PSYCHOPHYSIOLOGYCopyright © 1975 by The Society for Psychophysiological Research

Vol. 12. No. 3Printed in U.S.A.

The Cold Pressor Test and AutonomicFunction: A Review and Integration

WILLIAM LOVALLOUniversity of Oklahoma Health Sciences Center

ABSTRACT

The purpose of the present paper is to review the literature on the cold pressor test and to discuss the waysin which these investigations have pointed toward a model of homeostatic function. The research empioyingthe coid pressor test includes studies of psychosomatic iiiness, factors influencing autonomic response tostress, the relationship between psychological disorder and autonomic responsivity, as well as studies ofcortical influences on autonomic function, and the relationship between stress and task performance. Fromthese studies, several parameters of a potential model of homeostasis are selected and discussed. Particularemphasis is placed on the role of autonomic balance and initial values in influencing autonomic response tostress.

DESCRIPTORS: Cold pressor, Homeostasis, Autonomic balance. Law of Initial Values, Stress.

This paper has two purposes. The first is topresent the bulk of physiological andpsychological research employing the coldpressor (CP) test and the second is to discusshomeostasis. The second purpose derivesfrom the first since the majority of CP studiesreviewed either raise questions best answera-ble by a theory of homeostasis, or otherwiseprovide indications of how such a theorymight be constructed. The intent of this ap-proach is to provide an adequate account ofCP research and a reasonable frameworkwithin which to view the experimental results.

Physiological Responses to CP

The CP test was first used as a means ofexperimentally increasing blood pressures instudies of hypertension (Hines & Brown,1932). The test consisted of a half-hourbaseline period during which several bloodpressure (BP) measurements were made, fol-lowed by a 1-min immersion of a hand or foot

This review was supported in part by (NINDS) Grant05797. The author wishes to express his thanks to Dr.Oscar A. Parsons whose support and advice made thecompletion of this paper possible. Thanks are also inorder to Dr. Robert Edelberg who made many helpfulsuggestions during the early phases of the manuscriptpreparation.

Address requests for reprints to: William Lovallo, De-partment of Psychiatry and Behavioral Sciences, Univer-sity of Oklahoma Health Sciences Center, P. O. Box26901, 800 N. E. 13th Street, Oklahoma City, Oklahoma73190.

in ice water during which BP was taken twice.The patient's BP was monitored until it re-turned to baseline after withdrawal (Hines &Brown, 1932). The normal CP effect was arise in systolic and diastolic BP of about 10 to20 mm Hg, and it was later discovered that theeffect depended upon intact innervation fromthe immersed extremity. When a CP test usingthe foot was given to a patient with total anes-thesia below the fifth lumbar dermatome noBP change was noted, while immersion of thehand led to a normal rise (Sullivan, 1941; seealso Pickering, 1958). This point is of interestsince vascular responses to warm immersionsare thought to depend on a change in circulat-ing blood temperature at the hypothalamuswhile response to CP seems to be controlledlargely by neural impulses from the periphery(Appenzeller, 1970). While the main source ofthe rise in BP seems to be peripheral vasocon-striction, the contribution of increased cardiacoutput is also significant. This arises as a re-sult of hypothalamic discharge along the in-ferior cardiac nerve (Shapiro, 1961).

A number of studies have examined localvascular responses to CP. During long im-mersions of the hand, the capillary beds of agiven finger tip initially constrict and thenperiodically dilate for brief periods (Lewis,1929). The wave-like vascular response,known as the Lewis Wave, indicates aperiodic weakening of sympathetic tone (Ap-penzeller, 1970, p. 49). Another term for the

268

May. 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 269

phenomenon is cold-induced vasodilation(CIVD). Cier (1938) studied vasomotor activ-ity in dogs following functional removal of thespinal cord. Shortly after immersion of a pawin ice water, the paw was found to regain itsnormal temperature. Recovery was abolishedfollowing sciatic nerve section. The responsewas thought to be controlled by a paraver-tebral ganglionic reflex mechanism in whichsensory stimulation leads to vasodilation viaantidromic stimulation by way of the grayrami (Gellhorn, 1943). Greenfield, Shepherd,and Whelan (1950) have found the CIVD tooccur in chronically denervated fingers. Theevidence thus suggests that dilation to coldmay also occur as a peripheral phenomenon.Engel (1959) examined the effects of three du-rations of CP immersion: 2, 3, and 4 min.Measurements were taken at rest and duringthe final 2 min of immersion of the foot. Sys-tolic and diastolic BP, heart rate (HR), andpulse volume were each affected differently atthe end of the varying immersion times. Skinconductance level (SCL) increased initiallybut returned to baseline rather rapidly as didrespiration and finger temperature. Thus,when the initial generalized arousal to CP isignored, the major response to long immer-sions appears in the cardiovascular system.

The preceding studies indicate that: The ini-tial CP response is cutaneous vasoconstric-tion, increased HR, and increased arte-rial pressure. These responses are primar-ily neurogenic reflexes depending on intactperipheral innervation. It would appear thatthe cold stimulation excites pain and tempera-ture fibers which enter the spinal cord in thedorsal roots and run rostral via the lateralspinothalamic, anterior spinothalamic, andspinotectal tracts. The first two tracts proceedto the thalamus and thence to the somestheticcortex. In the medulla, they send collaterals tothe reticuiar formation. Fibers from thespinotectal tract enter the tectum which is, inpart, responsible for the elaboration of reflexmechanisms. Thus, CP stimulation affectscortical, subcortical, and probably limbicstructures via the reticuiar formation and tec-tum. Additionally, peripheral factors inter-vene to some extent. The sympathetic activa-tion varies with immersion time producing thewaxing and waning of the Lewis Wave. Fi-nally, while vasoconstriction is noted in theskin, blood vessels in muscle tissue are knownto dilate (Appenzeller, 1970).

Visceral Response Components-StimulusSpecificity and Stereotypy

The study of psychosomatic disease hasbeen based partly on determining the or-ganismic and environmental factors which in-teract over time to produce somatic illness. Inthis connection Alexander (1957) spoke of twoprinciples: Individual-Response Specificityand Stimulus-Response Specificity and specu-lated on their roles in the etiology ofpsychosomatic disorders. A number of studiesemploying the CP have been conducted to ex-amine these concepts. One of the first was byLacey, Bateman, and Van Lehn (1952) whoexamined HR, HR variability, and SCL in re-sponse to mental arithmetic, word associa-tion, hyperventilation, and CP. The resultsindicated a greater-than-chance occurrence ofa maximal score on the same variable for allstressors for a given 5. This interstressor re-sponse specificity suggested that for manypeople, one autonomic response tends to pre-dominate regardless of the stressor at hand. Asimilar study by Lacey and Lacey (1958) usingfemale 5 s also suggested the occurrence of m-terstressor response stereotypy; the 5 s tendedto show idiosyncratic patterns of responseacross stressors.

Sternbach (1960) compared several phys-iological responses to startle, epinephrine,norepinephrine, CP, and exercise. The resultsshowed that startle and epinephrine led toqualitatively similar responses. Comparisonof published response magnitudes for CP andexercise shows that virtually all physiologicalsystems measured changed in the same direc-tion for both stimuli and differed by less thanthe means for startle and epinephrine. Thusboth CP and exercise appear to produce anearly identical arousal pattern. Wenger andCullen (1958) reported a study of physiologicalresponses to 14 stimuli. They found that thepattern of response was nearly the same formental arithmetic and word fluency tasks,somewhat similar for CP, and different formost of the remaining stimuli. Thus, the dif-ference between stimuli also seems importantin determining the response, which suggestssome validity for the concept of stimulus-response specificity.

Wenger, Clemens, Coleman, Cullen, andEngel (1961) replicated the Lacey et al. (1952)study using the same stressors but measur-ing a larger number of physiological re-

270 WILLIAM LOVALLO Vol. 12, No. 3

sponses. They found less evidence forindividual-response specificity than Lacey etal. (1952) and still less of a stereotyped patternof response across stressors. Again, theyfound that the characteristics of the stimulusinfluence the response. Thus, stimulusspecificity as well as individual-responsespecificity and response stereotypy enter intoproducing the pattern of visceral response.

