditioning procedure required the animals to maintain prespecified diastolic blood pressure levels in order to obtain food and avoid shock. Five 1-g pellets were delivered to the animal for every I0 rain of accumulated time that the diastolic pressure exceeded criterion, as indicated by the white light. Conversely, the animal received a single, 8 mA, 0.25 sec electrical shock through stainless-steel electrodes applied to a shaved portion of the animal's tail for every 60 sec of accumulated time that the diastolic pressure was below criterion, as indicated by the red light. Additionally, each food delivery reset the shock timer postponing the delivery of shocks for an additional 60 sec, and each occurrence of an electric shock reset the food timer, postponing the delivery of food for at least an additional 10 rain. At the beginning of the training period, the criterion pressure levels were determined by the animal's preconditioning, resting, baseline diastolic pressure (approximately 75 mmHg). Increases in the criterion occurred at a gradual, progressive rate of about 1-2 mmHg per week over a period of 2 to 3 months. Conditioning sessions occurred daily during a 12-hr period from noon to midnight ("conditioning on") alternating with a 12-hr period from midnight to noon during which no programmed contingencies were in effect ("conditioning off"). Within the 2 to 3 month interval of training,
130
Goldstein et ai.
the baboons attained diastolic blood pressure levels during the conditioning sessions which were significantly above baseline levels, and they maintained these elevated pressures for more than 95°70 of each 12-hr session. Baroreflex sensitivity was determined using the method of Smyth et al. (1969) by measuring the extent of increase in interpulse interval per unit of increase in systolic blood pressure after intravenous injection of phenylephrine, a peripheral vasoconstrictor. These determinations were made in the baseline and training periods before, during, and after the conditioning sessions (11 AM, 11:45 AM, 12:15 PM, 6 PM, 11:45 PM, 12:15 AM, and 1 AM). For each determination, phenylephrine (7.5 gg/kg) was infused into the venous catheter and flushed over a 10-second period with 5 cc isotonic saline at room temperature. This procedure consistently produced increases in both systolic and diastolic pressure of about 25 mmHg which lasted for about 3 rain. Approximately 20 heart beats after the initiation of each flush, systolic, and diastolic pressures began to rise to a maximum over about an eight-beat interval, associated with a decrease in heart rate (Fig. 1). With each beat starting from the beat where a change in systolic pressure was first detected and ending with the beat where systolic pressure peaked, an interpulse interval was associated--the interval for the succeeding beat (Pickering and Davies, 1973). Interpulse interval in milliseconds was calculated by dividing the beat-to-beat heart rate into 60,000. For each infusion, the relationship between the interpulse interval and systolic pressure (the "baroplot," Fig. 2) was determined over the approximately eight-beat interval of pressure increase, and the slope of the resulting linear plot (refecting significant positive correlations between interpulse interval and systolic pressure for more than 95°7o of the injections) defined the baroreflex sensitivity. Dropped beats (i.e., beats for which the interpulse interval exceeded by more than 80°7o the interval for the proceding beat) were not included in the data analysis.
! HnT,.
- ~...~ ~
FLUSH
~
~
~..,,e'~,.-., --t-
CHARLI| 4-16-7S 11411PM PHI~IIYLEPI41tlNIE 1S0jU~*
Fig. 1. Polygraph recording showing the relationship between blood pressure and heart rate changes followingphenylephrine infusionin baboon "Charlie."
Baroreflex Sensitivity and Blood Pressure Conditioning 700
-
131
-
650-vt
Y-" 3.92 X - 39. 50 R: 0.95
~" 550
5O0
450
4001
1 [ J ..~__J-_ 130 140 150 160 170 SYSTOLIC PRESSURE, M M . OF HG
Fig. 2. The relationship between the interpuise interval and systolic pressure determined over the eight-beat interval of pressure increase shown in Figure t. The slope of the resulting linear plot defined the baroreflex sensitivity.
