PsY(ll Beat-by-Beat Cardiac Responses in Normals and Schizophrenics to Events Varying in Conditional Probability STUART R. STEINHAUER' -, J, RICHARD jENNtNGS^ DANIEL P, VAN K.\MMEN'-^, AND JOSEPH ZUBIN'-^ 'Department of Veterans Affairs Medical Center. Pittsburgh, and 'Western Psychiatric ln.stitute and Clinic. University of Pittsburgh SchcH^l of .Medicine ABSTRACT Anticipatory cardiac deceleration and poststimulus acceleration were studied in schizopbrenic inpatient,s and controls during performance of a counting task. Reduced cardiac responding has been reported for schizophrenic patients for paradiRms usinR relatively lonR intertrial intervals. DurinR a relatively fast rate of stimulus presentation (3-s interstimulus interval), changes in cardiac interbeat interval were measured in 20 inpatient male cbronic schizophrenics and 18 control volunteers. Subjects counted an infrequent tone whicb was always followed by at least one frequent tone. Control subjects showed significant anticipatory cardiac deceleration precedinR the unpredictable tones, wbereas patient,s did not sbow a differential cardiac deceleration. Control subjects sbowed poststim- ulus acceleration tbat was inversely proportional to tbe conditional probability of events, wbereas patients exhibited Rreatly reduced poststimulus acceleration: patterns for botb Rroups resembled findings previously observed for event-related potential and pupillary dilation data. .Analysis of cardiac cycle time indicated siRnificant variation in primary bradycardia associated witb the delay between stimuli and immediately preceding R-waves in controls (replicatins l^cey & Ijicey, 1980), wilb only an immediate bradycardia at stimulus reception for patients rcRardless of cardiac cycle time. The data reinforce tbe notion thai tbe manner in wbicb information is used by scbizopbrenies, as reflected by cardiac responsivity, differs botb quantitatively and qualitatively from that of controls. DESCRIPTORS: Heart rate. Cardiac cycle time, Schizopbrenia, Information processing. Pri- mary bradycardia. Cardiac responsiveness provides a major source 1986). Without questioning the value of these dif- of information on autonomic nervous system ac- ferences, it is also informative to examine phasic tivity in schizophrenics and normal controls (Zahn. differences in cardiac activity during stimulus pre- 1986). Considerable attention has been paid to ton- sentatton. Phasic responses appear sensttive to in- ic heart rate levels in schizophrenics, with elevated formation processing (Coles & Duncan-Johnson, heart rates typically reported among patients com- 1975; van der Molen, Somsen & Orlebeke 1985) pared to controls (Spohn & Patterson, 1979; Zahn, and the information processing of schizophrenics is known to differ from that of normals (Kietzman, The authors wish to thank Theodore P. Zahn and Her- ^P""8- & ^"bin, 1980), Phasic cardiac changes, bert E. Spohn for their helpful comments. This research ''^crefore, provide a noninvasive probe of physio- was supported by the Medical Research Service of the 'oK'^^al acttvities that reflect more centrally me- Departmcnt of Veterans Affairs, and by NIMH Grant diated processing. Cardiac responses during stim- *MH436I5, Dr, Joseph Zubin. one of the leading pro- ulus presentation have been recorded both during Ponents of theoretical and expcnmental approaches to active task performance in schtzophrenics and con- Psychopathology. diexl ot, December 18. 1990, His co- ,rols. and during the prx^sentatton of stmple st.mult authors gratefully acknowledge his role in all phases of ^^;,h „« ovnli MI .^d A this work, as well as the continuing influence of his teach- ^"f "^ "P'"^'''^^"^^ requ.rcd. ing and guidance to students and colleagues throughout '""*'-' ^'"'I'^s dealing wtth cardiac responses Ihe world, lo orienting stimuli (a nontask situation), in which Address requests for reprints to: Stuart R, Steinhaucr. deceleration is normally expected (Graham, 1979), Ph,D,. Biometrics Research, 151R. VA Medical Center, unmedicated schizophrenics have been reported to Highland Drive, Pittshurgh. PA 15206, exhibit cardiac acceleration (Zahn. Rosenthal & 223 224 Steinhauer, Jennings, van Kammcn, and Zubin Vol. 29. No. 2 Lawlor, 1968), Gruzelier (1975) reported accelera- that it may be of value