Lacey and Lacey (1962) extended thisanalysis by showing that CP responses in sev-eral autonomic measures were reproducibleover a 4-yr interval. Children were given 2CPs in one test session in a given year and 2CPs during the same week 4 yrs later. Resultsshowed a somewhat better than chance oc-currence of the same pattern of autonomic re-sponses in both test sessions. The Laceys re-ferred to this effect as intrastressor responsestereotypy, which appears to apply over longtime periods and in spite of intervening mat-urational effects. Engel (1960) published re-sults which support both Wenger and Cullen(1958) and Lacey et al. (1952). He measuredsystolic and diastolic BP, respiration rate,SCL, HR, HR variability, and finger tempera-ture in response to an automobile horn sound-ing, mental arithmetic, decoding scrambledproverbs, exercise, and CP. The data re-vealed that the patterns of response were dif-ferent across the various stimuli, supportingthe concept of stimulus-response specificity.Furthermore, response patterns appeared tobe idiosyncratic to the individual 5 s indicatinga measure of individual-response specificity.These findings are supported by the results ofanother study using normal females (Engel &Bickford, 1961).

More recently, Wieland and Mefferd (1970)studied long-term variation in physiologicalactivity. Three 5s were tested daily for 120days, and a part of the procedure includedmeasurement of activity at rest, during CP,and during recovery. In attempting to evaluatethe variables influencing their long-termmeasures at any one point, they proposedconditioning, environmental influences, activ-ity, alerting, circadian rhythms, and noveltyas being important. The most striking findingwas that over time, intraindividual variationwas as great as interindividual variation in ac-tivity both at rest and during stress. At thispoint, it is difficult to assess the bearing ofthese results on the status of the concepts ofstimulus and response specificity and the au-thors do not address this issue. The impor-

tance of these studies is their attempt tocharacterize the pattern of autonomic re-sponses in terms of their various environment-al and organismic components.

In the process of analyzing the componentsof autonomic responses, Lacey, Kagan,Lacey, and Moss (1963) proposed four con-structs: Interstressor Stereotypy, the ten-dency to respond maximally with one systemregardless of the stressor; IntrastressorStereotypy, which means that for a given 5,the pattern of response tends to be similarwith repeated exposure to a stressor; StimulusStereotypy, which is that different stimuli tendto elicit different responses; and they pro-posed Symptom Stereotypy by whichpsychosomatic patients show maximum reac-tion in the autonomic response consistent withtheir somatic complaint.

CP Studies of Psychosomatic Disorders

The studies discussed below have attempt-ed to evaluate tendencies toward responsestereotypy in psychosomatic patients. Thisresearch partly focuses on the Lacey group'sfourth construct, symptom stereotypy.

Hines and Brown's (1932) study revealedthat subjects could be distinguished into threecategories (normals, hyperreactors, andhypertensives) on the basis of baseline and CPresponses in BP. Normals showed the lowestbasal and response BP levels, hypertensivesthe highest, with hyperreactors in between.These results were confirmed by additionalstudies (Hines & Brown, 1933, 1936, 1939;and Hines, 1940). A number of other investi-gations (Briggs & Oerting, 1933; Pickering &Kissen, 1936; Ayman & Goldshine, 1938;Yates & Wood, 1936; Shapiro, 1961) con-firmed or found results very similar to thoseobtained by Hines and Brown. On the basisof CP test results, Dieckmann and Michel(1935) were able to separate pregnant womeninto normals, hyperreactors, and hyperten-sives. Randall, Murray, and Mussey (1935)suggested the possibility of using the test todetect cases of hypertensive toxemia. In re-cording the CP responses of pregnant women,Bak (1937) noted that some hypertensive BPresponse curves showed an abnormal fluctua-tion above and below baseline, and a slow re-turn to normal. He also noted that in cases ofhypertensive toxemia, the response is often adrop (a paradoxical response) in BP followedby damped fluctuations about the baseline.

Hines (1937) administered CP tests to 400

May, 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 271

school children and found that a portion ofthose showed hyperreactive BP responses.Further investigation revealed that nearly allof the hyperreactive children had a hyperten-sive in the immediate family. These results areof interest in suggesting the value of long-termstudies of psychosomatic disease.

Eyster, Roth, and Kieriand (1952) studiedperipheral vascular activity in patients withatopic dermatitis and found BP responses toCP which were in the hyperreactive range. Nocardiac acceleration was noted. The authorsalso observed that this and other abnormalcardiovascular responses became more nor-mal on remission of the dermatitis. Wenger,Clemens, and Cullen (1962) tested thehypothesis that patients with differentpsychosomatic disorders show different pat-terns of autonomic response to stress. The pa-tients had either gastric ulcers, non-ulcerativegastrointestinal disorders, neurodermatitis, orother dermatological disorders. Rest datashowed that all patient groups were higherthan normal in sympathetic activity. In re-sponse to CP, the neurodermatitis and non-ulcerative patients showed greater HR in-creases and smaller rises in diastolic BP thannormals. In general, the patient groups did notdiffer from one another, but differed fromnormals in the ways mentioned. The results ofthese two experiments are in direct disagree-ment with respect to HR and BP responses.Wenger et al. (1962) used only male 5s andsuggested that the use of male and female 5 s inthe earlier study contributed to the discrep-ancy.

It appears that experiments in which re-sponse to the CP has been used to differen-tiate between patients with various psy-chosomatic disorders have met with lim-ited success. This may result from twocauses. The number of factors influencingsuch autonomic responses is large and theirinteractions are poorly understood. Also, theamount of effort expended in this direction hasnot been great enough to provide a systematicbreakdown of the relevant subject and en-vironmental parameters.

CP Studies on Psychological Disorders

A number of studies have evaluated theproposition that various psychological disor-ders are accompanied by differential re-sponses to cold immersion. White and Gildea(1973) compared CP responses in controls,hypertensives, and anxiety reactive 5s. The

hypertensives showed the highest BP re-sponse of any group, and the slowest return tobasal level. The responses of anxiety patientsfell between normals and hypertensives onHR and BP measures. Rosen (1951) comparedCP response of hysterics to those of severalother patient groups. The mean responsemeasures for the non-hysterical patientsshowed normal BP increases. Observations ofchanges from basal levels among hystericsshowed an average decrease of 22.0 mm Hgsystolic and an increase of 1.0 Hg diastolic forthe first min, and decreases of 47.0 mm Hgsystolic and 0.7 mm Hg diastolic for the sec-ond min. These paradoxical (i.e., responsechanges opposite to normal) results for hys-terics were interpreted by Rosen as due to adenial of expressions of anxiety, accompaniedby inhibition of vasomotor responses tostressors or painful stimuli such as the CP.The results of an earlier study by Reider(1938) tend to confirm Rosen's findings. Atwo-part study published by Igersheimer(1953) also compared CP responses amongseveral patient groups and attempted to de-termine the CNS level which controls the re-sponse to CP stimulation. Normals showedgreater BP increases than schizophrenics andmanic-depressives, and increases slightlyhigher than the neurotics. Some schizo-phrenics showed the paradoxical response. Ingeneral, the psychoneurotics showed CP re-sponses intermediate to those of the normalsand schizophrenics. His results were discuss-ed in terms of Gellhom's (1953, pp. 429-432)theory.

Wenger, Clemens, and Cullen (Note 1) re-ported results of a study comparing normalswith schizophrenics and non-psychotic pa-tients in a controlled resting state and in re-sponse to CP. During rest, the schizophrenicsshowed BPs which were not different fromthose of other psychiatric patients. Their au-tonomic balance (A) scores in fact showed apredominance of sympathetic activity as com-pared to normals. In response to CP, schizo-phrenics failed to differ from the other patientsin the study. In general, these results fail tosupport Igersheimer's (1953) findings.Changes in drug therapy for schizophrenics,especially the introduction of the use of chlor-proniazine in the mid 195O's, may be asignificant variable underlying the differencesin these results.

Clearly, disorders of a functional psy-chological nature are accompanied by related

272 WILLIAM LOVALLO Vol. 12, No. 3

changes in physiological function, althoughvarious forms of psychological disturbance donot appear to be associated with unique pat-terns of CP response. An explanation of theseresults would seem to require some account ofaltered perception of the stimulus, of alteredsymbolic meaning of the stimulus, and a shiftin the responsivity of the ANS. That thesepossibilities exist will be discussed in connec-tion with pain.