RESULTS
Baseline Baseline blood pressure and baroreflex sensitivities are shown in Table I. Analyses of variance conducted on these data showed no significant trends over the listed times for any of these variables. No significant differences were obtained between the values for these variables during the times corresponding to Conditioning Off (11 AM, 11:45 AM, 12:15 AM, and 1 AM) and the times corresponding to Conditioning On (12:15 PM, 6 PM, and 11:45 PM).
Training Table II shows the 40-min averaged systolic and diastolic blood pressures and baroreflex sensitivities during training. All the subjects showed clear-cut, consistent elevations in systolic and diastolic pressure during Conditioning On as opposed to Conditioning Off (t = 8.27, df = 5,
132
Goldstein et al.
o
~
~: ~
~
~
~
~
~
o
E E
=
lo.
..
,~-~ Jl
II °
Baroreflex Sensitivity and Blood Pressure Conditioning
133
.=,
•
=
~
~
~
~
?
N •
N
~
~
~
~
~.
~
t~ -
~'~
o,t
N ,~
o
0
&
E~ .p.
Baroreflex Sensitivity during Operant Blood Pressure Conditioning D. S. Goldstein, A. H. Harris, and J. V. Brady The Johns Hopkins University School o f Medicine
Baroreflex sensitivity was measured in baboons operantly conditioned to increase their diastolic blood pressure in daily, 12-hr sessions, by using the extent o f increases in interpulse interval per unit o f increase in systolic pressure after intravenous phenylephrine injection as an index o f baroreflex sensitivity. Following training, baroreflex sensitivity increases averaging 32% were observed before and after the 12-hr conditioning sessions. During the conditioning sessions, however, consistent diastolic blood pressure elevations averaging 1707o (14 mmHg) were accompanied by significant decreases in baroreflex sensitivity averaging 44% relative to the increased "'before" and "after'" sensitivity levels. The results suggest that changes in baroreflex sensitivity participate in operantly conditioned blood pressure changes. Several reports have described operantly conditioned blood pressure changes using biofeedback on both laboratory animals (Harris, Gilliam, Findley, & Brady, 1973; Miller, 1969) and humans (Benson, Shapiro, Tursky & Schwartz, 1971; Shapiro, Schwartz, & Tursky, 1972). Adequate understanding of the physiologic processes underlying these conditioned changes requires consideration of the complex system of homeostatic mechanisms controlling blood pressure (Korner, 1971; Sagawa, Kumada, & Schramm, 1974). Changes in arterial baroreceptor activity are of importance in this regard, because they constitute the major source of afferent information to the brain about changes in the systemic circulation (Korner, 1972) and because the baroreceptor reflex (baroreflex) is a particularly powerful and rapidly acting circulatory homeostatic mechanism (Forsyth, 1974). The arterial baroreceptors are nerve endings located in the adventitia of the walls of large blood vessels--particularly in the carotid sinus and aortic arch--which respond to stretching of the vessel walls by generating 127 3"his journal is c o p y r i g h t e d by Plenum. Each article is available for $7.50 f r o m Plenum Publishing C o r p o r a t i o n , 227 West 17th Street, New Y o r k , N.Y, 10011.
128
Goidstein et al.
impulses along afferent neural pathways to the brainstem (Sagawa et al., 1974). Barosensory signals impinge on suprabulbar sites including the central gray of the midbrain (Hockman, Livingston, & Talesnik, 1972), the anterior hypothalamus (Spyer, 1972), and the amygdala (Hockman et al., 1972). The final efferent pathway of the reflex includes both sympathetic adrenergic (Heymans & Neil, 1958) and cardiac vagal (Jewett, 1964) nerves. In response to phasic increases in blood pressure, the baroreflex produces decreases in heart rate (Kezdi, 1967), cardiac contractility (DeGeest, Levy, & Zieske, 1964), cardiac output (Kumada & Iriuchijima, 1965), and arterial resistance (Heymans & Neil, 1958). All these effects tend to oppose the initial perturbations in blood pressure, so that the baroreflex acts to buffer phasic blood pressure changes in a homeostatic fashion. In 1969, Smyth, Sleight, and Pickering reported a simple method for measuring the sensitivity of the cardiac limb of the baroreflex which would not disturb the conscious, behaving animal. The index of sensitivity was the extent of increase in interpulse interval per unit of increase in systolic blood pressure after intravenous injection of a peripheral vasoconstrictor having no direct effect on the heart. We used this technique to assess baroreflex sensitivity in baboons trained using operant conditioning and biofeedback to increase their diastolic blood pressures in daily, 12-hr sessions. Decreases in baroreflex sensitivity accompany some experimental models of hypertension in laboratory animals (Aars, 1968; Alexander & DeCuir, 1966; McCubbin, Green, & Page, 1956), and clinical hypertension in humans (Bristow, Honour, Pickering, Sleight, & Smyth, 1969; Gribbin, Picketing, Sleight, & Peto, 1971; Takeshita, Tanaka, Kuroiwa, & Nakamura, 1975). Our working hypothesis therefore was that baroreflex sensitivity would decrease during periods of operantly conditioned hypertension.