to vary task characteristics tion in schizophrenics who also showed a substan- within a single study of schizophrenics and normal tial skin conductance response (responders). but not controls. For example, signal stimuli could be var- in those patients who were nonresponders. In con- ied in informational complexity and predictability. trast, Bartfai, Levander, Edman. Schalling, and Sed- Finally, changes of the heartbeat just following vall (1983) observed no cardiac acceleration among stimulus presentation may bc a sensitive index of schizophrenics, but identified poststimulus decel- stimulus registration. The phenomenon of primary eration to at least the first tone in 92% of patients bradycardia (Lacey & Lacey, 1980) involves a rel- and 92% of controls. atively early deceleratory response, which has been These only partially consistent results from non- attributed to stimulus reception, rather than antic- task situations can be compared to those in which ipation. According to the Laceys, the time at which patients are asked to perform a simple task. When bradycardia can be observed is dependent on the a signal stimulus can bc anticipated in time, cardiac latency ofthe stimulus from the previous R-wave, deceleration is typically observed beginning prior When this latency is short, an immediate brady- to stimulus onset (anticipatory deceleration); after cardia can be seen, with a lengthening of interbeat stimulus delivery, cardiac acceleration tends to oc- interval during the very interval in which the stim- cur over the next several beats (Lacey & Lacey. ulus occurred (Interbeat Interval 0), If the stimulus 1974), These phenomena can be readily observed occurs late in the interbeat interval, then deceler- dunng reaction time tasks, and have been concep- ation is delayed until the succeeding interval (In- tually related to the development and dissipation terbeat Interval +1) rather than the stimulus in- of anticipatory attention (Coles & Duncan-John- terval. The presence of primary bradycardia among son. 1975; van der Molen et al,. 1985). schizophrenics, even in the absence of other major Several studies have examined cardiac activity cardiac changes, would suggest that there has been during active task performance by schizophrenics. registration of stimulus reception. To date, this phe- Schizophrenics did not show typical cardiac decel- nomenon has not been examined in schizophrenic eration between a warning signal and an aversive patients. tone, which could be terminated by a response fol- The present study involved the adoption of a lowing the aversive tone (Waddington, Maccolloch, paradigm that was originally employed for inves- Schalken. & Sambrooks. 1978), Similarly, unme- tigation of event-related potential and pupillary re- dicated schizophrenics have been reported to show sponses (Clark. Niemi. & Balogh. 1985; Steinhauer decreased anticipatory deceleration during simple & Zubin. 1982). in which probability was varied to reaction time tasks (Gray. 1975; Zahn. Carpenter. produce both predictable and unpredictable events. & McGlashan. 1981a, 1981b). Gruzelier (1975), In the typical oddball paradigm, one of two stimuli employing an unwarned tone discrimination task, occurs less frequently than the other, with the in- also failed to find deceleration; he reported accel- frequent stimulus producing larger amplitude psy- eration to signals by all patients, with schizophrenic chophysiological responses, A modification of the respwnders (defined by skin conductance response) paradigm restricts the occurrence of infrequent exhibiting cardiac acceleration to nonsignals as events to no more than one in a row. By informing well. In contrast, during performance of a span of the subject that this contingency is in effect, the apprehension task, Spohn. Thetford. and Wood- occurrence ofthe infrequent stimulus permits the ham (1970) found that both patients and controls attentive subject to exp)cct that the next stimulus exhibited pre- and poststimulus deceleration. will always be the frequent stimulus. In contrast, The poststimulus cardiac response has also been when a frequent stimulus occurs, it is not possible studied in size estimation tasks, with varying results to predict accurately whether a frequent or infre- dep)endingon the classification