The Role of Central Processes inControlling Responses to CP

A number of papers discussed have sug-gested that cortical processes can modifyphysiological responses and pain sensations toCP stimulation.

Cortical Studies. Few articles have dealt di-rectly with the involvement of the cortex inCP responses. The earliest paper that hascome to the attention of the reviewer is that byRizzolo (1929). Dogs' foreiimbs were placedin ice water for 3 to 10 min, and cortical ex-citability was measured pre- and post-immersion at points on the motor cortex. Thecortical motor point of the right forelimb in-creased in excitability with immersion timesup to 3 min and then decreased. The effectwas less pronounced at other motor points. Itwould be interesting to examine the possibilityof parallel courses in the habituation of thecortical excitability effect and the peripheralvasoconstrictive response. Nakao, Ballin, andGellhorn (1956) showed that decreases in BPled to increased activity in the hypothalamusand on the cortex and that increased BP led tothe opposite change in these measures. Thecentral inhibitory role of the aortic and carotidbaroceptor afferents will be discussed below,but it is apparent that these influences can di-rectly or indirectly affect cortical activity.

Igersheimer (1953) also investigated the roleof the cortex in the CP response mechanism.In order to determine the CNS level mediatingthe CP response, 15 medical students weregiven a CP test under deep levels of anes-thesia, during which the cortex is depressedand presumably any remaining peripheral re-sponse would be subcortically controlled. Atsuch degrees of anesthesia, a response decre-ment of 26% was observed, indicating a sub-stantial degree of cortical control, assumingno anesthesia effects on brainstem structures.The overall results of the Igersheimer studieswere interpreted as indicating that the CP re-sponse is controlled largely by subcorticalstructures, particularly the hypothalamus and

medulla, and is based on a neurogenic reflex.A more recent study by Griffin (1963) sug-

gests that neocortical areas influence the mag-nitude of the CP response. Comparisons weremade between habituation of CP responses inleucotomized schizophrenic patients, nor-mals, and non-leucotomized schizophrenics.Leucotomized patients showed higher basalBP levels and required significantly more im-mersions before habituating than did the nor-mals or schizophrenics. Other work on CPhabituation is reported by Glaser and Whitlow(1957), Glaser, Hall, and Whitlow (1959), andGreenwood and Lewis (1959). Sterman,Clemens, and Wenger (1966) noted that uponCP immersion, EEG alpha rhythm (8-12 Hz)was blocked but that for some 5 s alpha recov-ered during the immersion period, while forothers recovery did not occur until after with-drawal from the water. In the "rapid recov-ery" group, HR, BP, and log palmar conduct-ance showed more rapid recovery during im-mersion than for the "slow recovery" 5s.However, rest levels of these measures werenot so simply related to alpha recovery la-tency or to recovery trends. According to theauthors, "it is likely" that some active centralprocess leads to response reversals duringimmersion. They cite studies indicating thatstimuli acting on mesencephalic and hy-pothalamic structures which produce ex-citatory responses also activate bulbar struc-tures which tend to counteract such re-sponses. Forebrain efferents to the lowerstructures mediate cerebral regulation of theseareas and also allow for the expression of theeffects of prior conditioning (Clemente, Ster-man, & Wyrwicka, 1963). Such cortical ef-fects are interesting in light of Rizzolo's (1929)finding of a lowered excitability threshold ofthe motor point of the immersed limb withlong immersion times, and the results of Grif-fin (1963) emphasizing the role the frontal cor-tex plays in the establishment of habituation ofresponse to CP. Lovallo, Parsons, and Hol-loway (1973) found that the peripheral vas-omotor response to and recovery from CP isaltered in patients who have suffered corticallesions.

In general, then, effects of CP stimulationcan be modified by inhibitory action of thefrontal-basal forebrain and bulbar brainstemstructures. The cortex may also affect the CPresponse more indirectly via memory effectsrelated to affective associations to the painfulstimulation. These studies are discussed in thenext section.

May. 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 273

Pain and Pain Modification. A review ofstudies of cold stress indicates that cold-immersion effects cannot be understood with-out considering affective states such as pain.Appenzeller (1970) points out that the re-sponse to CP is partly due to the pain inducedby the ice water. Edes and Dallenbach (1936)took subjective reports by 5 s which indicatedthat the sensation passed through severalstages from a feeling of cold, increased cold,pain, to increased pain, which leveled off, de-creased pain, and then cold alone. This evi-dence indicated that the sensations of cold andpain were mediated by separate classes offibers since the cold sensation was presentthroughout while pain had a later onset andearlier disappearance. Wolf and Hardy (1941)reported a pain study which provided reportsof the progression of pain states occurringwith various water temperatures. It was foundthat BP increased, leveled off, and decreasedin direct proportion ofthe reported pain, andso the BP response seems to be associatedwith the pain rather than the cold sensationper se (see also Wolf & Hardy, 1942). Impor-tantly, the two sensations (cold and pain) canbe perceived and independent and the percep-tion of pain is directly correlated with changesin peripheral response. This view is supportedby Hilgard (1969).

Some evidence suggests that the reportedsensation of cold-pain may be modified.Barber and Hahn (1962 and 1964) tested theeffects of hypnosis on pain reports and onphysiological response magnitude. Followingpain suggestion, a hypnotized group reportedpain to immersion in room temperature water.Their pain-state reports were the same asthose of a waking control group given a CPindicating that hypnosis is capable of modify-ing subjective pain states. The BP measure-ments were suggestive of a hypnosis-inducedBP increase, but were nonconclusive. Hil-gard, Cooper, Lenox, Morgan, and Voe-vodsky (1967) measured pain states using theWolf and Hardy (1941) rating scale. 5s in asuggested analgesia group had significantlylower pain states to CP than control 5s. TheHR measurements also indicated that hyp-nosis was responsible for smaller responses inthe analgesia-hypnosis group. The results ofboth of these studies indicate that the experi-ence of pain and the related autonomic arousalmay be subject to hypnotic control. Doupe,Miller, and Keller (1939) studied the hypnosiseffect on vasomotor responses to both stimuli.In addition, they showed a lack of affective

reaction (squirming, grimacing) to the stress.This partial response decrement led the au-thors to conclude that suggested analgesiafailed to block the physical representation ofthe stimulus but reduced the affective reaction(pain).

The hypnosis studies indicate that vas-omotor responses to stress can be modified. Itis reasonable to assume that suggestion ismediated cortically. In this manner, corticalinfluences may be able to modify the mag-nitude of the CP response. Considered withthe other studies cited, these results suggestthat two mechanisms mediate the CP re-sponse. The first part ofthe response is due tothe physical cold stimulus per se and is mostlikely controlled subcortically. The secondphase ofthe response is due to the cold pain.The pain state, having arousing and negativeaffective properties, accounts for a portion ofthe sympathetic arousal and may act corticallyand subcortically. Hypnosis and other ver-bally mediated stimuli appear able to affect thecortically mediated, pain-related portion ofthe response. Murphy and Gellhorn (1945)demonstrated \Videspread interaction betweencortex and the hypothalamus. Given such ex-tensive cortical, limbic, and diencephalic in-terrelationships, cortical effects on the re-sponse to noxious stimulation may besignificant.

The Law of Initial Values and thePrinciple of Autonomic Balance in

Models of HonieostasisAs a means of conceptualizing the above

research, this section will first review multiplestressor studies, then discuss the Law of Ini-tial Values (LIV), and then present researchdemonstrating a relationship between CP re-sponse and task performance under stress.Finally, Teichner's (1968) approach to theproblem of initial values will be presented fol-lowed by Wenger's (1948) and Gellhorn's(1968a, 1968b, 1969, and 1970) systems ofcharacterizing autonomic balance.

Multiple Stressor Studies

Studies using more than one stressor haveproduced results of interest in studying theLIV (Wilder, 1957). An understanding of re-sponses to multiple stimuli is essential sinceenvironmental stimuli never occur in isolationbut usually appear in complex patterns (Gold-berg, 1966). Blitz and Dinnerstein (1968) per-formed experiments which demonstrate thatthe CP effect (an immersion of one hand)

274 WILLIAM LOVALLO Vol. 12, No. 3

could be blocked by prior, contralateral CPstimulation but the authors did not suggest anyexplanation for the effect. Blitz and Dinners-tein suggest that subcortical effects may re-duce afferent pain stimulation and thus alsoreduce sympathetic activity during the secondimmersion. This finding seems interesting anddeserves to be replicated.