METHODS
Six adult male baboons (Papio anubis), each weighing approximately 20 kg, served as subjects. Each animal was maintained in a primate restraining chair (Findley, Robinson, & Gilliam, 1971), and housed in a sound-reducing experimental chamber provided with stimulus lights and an automatic food dispenser, as described in a previous report (Harris et al., 1973). Each animal was surgically prepared (Werdegar, Johnson, & Mason, 1964) with two silicone-coated polyvinyl catheters, the arterial catheter implanted either into a femoral or carotid artery and the venous catheter into the corresponding vein. The distal end of each catheter was tunnelled under the skin and exited in the interscapular area, fitted with an 18-gauge Luer stud adapter, and the arterial catheters connected to a Statham transducer (P23De) mounted on the outside-top of the experimental chamber. The proximal end of each femoral catheter was advanced to just above the iliac
Baroreflex Sensitivity and Blood Pressure Conditioning
129
bifurcation, while that of each neck catheter was advanced until intrathoracic. Patency of the arterial catheter was maintained by continuous infusion of lightly heparinized saline (5,000 USP units per liter) at a constant rate of about 4 ml/hr and by a more rapid " f l u s h " once each day. Daily calibration of the system was accomplished without dismantling the components by integration of a mercury manometer through a series of three-way valves (Findley, Brady, & Robinson, 1971). Pressure signals from the transducer were amplified and displayed on an Offner polygraph (Type R) which provided continuous heart rate and beat-by-beat blood pressure recordings. In addition, the pressure and rate signals were analyzed by an electronic averager (Swinnen, 1968) which provided on-line printout of heart rate and systolic and diastolic blood pressures over consecutive 40-rain intervals. Throughout the extended course of the experiment, blood pressure and heart rate were measured continuously, 24 hours each day, and adjustable meter relays integrated with the physiologic recording system provided for selection of criterion pressure levels and automatic programming of contingent food and shock events. Two "feedback" lights, one white and one red, were mounted in front of the animal and signalled when diastolic blood pressure was above or below the prescribed criterion, respectively. After recovery from the surgery, a baseline period lasting approximately 6 weeks provided for continuous 12-hr presentations of the white signal light beginning at noon each day during which five 1-g food pellets were delivered every 10 min. The instrumental conditioning procedure required the animals to maintain prespecified diastolic blood pressure levels in order to obtain food and avoid shock. Five 1-g pellets were delivered to the animal for every I0 rain of accumulated time that the diastolic pressure exceeded criterion, as indicated by the white light. Conversely, the animal received a single, 8 mA, 0.25 sec electrical shock through stainless-steel electrodes applied to a shaved portion of the animal's tail for every 60 sec of accumulated time that the diastolic pressure was below criterion, as indicated by the red light. Additionally, each food delivery reset the shock timer postponing the delivery of shocks for an additional 60 sec, and each occurrence of an electric shock reset the food timer, postponing the delivery of food for at least an additional 10 rain. At the beginning of the training period, the criterion pressure levels were determined by the animal's preconditioning, resting, baseline diastolic pressure (approximately 75 mmHg). Increases in the criterion occurred at a gradual, progressive rate of about 1-2 mmHg per week over a period of 2 to 3 months. Conditioning sessions occurred daily during a 12-hr period from noon to midnight ("conditioning on") alternating with a 12-hr period from midnight to noon during which no programmed contingencies were in effect ("conditioning off"). Within the 2 to 3 month interval of training,
130
Goldstein et ai.