of patients based on quent stimulus event will occur next. The design additional diagnostic criteria, Warner (1973) re- enables the evaluation of differential anticipatory ported cardiac deceleration for schizophrenics with deceleration to predictable and unpredictable good premorbid histories, but cardiac acceleration events, poststimulus primary bradycardia, and among those with pxKir premorbid histories, poststimulus acceleratory activity. In addition, it McCormick and Broekema (1978) repwrted that allows comparison with data previously collected cardiac deceleration was greater in both paranoid on pupillary and event-related potential respond- and nonp>aranoid patients than in controls. ing, which indicated atypical responsivity among This brief review suggests that different tasks schizophrenic patients related to stimulus change, may alter pre- and poststimulus cardiac responding rather than to overall or conditional probability ef- in schizophrenics and normals. The findings suggest fects (Steinhauer & Zubin, 1982), March, 1992 Cardiac Activity in Schizophrenia 225 Metbod there would never be two high tones in succession. Subjects Prior to recording, subjects were asked what would occur next if a high tone was heard; all subjects could Twenty male patients from the Highland Drive VA correctly answer that a low tone would occur next. The Medical C enter, meeting Research Diagnostic Criteria sequence of tones was generated randomly by com- for chronic schizophrenia, were recruited for the study. puter to produce high tones at an overall probability Patient subjects were recently hospitalized. Mean of ,25. restricted only by the requirement that two high age±SD ofthe patients was 33,6 ±8,0 years. All but tones not occur in succession. 2 patients were receiving neuroleptic medication. Elev- Subjects sat in a separate testing room and com- en patients were receiving haloperidol, 4 received thi- municated with the exp)crimenter by intercom. The othixcne, and I each were receiving thioridazine, tri- room was maintained in darkness (because of simul- fluoperazine, or fluphenazine decanoate. This provid- taneous measurement of pupillary diameter); subjects ed a daily median dose of 875 chlorpromazine equiv- used a head rest and focused on an array of four red alents (CPZeq). with a range of 300-2000 CPZeq for LEDs to maintain fixation and minimize head move- 17 subjects, and I patient on haloperidol receiving ments. 3500 CPZeq, Seven patients were receiving antichol- Subjects were rewarded with a bonus of 25 cents inergics: I patient on haloperidol received 30 mg/day when they counted accurately, with 10 cents if correct of amantadine. while 3 patients on haloperidol. 2 on to within one or two counts, but no bonus for three thiothixene. and the patient on trifluoperazine each or more errors. The bonus and number of errors were received 2-4 mg benztropine, told to the subject after his report on each block. Male control volunteers, recruited primarily through local employment agencies, were screened for Data Reduction and Analysis psychiatric disorder using the Renard Diagnostic In- To ensure at least minimal attention to the task, terview (Helzer et al,, 1977), Eighteen control subjects data were included for analysis only when the subject meeting RDC criteria for no psychiatric disorder were reported with no more than thrce errors on a block; tested. Mean age ofthese subjects was 28,4 ± 5,1 years, in some cases, additional blocks were run to provide significantly younger than the patient group (/>=,025); four blocks of usable data per subject. However, most subsequent analyses were p)erfbrmed using age as a subjects made relatively few errors throughout the covariate. Controls were drug-free. Controls received task. Trials on which the subject moved or blinked $5/hour. in addition to a small bonus associated with excessively were coded and excluded from the data task performance. Patients received only the bonus analysis. The detection of multiple apparent R-waves associated with the task. Informed consent was ob- within 400 ms, or large changes in interbeat interval tained from all patient and control subjects prior to (>200 tns) from one interval to the next, were inter- testing, preted as artifacts by the data analysis program. To examine cardiac effects that both preceded and Psychophysiological Recording followed stimulus events, a trial was included for The electrocardiogram was derived from large analysis only when that trial, as well as the preceding Beckman Ag/AgCI electrodes placed under the lcf\ rib and subsequent trials, were all free of artifacts, pro- cage and on the right shoulder, with a small ground viding an overall time window of 9 s for examining electrode on the forehead. The signal was amplified by cardiac changes. a Grass Model 78 polygraph, and a peak detection Although the overall probability was ,25 fbr targets circuit (Shimizu. 