Teichner (1965) using multiple stressorsstudied the relationship between CIVD la-tency and emotional stress. It is known thatCIVD latency normally decreases upon eachsuccessive immersion. All subjects were given1 CP test per day for 4 days plus no-shock,weak-shock, or strong-shock on days 2 and 3.The strong-shock group showed an increase inCIVD latency on days 2 and 3 and also on day4 after being told that no shock would begiven. The weak-shock group increased theirlatency on days 2 and 3, but the latencydropped on day 4. The no-shock groupsshowed the expected continued decline inCIVD latency. Assuming that anticipatoryemotional arousal occurred in the strong-shock group on day 4, it can be concluded thatthe emotional reaction increased the CIVDlatency. This in effect suggests increased SNSactivation which is mediated cortically.

Mefferd and Wieland (1965) tested thehypothesis that cognitive activity could mod-ify physiological response. The dependentvariable was electrodermal activity, whichpeaks about 40 sec following CP immersionbut which reaches a maximum just after thebeginning of cognitive tasks. To overlap themaximum sympathetic response to bothstimuli, Mefferd and Wieland began a wordassociation task 30 sec after immersion. Theresults were that the galvanic skin response(GSR) during immersion dropped im-mediately upon the start of the word associa-tion to levels no greater than word associa-tions alone. These results were confirmed inlater experiments using: passive exposure to apassage being read plus CP (Sadler, Wieland,Mefferd, Benton, & McDaniel, 1967) and areaction time task plus CP (McDaniel, Mef-ferd, Wieland, Sadler, & Benton, 1968) allwith the same paradigm as above. The resultsof these experiments were attributed to theeffect of cognitive activity distracting the sub-jets' attention from the pain stimulus, sincecognitive activity was ongoing when cold-painreached a maximum at 40 sec.

Sadler, Mefferd, Wieland, Benton, and

McDaniel (1969) tested the conclusion thatcognitive activity alone could distract subjectsfrom their pain state and reduce sympatheticactivation. Subjects began a passive listeningtask prior to CP thus separating maximalarousal to the two. The ongoing cognitive ac-tivity failed to reduce activation to CP. Theauthors concluded that cognitive activityalone did not reduce activation and suggestedthat results of past studies were due to subcor-tical influences. Without underestimating theimportance of brainstem mechanisms, an al-temative explanation is possible. In the earlierstudies, the superimposition of maximalarousal to CP and the onset of a cognitive taskshould have produced an orienting response tothe task onset. This could be expected to pro-duce a narrowed attentional field and thus lessattention to other afferent stimuli, and a shiftin attention from the cold-pain. Attentionalmechanisms may in fact be important here,not because of cognitive activity, but becausethe subject was required to reorient himself tothe beginning of a new task thus bringing at-tentional mechanisms into play.

In addition, the reading or listening tasksemployed did not require active, overt re-sponses of the subject, and also required lessinformation transformation than a word as-sociation task does. This would render theSadler et al. tasks less distracting than anorienting response to the onset of a task.Germana (1968 and 1969) clearly distinguishesbetween cortical-autonomic effects in caseswhere the task is passive and where it requiresformulation of overt responses. The ongoingcognitive activity 40 sec after the initiation ofthe reading task could not have had as strongan attentional effect as the orienting responseto the onset of a word association task, asused by Mefferd and Wieland (1965).

It should be clear from the above studiesthat sympathetic arousal is not simply addi-tive, but the arousal due to two or morestimuli interacts cortically and/or subcorti-cally and can modify output to the final com-mon pathway. In any event, the possibilitythat cortically mediated activity is importantin these results cannot be ruled out. A remain-ing problem is the development of a system ofcharacterizing autonomic response which canencompass results of studies showing de-creased response to multiple stressors (e.g.,Mefferd & Wieland, 1965) and those showingan increase (e.g., Teichner, 1965).

May. 1975 THE COLD PRESSOR TEST AND HOMEOSTAStS 275

The LIVWilder (1957) discussed what he called the

LIV:Not only the intensity but also the direction of a responseof a body function to any agent depend to a large degreeon the initial level of that function at the start ofthe exper-iment. The higher this initial level, the smaller is the re-sponse to function-raising, the greater is the response tofunction-depressing agents. At more extreme initiallevels, there is a progressive tendency to no response andto paradoxical reactions, i.e., a reversal ofthe usual direc-tion ofthe response.

Lacey and Lacey (1962) tested the LIV bymeasuring response to CP in three ways: ab-solute change, change expressed as the per-centage of the pre-stimulation level, andchange measured by the T score. The latter asa measure of response is based on the stand-ard deviation of attained stress level from theexpected stress level derived by a regressionanalysis. They found support for the LIV andstated that of the three response measures,data transformed to T scores were most effec-tive in demonstrating the effect of initial val-ues on response values.

Relationships Between CP Responses andTask Performance. In a second experiment inTeichner's 1965 paper, subjects were classedas fast and slow dilators based on their CIVDlatencies. All subjects were then given achoice response task which became increas-ingly conflict arousing (and, hence, sym-pathetically activating) over trials. If the slowdilators are considered high in sympathetictonus, the arousal due to conflict allowedthem to improve their performance initially,until overall arousal reached non-optimallevels and their performance declined. If thefast dilators can be considered low in chroniclevels of sympathetic tonus, the conflict-induced arousal allowed them to improve theirperformance continuously within the range ofstress provided. The CIVD latencies of eachgroup predicted behavior under stress in amanner something like the inverted U func-tion. Thus, autonomic activity may be relatedto performance on certain tasks when stresslevels are defined. Teichner (1966) tested thehypothesis that CIVD latency could predictbehavior on a task which was sensitive toarousal levels and attention. Subjects wereseparated into short- and long-latency dilatorsand tested in either a cold room (55°F) orwarm room (80°F) on a monitoring task. Fastdilators showed relatively better target detec-tion performance at 55° than at 80°, while slow

dilators showed the reverse. Again, these re-sults show that cold immersion responsecharacteristics are related to arousal levelsand that when performance is studied in astressful situation, autonomic arousal in-teracts with attention and performance.Lovallo et al. (1973) and Lovallo and Zeiner(Note 2) have studied peripheral vasomotorresponses to CP. In the former paper, it wasnoted that some 5s appeared to show in-creased finger volume to CP while othersshowed a decrease. The latter experimentsevaluated the possibility that the direction ofchange in digital volume was related to somemeasure of peripheral vasomotor activity atrest. It was found that 5s who showed de-creased digital volume to CP had large pulsewave amplitudes at rest while 5s who showedincreased finger volume had small pulse waveamplitudes at rest. It thus appeared that de-gree of peripheral vasoconstriction at rest wasrelated to direction of change in finger volumeto CP. If so, these results demonstrate theeffect of initial values on the direction ofresponse." Gellhorn and Louffbourrow (1963)summarize experiments which have shownclearly that the response to stress is a functionof the stimulus and, more importantly, ongo-ing levels of autonomic activity. Teichner andLevine (1968) presented a study in which fastand slow dilators were classically conditionedm rooms at either 50° or 90°F. Conditioningprocedures included a light paired with shock.The conditional response was vasomotor ac-tivity. It is well known that electric shock is asympathetically arousing stimulus and as suchelicits vasoconstriction in control conditions.However, subjects conditioned under 50°showed vasodilation to the shock and later asa conditional response to the light alone. The90° group showed the normal constrictive re-sponse. The CR was in fact a change in ongo-ing level of activity, rather than a unidirec-tional response, the relevant parameter beingbackground arousal level.

'Data from our laboratory (Lovallo & Zeiner, Note 3)have shown that within a given subject, the direction ofvasomotor response can be manipulated. Each S wasgiven 3 CP tests on 3 separate days with the room temper-ature at 12°C, 22°C, and 32°C. Under the warm tempera-ture, mitial pulse wave amplitude was large and CP re-sponse was a drop in blood volume level. When the roomtemperature was cold the initial pulse wave amplitude wassmall and in response to the CP, blood volume levelsinc reased or failed to change. When the 5s were tested at22°C, intermediate rest levels and responses were ob-served.