the baboons attained diastolic blood pressure levels during the conditioning sessions which were significantly above baseline levels, and they maintained these elevated pressures for more than 95°70 of each 12-hr session. Baroreflex sensitivity was determined using the method of Smyth et al. (1969) by measuring the extent of increase in interpulse interval per unit of increase in systolic blood pressure after intravenous injection of phenylephrine, a peripheral vasoconstrictor. These determinations were made in the baseline and training periods before, during, and after the conditioning sessions (11 AM, 11:45 AM, 12:15 PM, 6 PM, 11:45 PM, 12:15 AM, and 1 AM). For each determination, phenylephrine (7.5 gg/kg) was infused into the venous catheter and flushed over a 10-second period with 5 cc isotonic saline at room temperature. This procedure consistently produced increases in both systolic and diastolic pressure of about 25 mmHg which lasted for about 3 rain. Approximately 20 heart beats after the initiation of each flush, systolic, and diastolic pressures began to rise to a maximum over about an eight-beat interval, associated with a decrease in heart rate (Fig. 1). With each beat starting from the beat where a change in systolic pressure was first detected and ending with the beat where systolic pressure peaked, an interpulse interval was associated--the interval for the succeeding beat (Pickering and Davies, 1973). Interpulse interval in milliseconds was calculated by dividing the beat-to-beat heart rate into 60,000. For each infusion, the relationship between the interpulse interval and systolic pressure (the "baroplot," Fig. 2) was determined over the approximately eight-beat interval of pressure increase, and the slope of the resulting linear plot (refecting significant positive correlations between interpulse interval and systolic pressure for more than 95°7o of the injections) defined the baroreflex sensitivity. Dropped beats (i.e., beats for which the interpulse interval exceeded by more than 80°7o the interval for the proceding beat) were not included in the data analysis.
! HnT,.
- ~...~ ~
FLUSH
~
~
~..,,e'~,.-., --t-
CHARLI| 4-16-7S 11411PM PHI~IIYLEPI41tlNIE 1S0jU~*
Fig. 1. Polygraph recording showing the relationship between blood pressure and heart rate changes followingphenylephrine infusionin baboon "Charlie."
Baroreflex Sensitivity and Blood Pressure Conditioning 700
-
131
-
650-vt
Y-" 3.92 X - 39. 50 R: 0.95
~" 550
5O0
450
4001
1 [ J ..~__J-_ 130 140 150 160 170 SYSTOLIC PRESSURE, M M . OF HG
Fig. 2. The relationship between the interpuise interval and systolic pressure determined over the eight-beat interval of pressure increase shown in Figure t. The slope of the resulting linear plot defined the baroreflex sensitivity.
RESULTS
Baseline Baseline blood pressure and baroreflex sensitivities are shown in Table I. Analyses of variance conducted on these data showed no significant trends over the listed times for any of these variables. No significant differences were obtained between the values for these variables during the times corresponding to Conditioning Off (11 AM, 11:45 AM, 12:15 AM, and 1 AM) and the times corresponding to Conditioning On (12:15 PM, 6 PM, and 11:45 PM).
Training Table II shows the 40-min averaged systolic and diastolic blood pressures and baroreflex sensitivities during training. All the subjects showed clear-cut, consistent elevations in systolic and diastolic pressure during Conditioning On as opposed to Conditioning Off (t = 8.27, df = 5,
132
Goldstein et al.
o
~
~: ~
~
~
~
~
~
o
E E
=
lo.
..
,~-~ Jl
II °
Baroreflex Sensitivity and Blood Pressure Conditioning
133
.=,
•
=
~
~
~
~
?
N •
N
~
~
~
~
~.
~
t~ -
~'~
o,t
N ,~
o
0
&
E~ .p.