276 WILLIAM LOVALLO Vol. 12, No. 3

Scherrer (1959 and 1962) demonstrated theeffect of basal BP on BP response tohypothalamic stimulation. Low initial BP, fol-lowed by hypothalamic stimulation leads toincreased BP and increased firing of thesplanchnic and renal nerves. High initial BP,followed by hypothalamic stimulation leads toreduction in these measures. Long-durationstimulation may elicit an increased BP fol-lowed by a decrease during stimulation. Theform of this response is suggestive of theLewis Wave discussed earlier. He concludedthat post excitatory inhibition can followspontaneous or elicited sympathetic activityand that this inhibition of ongoing activity iscentral in origin and is inherent in central au-tonomic function. This research stems fromthe earlier work of Bronk, Pitts, and Larrabee(1940). A possible neurophysiological descrip-tion of the effect of initial levels on cardiovas-cular responses to CP would be as follows.Immersion of an extremity in ice water leadsto afferent stimulation of the spinothalamicand spinotectal tracts which ascend via themedulla and send collaterals to the reticularneurons which are in intimate association withcardiovascular regulating areas in the medulla.Their excitation causes increased cardiac out-put and vasconstriction in the skin and henceincreased BP. The baroceptors in turn in-crease the rate of afferent firing via buffernerves to the same medullary areas in order toinhibit the ongoing responses. This negativefeedback system thus serves to maintainhomeostasis. If in a normal organism, levels ofBP are high prior to CP stimulation, afferentfeedback via baroceptors will serve to provideinitially high levels of inhibitory activity.Stimuli which tend to further increase BP willthen cause greater inhibitory than excitatorytendencies and the response will show little ifany rise in BP or perhaps a sudden drop (re-versal). Lacey (1956) in discussing the LIVwrote:. . . the more powerful influence of parasympathetic in-hibitory mechanisms, and their potentiation bysympathetic excitation, suggest the following interpreta-tion ofthe dynamics of a cardiovascular response. As thepre-stimulus level of functioning increases, there is a dis-proportionately greater homeostatic restraint, both in in-creased magnitude and decreased latency and, as themagnitude of induced activation increases, there is a dis-proportionately greater increment in counter reaction.

Stress Response, Performance, and theLIV. The idea that ongoing levels of au-tonomic activity change due to stress, and thatthe direction of the change is a function of

ongoing levels of activity provides the basisfor Teichner's (1968) model of stress re-sponse. The model also takes account of fluc-tuations of attention in response to varyingrates of information input (seen as stressors)from the environment. The information-loadparameters are under control of mechanismsparallel with, and similar to, those regulatingphysical stressors. The essential feature ofthis system is the relationship between en-vironmental stimuli (inputs) and states ofcompensatory reactions — these includestressful inputs and stress responses, respec-tively. Environmental stimuli (whether physi-cal or symbolic) change over time. Thus, theyhave a rate-of-change parameter. The rate ofchange could be conceived of as rates ofchange of firing in afferent neural inputs.Compensatory responses, or outputs, tostimuli at the CNS or peripheral level have asimilar rate of change which can be expressedas the rate of change of efferent neural dis-charge. The rate of change is viewed here asthe critical response parameter. Under normalenvironmental stimulation and normal or-ganismic states, the rate of change of input inratio with the rate of change of the output isreferred to as the Input rate/Output rate ratio.According to Teichner (1968) this ratio hassome normal range of values within which thechange in input is compensated for by thechange in output. For example, a drop of 3°Fin environmental temperature (a change ininput rate) would be matched by some outputphenomena represented peripherally by mildvasoconstriction, reduced sweating, andperhaps some increase in muscle tension.When the input and output are nearly equal,the ratio has a value somewhere near one.When the rate of change in input changesmore rapidly than output phenomena are ableto cope with (i.e., match with a similarchange), the value of the ratio exceeds someoptimal value and a stress reaction occurs.

Since the physiological systems in questionare assumed to operate within certain limits ofthe ratio, the output must maintain a changerate similar to that ofthe input. To do so underrapid rates of input change, reversal reactionsmay occur. Thus when a hand is placed in icewater, the change in output is a vasoconstric-tion and increased BP. At some point, thevasoconstrictive output cannot change its rate(e.g., under maximal vasoconstriction). Soduring long immersions, the change in outputtoward vasoconstriction fails to match the rate

May. 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 277

of input change. At this point, a reversal reac-tion occurs, accompanied by vasodilation anddecreased BP. The rate of change of outputcan again increase to a level which can matchthe rate of the input and the value of the ratiodecreases to within normal limits. It should beremembered that in this model the rate ofchange is more important than the direction ofchange.^

Another concept of importance is atten-tional bandwidth, which refers to the propor-tion of incoming stimuli (inputs) which are at-tended to. During information overload (astressor) the attentional bandwidth narrows toeliminate unnecessary stimuli and decreaserates of input change. Bandwidth changes aresaid by Teichner (1968) to be directly relatedto the amount of cortical projection area de-voted to processing inputs. Attentional band-width could have implications for understand-ing the effect of attention and pain in influenc-ing CP responses. This concept is also di-rectly applicable to Teichner's (1965 and 1966)results in which physiological response in-teracted with performance.

Initial level of physiological activity is ap-parently important in determining magnitudeand direction of a given response. This con-sideration is certainly important for an under-standing of responses to multiple stimulationsuch as that employed by Mefferd and Wie-land (1965). The interesting results of thestudies by Teichner (1965 and 1966) andTeichner and Levine (1968) suggesting the ef-fects of ongoing level of response on task per-formance point to the importance of under-standing the relationships and mechanisms in-volved. The fact that many hypertensiveshave not only high resting BP but also showabnormally large BP increases to CP perhaps

In connection with Teichner's concept on the rate ofchange ofthe output matching that ofthe input, the sameconcept exists in psychophysics. Consider that the per-ceived intensity of a stimulus is the output (the percept)and that the input is the afferent neural firing rate from thereceptor. When inputs are near normal environmentallevels, the output matches the input quite closely. This isexpressed another way in that within normal environment-al levels of stimulation j.n.d.'s are small. This must be sowhen the change in input is matched closely by the changein output. At levels of stimulation increasingly distantfrom normal levels, j.n.d.'s become larger. That is theinput change is marked less well by the change in output.When the change in input can no longer be matched bychanges in output, stress reactions occur. This psycho-physical concept seems closely parallel to Teichner'sconcept of response to physiological and symbolic stimuliunder normal and stressful levels of stimulation.

indicates some fundamental change inmechanisms controlling the relationship be-tween initial levels and final response. Morestudies ofthe LIV are needed to better under-stand stimulus-response relationships for var-ious physiological systems.

Autonomic BalanceAnother determinant of physiological re-

sponsivity is autonomic balance. The termsvagotonia and sympathicotonia (Eppinger &Hess, 1917) were originally used to charac-terize individuals who were relatively para-sympathetic or sympathetic in responseto stimulation. Various attempts have beenmade to systematize such differences in re-sponse tendency as a means to better under-stand normal and abnormal states. Followingearlier work with children (c.f., Wenger &Cullen, 1972), Wenger (1948) described amethod by which a profile of resting au-tonomic activity was obtained from a largegroup of normal male 5s. A number of au-tonomic variables were measured and dias-tolic BP, log change in palmar skin conduct-ance, salivary output, volar forearm skinconductance, palmar skin conductance level,oral temperature, and heart period were foundto be rnost useful in describing a factor of au-tonomic function. These measures were thenemployed to develop a score of autonomicbalance (an estimate of the autonomic factor)which was derived separately for each 5. Thedistribution of these scores was symmetricalabout a mean of 70 and had a standard devia-tion of 8. This system has allowed individualsto be tested at rest and their autonomic bal-ance (or imbalance) determined by compari-son with the normative data. In this way, therelative parasympathetic or sympatheticdominance of a given individual can beknown. This line of research has provided animportant tool for evaluating some changes inautonomic function accompanying psychoso-matic disorders and other states. The mostthorough review of this extensive line of re-search is given by Wenger and Cullen (1972).

Gellhorn (e.g., 1970) has attempted to con-ceptualize autonomic balance in a somewhatdifferent way, but again with the purpose ofunderstanding normal and abnormal states ofsystemic physiology. According to Gellhorn(1968a, 1968b, 1969, and 1970), a stimuluswhich excites the sympathetic division of theANS will have certain peripheral and centraleffects. Peripheral effects would be increased

278 WILLIAM LOVALLO Vol, 12, No. 3

HR, vasoconstriction, increased BP, pupil-lary dilation, increased skin conductance andincreased somatic activity. Centrally, corticaldesynchronization is noted in the EEG. Acti-vation of the parasympathetic system wouldresult in the opposite changes in most of thesemeasures. Activation ofthe SNS and relatedsomatic structures is referred to as ergotropicactivity. In the same way, activation of thePNS and its accompanying effects is referredto as trophotropic activity. In general, activityofthe ergotropic system is associated with re-sponses which are action-oriented. Tropho-tropic activity is associated with non-action-oriented behaviors.

An important concept in Gellhorn's systemis that of tuning which refers to the ability ofthe ergotropic or trophotropic system to re-spond to a given stimulus. In normals, ergo-tropic and trophotropic tendencies are bal-anced so that the appropriate autonomic re-sponse occurs to a given stimulus. When thebalance of tuning shifts, the reciprocal rela-tionship is less effective and the organism ismore likely to respond according to its state oftuning. An ergotropically tuned systeni istherefore more sympathetically responsive,and less likely to show appropriate parasym-pathetic rebound. When imbalance is severe,the LIV may fail to hold. For example, inhypothalamic stimulation studies (Gellhorn,1959), application of a sympathetic stimultjsduring rebound to a prior sympatheticstimulus led to paradoxical responses charac-teristic of decreased sympathetic activation,indicating an induced shift in the balance oftuning and consequent failure of the LIV.Most studies of CP effects in hypertensionhave found high resting BP and abnormallylarge responses. The LIV would predict thathypertensives, having higher initial values,would show smaller BP increases than nor-mals. The Hines and Brown's studies indicatehypertensives have larger responses. An ex-ception is indicated in the study by Bak (1937)who found paradoxical BP responses to CP incases of hypertensive toxemia. This serves toindicate the need for additional work on theinteraction of autonomic balance and initialvalues in determining physiological responses.Gellhorn's (1970) system seems valuable as aninitial attempt to assess the role of changes inautonomic balance in understanding variousstates of disorder.

The above discussion is an attempt to inte-grate the efforts of Wilder (1957) and Teichner

(1968) on the LIV and those of Wenger et al.(1962 and Note 1) and Gellhorn (1970) on au-tonomic balance. To summarize the collectivecontributions of these workers: three impor-tant parameters which determine stress re-sponses are dynamic range ofthe system, itsset point, and the initial value of functionwhen the stress is applied. The dynamic rangeis simply the range of values over which theautonomic system in question can efficientlyoperate. Set point is analogous to the settingof a thermostat. When the operating level ofthe system exceeds the value tolerable by theset point, countervailing tendencies arebrought to play. This may lead to a reversal.The work of Gellhorn and Wenger aboveseems most relevant to this construct. Third,given a dynamic range and set point, the initiallevel of function from which change occurs isimportant. When the initial value is near theset point and the function moves toward thelimit ofthe dynamic range, a reversal is morelikely than if the function moves toward theset point. The work of Wilder, Lacey, andTeichner seems most relevant to this aspect ofautonomic function.

Conclusions

While no unified model or theory yet existsto systematically account for the results dis-cussed here, reasonable progress has beenmade in outlining the requirements or featuresthat such a theory might contain. A theory ofhomeostasis would need to account for certainrelevant variables. First, the nature of thestressor is critical. That is, the response likelyto be elicited by the stimulus should be under-stood in terms of the number of systems itaffects and the degree to which each may beactivated. The work of Engel (1960) onstimulus-response specificity and ofthe Laceygroup (Lacey et al., 1963, and Lacey, 1967) aswell as Wenger and Cullen (1958) on responsepatterns provide a basis for further research inthis area. A second variable is individual dif-ferences in autonomic response tendencywhich need to be specified to determine thekinds of response pattern differences notedbetween 5s. The papers by Lacey and Lacey(1958) on response-stereotypy, and by Engel(I960; Engel & Bickford, 1961) on individual-response specificity as well as those ofWenger et al. (1961, 1962, and Note 1) areforays in this direction. Also relevant to theissue of individual differences is Gellhorn's(1968a, 1968b, and 1970) system and the

May, 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 279

autonomic balance measurement of Wenger(Wenger & Cullen, 1972).

A third and much-related factor is the mean-ing the stimulus has for the organism. For ex-ample, stimuli, regardless of intensity, whichhave rapid onset can elicit flight reactions inanimals and startle reactions in humans. Cog-nitive factors, learning, and expectancies areimportant here. The works discussed aboveon pain and its modification and the influenceof cortically-mediated effects on CP responseare relevant to this issue. Fourth, an ultimatemodel of homeostasis must incorporate theLIV. The importance of the LIV is indicatedin Lacey and Lacey (1962) and in Teichner's(1968) system. It is not suggested here thatthese factors are independent. Instead, theymost likely interact in rather complex ways.Neither is it suggested that this list is exhaus-tive, but it is intended to emphasize some ofthe major themes represented in researchusing the CP. These would appear to be someof the pertinent variables participating in andinfluencing homeostasis at the systemic level.

The foregoing discussion, while far from in-tegrating the results of the clinical observa-tions and experiments employing the CP test,may hopefully provide an understanding of afew of the important themes arising fromstudies of systemic physiology. The emphasisplaced on the LIV and autonomic balance isrelated to three important issues. The firstissue is the nature of the servomechanism un-derlying homeostasis and its dynamic charac-teristics. The second is the way in which asupposed servomechanism is set and how it canbe reset. Normally, the mechanism is bal-anced, but studies of disordered states suggestthat at times and in given individuals themechanism becomes differentially sensitive,i.e., the set point is changed. A third questionemerging from the first two is, how do the LIVand varying degrees of autonomic imbalanceinteract? Additional knowledge of the ways inwhich cortical, limbic, and diencephalicmechanisms interact and act upon mesence-phalic mechanisms would assist greatly inanswering the above questions.

SummaryClinical observations and experiments using

the CP as a stressor have proven valuable inthe advancement of psychophysiology. This is

particularly true in evaluation of autonomicactivity in normal and abnormal states. Thephysiological effects of the CP are discussedincluding early cardiovascular studies ofhypertension, studies of mechanisms mediat-ing response to CP, and generalized responsesto stressors, including CP studies which at-tempted to define the parameters of autonomicresponses. It has been suggested that both in-dividual differences in response to stress and1;he characteristics of a given stimulus interactin determining autonomic responses. A dis-cussion of these factors in psychosomatic dis-ease is presented including a study of theetiology of hypertension and an analysis of re-sponse stereotypy. Along similar lines,studies of autonomic responsivity in psychiat-ric disorders are discussed.

The role of the cerebral hemispheres inmediating the response to CP and the effectsof CP on these structures is considered. Theseinclude studies of cortical excitation, the roleof the cortex and forebrain in mediating in-hibitory tendencies, and studies of pain and itsmodification. A number of studies are re-viewed in which CP has been administered inconjunction with other stimuli to determinethe effect of multiple stimulation. The impor-tance of both cortical and subcortical influ-ences is emphasized in analyzing the effects ofmultiple stimuli. Three studies are presentedin which the CIVD latency during CP is usedas an independent variable to predict perform-ance in stressed and unstressed conditions.

As a means of elucidating principles impor-tant in understanding homeostasis and its dis-orders, discussion of the LIV and of au-tonomic balance are presented. The LIV isviewed as important to evaluation of mag-nitude and direction of response, to under-standing multiple stressor effects, and to in-teraction of base levels of activity and taskperformance under stress. The concept of au-tonomic balance is utilized in characterizingthe propensity of a system to respond either ina sympathetic or parasympathetic direction.The interaction of initial values and autonomicbalance is emphasized.

In light of the lack of a comprehensivemodel of homeostasis with which to integratethis literature, an attempt is made to delineatethe basic requirements of such a model.

280 WtLLIAM LOVALLO Vol, 12, No, 3

REFERENCESAlexander, F, Psychosomatic medicine, its principles and

application. New York: Norton and Company, 1957,Appenzeller, O, The autonomic nervous system. New

York: American Elsevier, 1970,Ayman, D,, & Goldshine, A, D, Cold as a standard

stimulus of blood pressure: A study of normal andhypertensive subjects. New England Journal ofMedicine, 1938,2/9,650-655,

Bak, M, "Kalt-Test" untersuchungen an Schwangeren,Gebarenden und Wocherinnen, Monatschrift FiirCehurtschaft und Gynakologie, 1937, 105, 24-29,

Barber, T, X,, & Hahn, K, W,, Jr, Psychological andsubjective responses to pain-producing stimulationunder hypnotically suggested and waking-imagined"analgesia," Journal of Abnormal & Social Psychol-ogy, 1962,65, 411-418,

Barber, T, X,, & Hahn, K. W,, Jr, Experimental studiesin "hypnotic" behavior: Physiologic and subjective ef-fects of imagined pain. Journal of Nervous & MentalDisease, 1964, 139, 416-425.

Blitz, B., & Dinnerstein, A. J. Pain attenuation by con-tralateral cold stimulation. Psychonomic Science. 1968,10, 395-396.

Briggs, J. F., & Oerting, H. Vasomotor response of nor-mal and hypertensive individuals to thermal stimulus(cold). Minnesota Medicine, 1933, 16, 481-486.

Bronk, D. W., Pitts. R. F., & Larrabee, M. G. Role ofthe hypothalamus in cardiovascular regulation. Re-search Association for Nervous and Mentat DiseaseProceedings, (1939), 1940,20, 323-341.

Cier, J. F. Thesis, Lyon, 1938.Clemente, C. O., Sterman, M. B., & Wyrwicka, W.

Forebrain inhibitory mechanisms: Conditioning ofbasal forebrain induced EEG synchronization andsleep. Experimentat Neurology, 1963, 7, 404-417.

Dieckmann, W. J., & Michel, H. L. Thermal study ofvasomotor ability in pregnancy: Preliminary report.Archives of Internal Medicine. 1935, 55, 420-430.

Doupe. J., Miller, W. R., & Keller, W. K. Vasomotorreactions in the hypnotic state. Journal of Neurology &Psychiatry, 1939,2,97-106.

Edes, B., & Dallenbach, K. M. The adaptation of painaroused by cold. American Journal of Psychology,1936,45,307-315.

Engel, B, T. Some physiological correlates of hunger andpain. Journal of Experimental Psychology, 1959, 57,389-396.

Engel, B. T. Stimulus-response and individual-responsespecificity. Archives of General Psychiatry, 1960, 2,305-313.

Engel, B. T., & Bickford, A. F. Stimulus-response andindividual-response specificity in essential hyperten-sives. Archives of General Psychiatry, 1961, 5,478-489.

Eppinger, H., & Hess, L. Die Vagotonie, Berlin, 1910.(Translation) Vagotonia, New York: The Nervous andMental Disease Publishing Company, 1917.

Eyster, W. H., Jr., Roth, G. M., & Kierland, R. R,Studies on the peripheral vascular physiology of pa-tients with atopic dermatitis, Journat of InvestigativeDermatology. 1952, 18, 37-46,

Gellhorn, E, Autonomic regulations. New York: Inter-science, 1943,

Gellhorn, E, Physiological foundations of neurology andpsychiatry. Minneapolis: University of MinnesotaPress, 1953,

Gellhorn, E, Sympathetic and parasympathetic summa-tions,/4cra NettrovegaranVa, 1959,20, 181-194.

Gellhorn, E. Conditioning, sensations, and theergotropic-trophotropic balance. Conditional Reflex,1968, i, 34-44. (a)

Gellhorn, E. Central nervous system tuning and its impli-cations for neuropsychiatry. Journal of Nervous &Mental Disease, 1968, 147, 148-162. (b)

Gellhorn, E. Further studies on the physiology andpathophysiology of the tuning of the central nervoussystem. Psychosomatics, 1969, 10, 94-104.

Gellhorn, E. The emotions and the ergotropic andtrophotropic systems. Psychologische Forschung,\910,34, 48-94.

Gellhorn, E., & Louffbourrow, G. N. Emotions andemotional disorders: A neurophysiological study. NewYork: Harper & Row, 1963.

Germana, J. Response characteristics and the orientingreflex. Journal of Experimental Psychology, 1968, 78,610-616.

Germana, J. Central efferent processes and autonomic-behavioral integration. Psychophysiology, 1969, 6,78-90.

Glaser, E. M., Hall, M. S., & Whitlow, G. C. Habitua-tion to heating and cooling ofthe same hand. Journal ofPhysiotogy, \959, 146, 152-164.

Glaser, E. M., & Whitlow, G. C. Retention in a warmenvironment to adaptation to localized cooling. Journatof Physiotogy, 1957, 136, 98-111.

Goldberg, D. H. The physiological effect of multiplestressors. Behaviorat Science, 1966, / / , 438-443.

Greenfield, A. D. M., Shepherd, J, T,, & Whelan, R, F,Average internal temperature of fingers immersed incold water, Ctinical Science, 1950, 9, 349-354,

Greenwood, R, M,, & Lewis, P, D, Factors affectinghabituation to localized heating or cooling. Journal ofPhysiotogy, \959, 146, 10-11,

Griffin, J, P. The role ofthe frontal areas ofthe cortexupon habituation in man. Ctinicat Science, 1963, 24,127-134.

Hilgard, E. R. Pain as a puzzle for psychology andphys\o\o$y. American Psychologist, 1969,24, 103-113.

Hilgard, E. R., Cooper, L. M., Lenox, J., Morgan, A.H., & Voevodsky, J. The use of pain-state reports inthe study of hypnotic analgesia to the pain of ice water,Journat of Nervous & Mental Disease, 1967, 144,507-513.

Hines, E. A., Jr. Reaction ofthe blood pressure of 400school children to a standard stimulus. Journat of theAmerican Medicat Association, 1937, 108, 1249-1250.

Hines, E. A., Jr. The significance of vascular hyperreac-tion as measured by the cold-pressor test. AmericanHeart Journat, 1940, 19, 408-415.

Hines, E. A., Jr., & Brown, G. E. Standard stimulus formeasuring vasomotor reactions. Its application in thestudy of hypertension. Proceedings ofthe Staff Meet-ing ofthe Mayo Ctinic, 1932, 7, 332-335.

Hines, E. A., Jr., & Brown, G. E. Standard test formeasuring variability of blood pressure: Its significanceas index of prehypertensive state. Annats of InternatMedicine, 1933, 7, 209-217.

Hines, E. A., Jr., & Brown, G. E. The cold pressor testfor measuring the reactibility of the blood pressure:Data concerning 571 normal and hypertensive subjects.

American Heart Journat, 1936, / / , 1 - 9 .Hines, E. A., Jr., & Brown, G. E. Technic ofthe cold-

pressor test. Proceedings of the Staff Meeting of theMayo Ctinic, 1939, 14, 185-187.

lgersheimer, W, W, Cold pressor test in functional

May. 1975 THE COLD PRESSOR TEST AND HOMEOSTASIS 281

psychiatric syndromes. Archives of Neurology &Psychiatry, 1953, 70, 794-801,

Lacey, J. I. The evaluation of autonomic responses: To-ward a general solution. Annals of the New YorkAcademy of Sciences, 1956,67, 123-164.

Lacey, J. 1. Somatic response patterning and stress: Somerevisions of activation theory. In M. A. Appley & R,Trumbull, (Eds,), Psychological stress: Issues inresearch. New York: Appleton, Century, Crofts 1967Pp, 14-42.

Lacey, J. I., Bateman, D. E., & Van Lehn, R, Au-tonomic response specificity and Rorschach color re-sponses. Psychosomatic Medicine, 1952, 14, 256-260.

Lacey, J, I,, Kagan, J,, Lacey, B. C , & Moss, H. A. Thevisceral level: Situational determinants and behavioralcorrelates of autonomic responses. In P. H. Knapp(Ed.), Expression ofthe emotions in man. New York:International Universities Press, Inc 1963 PD161-196. " • ^'

Lacey, J. I., & Lacey, B. C. Verification and extension ofthe principle of autonomic response-stereotypy./I/nen-can Journat of Psychotogy, 1958, 71, 50-73.

Lacey, J. I., & Lacey, B. C. The law of initial value in thelongitudinal study of autonomic constitution: Repro-ducibility of autonomic responses and response pat-terns over a four year interval. Annats ofthe New YorkAcademy of Sciences, 1962,38, 1257-1290.

Lewis, T. Observations upon the reactions ofthe vesselsofthe human skin to cold. Heart, 1929, 15, 177-208.

Lovallo, W., Parsons, O. A., & Holloway, F. A. Au-tonomic arousal in normal, alcoholic, and brain-damaged subjects as measured by the plethysmographresponse to cold pressor stimulation. Psychophys-iology. 1973,10. 166-176.

McDaniel, C. D., Mefferd, R. B., Jr., Wieland, B. A.,Sadler, T. G., & Benton, R. G. Modification of gal-vanic skin responses by reaction time measurement.Psychophysiology, 1968,4, 340-341.

Mefferd, R. B., Jr., & Wieland, B. A. Modification inautonomicaily mediated physiological responses to coldpressor by word association. Psychophysiotogy, 1965,

Murphy, J. P., & Gellhorn, E, Influence of hypothalamicstimulation on cortically induced movements and actionpotentials of the cortex. Journat of Neurophysioloev1945,5,341-364.

Nakao, H,, Ballin, H, M,, & Gellhorn, E, The role ofthesino-aortic receptors in the action of adrenalin, norad-renaline, and acetylcholine on the cerebral cortex,Electroencephalography & Clinicat Neurophysiolosv1956,5,413-420,

Pickering, G, W. Regulation of body temperature inhealth and disease. Lancet I. 1958, 1-9,

Pickering, G, W,, & Kissen, M, The effects of adrenalineand of cold on the blood pressure in human hyperten-sion, Ctinical Science, 1936,2, 201-207,

Randall, L, M,, Murray, S, E,, & Mussey, R. D. The"cold test" in pregnancy: Preliminary report of its usein prenatal care, American Journal of Obstetrics &Gynecology, 1935, 29, 362-365.

Reider, N. Blood pressure studies on psychiatric patients.Bultetin ofthe Mennlnger Ctinic. 1938, 2, 65-67.

Rizzolo, A. Cold stimulation of a peripheral region andthe excitability ofthe cerebral cortex. British Journat ofMedicat Psychology, 1929, 9, 258-262.

Rosen, S, R. Vasomotor response in hysteria. yo«/-«a/o/the Mount Sinai Hospital. New York. 1951 18179-180.

Sadler, T. G,, Wieland, B. A., Mefferd, R. B., Jr., Ben-ton, R, B., & McDaniel, C. D, Modification in au-tonomicaily mediated physiological responses to coldpressor by cognitive activity: An extensionPsychophysiology, 1967, 4, 229-239.

Sadler, T. G., Mefferd, R. B., Jr., Wieland, B. A., Ben-ton, R. G., & McDaniel, C. D. Physiological effects ofcombinations of painful and cognitive stimuliPsychophysiology, 1969,5, 370-375.

Scherrer, H. Hypothalamic stimulation and blood pres-sure homeostasis. Acta Neurovegetativa, 1959, 20,205—218.

Scherrer, H. Effect of basal blood pressure levels onblood pressure responses to central nervous stimula-tion. Annals ofthe New York Academy of Sciences1962,98, 1257-1290.

Shapiro, A. P. An experimental study of comparative re-sponses of blood pressure to different noxious stimuli.Journal of Chronic Disease. 1961, 13, 293-311,

Sterman, M, B,, Clemens, T, L,, & Wenger, M, A. Au-tonomic and clinical correlates of differential elec-troencephalogram response to cold pressor stimulation.Psychophysiotogy, 1966,2, 327-337.

Stembach, R. A. A comparative analysis of autonomicresponses in startle. Psychosomatic Medicine 1960 22204-210. ' '

Sullivan, J. O. Dependence ofthe cold pressor reactionon peripheral sensation, Journat of the American Medi-cal Association, 1941, 117, 1090-1091.

Teichner, W. H. Delayed cold-induced vasodilation andbehavior. Journal of Experimental Psychology 196569, 426-432.

Teichner, W. H. Individual thermal and behavioral fac-tors in cold-induced vasodilation. Psychophysioloey1966,2,295-304. -^ y y sy^

Teichner, W. H. Interaction of behavioral and physiolog-ical stress reactions. Psychotogicat Review, 1968, 75271-291.

Teichner, W. H., & Levine, J. M. Digital vasomotorconditioning and body heat regulation.Psychophysiotogy, 1968, 5, 67-76.

Wenger, M. A. Studies of autonomic balance in Army AirForces personnel. Comparative Psychotogy Mono-graphs, Vol. 19, No. 4. Berkeley: University ofCalifornia Press, 1948.

Wenger, M. A., Clemens, T. L., Coleman, D. R., Cul-len, T. D., & Engel, B. T, Autonomic responsespecificity. Psychosomatic Medicine, 1961 23185-193.

Wenger, M. A., Clemens, T. L., & Cullen, T. D. Au-tonomic functions in patients with gastrointestinal anddermatological disorders. Psychosomatic Medicine1962, 24, 267-273.

Wenger, M. A,, & Cullen, T, D, ANS response patternsto fourteen stimuli, American Psychologist, 1958, 13423,

Wenger, M. A., & Cullen, T. D. Studies of autonomicbalance in children and adults. In N. S. Greenfield & R.A. Stembach (Eds.), Handbook of psychophysiology.New York: Holt, Rinehart and Winston, 1972 Pp535-570.

White, B. V., & Gildea, E. F. "Cold pressor test" intension and anxiety: A cardiochronographic study.Archives of Neurology & Psychiatry, 1937, 38964-984.

Wieland, B. A., & Mefferd, R. B., Jr. Systematic changesin levels of physiological activities during a four monthperiod. Psychophysiology. 1970, 6, 669-689.

282 WtLLIAM LOVALLO ^ot. 12. No, 3

Wilder, J. The law of initial values in neurology and on pain due to local cooling on f f "'•'^/"^"'''f'" |J^psychiatry Journat of Nervous & Mentat Disease, "cold pressor" effect. Research Pubtcat.ons of the?957/i 73-86. Association for Nervous and Mentat Disease, 1942,25,

Wolf S & Hardv J D Studies on pain; observations on 123-142.; L duf totocai cooking and on factors involved in the Yates, M. R., & Wood, J. E Jr. V-omotor response ofcold pressor effect. Journat of Ctinicat Investigations, non-hypertensive '"'^'^;^,"^'^^'° ^ J'^-^J^ t i1941 20 521-533, stimulus. Proceedings ofthe Society Jor

Wolf, S,, & Hardy, J. D, Studies on pain. Observations Biotogy and Medicine, 1936,34. 560-562.

REFERENCE NOTES

2. Lovallo, W., & Zeiner, A. R. Cutaneous vasomotor 1974.responses to cold pressor stimulation. Paper presentedat the 20th Annual Meeting of the SouthwesternPsychological Association, Dallas, 1973.

Announcement

THERMISTORS FOR BIOMEDICAL USE

Reprints ofthe 13-page paper, "Thermistors for Biomedical Use," by MeyerSapoff of Thermometrics, Inc., 15 Jean Place, Edison, NJ 08817, given at theFifth Symposium on Temperature, are now available upon request.

The various types of thermistors which have been found to be most suitable forbiomedical use are reviewed. Fabrication techniques employed for each basicthermistor type are described with particular emphasis placed on their resultantadvantages and limitations. For each thermistor structure considered, the rela-tionships which exist between the structure, its fabrication methods, and the costof maintaining close tolerances and interchangeability are also presented. Since ahigh degree of accuracy and the rapid measurement of very small temperaturedifferences are usually required, the testing and calibration problems peculiar tosuch biomedical sensors are discussed. A qualitative discussion of the factorswhich affect the stability of thermistors is then presented along with quantitativedata pertaining to the capabilities of each thermistor type. It is shown thatspecified requirements often exceed the practical capabilities ofthe devices used.Anticipated future developments of thermistors for biomedical use are also dis-cussed.