1978) was used to detect the R-wave, and .75 for nontargets. stimulus events were further The time of each R-wave, to the nearest ms. was stored characterized by their second-order sequence (condi- on floppy disks of a PDP-11/03 (MINC) laboratory tional probability). Those nontargets preceded by a minicomputer. Interbeal interval in ms and heart rate target, symbolized by Nt, were completely predictable, in bpm were computed off-line. thus having a conditional probability of 1,00, All other stimulus events were not perfectly predictable. Fol- Procedure lowing any nontarget, the next stimulus would be a target on one-third of all trials (T, conditional prob- Each subject was presented with high-pitched tones ability " .33), and conversely, a nontarget on the re- (1500 Hz) and low-pitched tones (800 Hz) at 65dB. 40 maining two-thirds of such trials (Nn, conditional ms in duration (with rise/fall times of 4 ms). through probability — ,67; note that this represents the rep)e- a speaker placed in front him. One tone was presented tition of a nontarget following the preceding nontar- every 3 s for a total of 80 tones (trials) per block. At get). Thus, data were analyzed for each of the three least 4 blocks of tones were presented to each subject. conditional events labeled as T(,33). Nn(,67). and Subjects were first asked to identify sample tones Nt( 1,0), Note that the labels employed, based on con- as "high" or "low" in pitch. The subject was instructed ditional probabilities, do not necessarily reflect the to count silently the number of high tones (targets), subjective probabilities of each event for the controls and report the number of high tones (but not low and patients. tones) at the end of each block. Subjects were told that: The initial series of analyses was based on the av- I) there would bc fewer high than low tones, and 2) erage interbeat intervals across trials for each of the 226 Steinhauer, Jennings, van Kammen, and Zubin Vol. 29. No. 2 three probability conditions, beginning at two intervals Schizophrenics prior to stimulus onset (Interbeat Interval —2). through the interval of stimulus delivery (Interbeat T Ip- 33) Nn (p> 87] Interval 0). and continuing for four intervals following Nt jp-l stimulus onset (Interbeat Interval +4), Analysis of variance was employed using a repnrated measures design, wtth Diagnostic Group (control vs, schizophrenic) X Fixed Factorsof Probability (3 stim- ulus events) X Interbeat Interval (7 intervals). Where appropriate, degrees of freedom were reduced accord- ing to the Greenhouse-Geisscr correction (Jennings & Wood. 1976), The data were also analyzed for each subject by converting tnterhwat intervals on individuals trials to heart rate in bpm. The average heart rate was calcu- 2 3 4 lated for 14 periods of ,5 s each, beginning 2,5 s pnor Successive Interbeat Intervals to stimulus onset. Each interval bridging two separate Figure 2. Interbeat Interval data for schizophrenic pa- interbeat intervals was interpolated so that the rate tients (n —20) for each conditional probability event, measure was weighted by the proportion of time spent scaled to same range as in Figure I, in each interbeat interval. Because these analyses re- sulted in similar findings, details ofthe heart rate anal- yses have been omitted. prestimulus change, with acceleration rather than anticipatory deceleration prior to stimulus delivery, Results and some deceleration following the stimulus Beat-by-Beat Analysis event. The average interbeat interval for schizo- Interbeat interval data are plotted as a function phrenics (718,53 ms, equal to a heart rate of 83,50 of event probability for controls (Figure 1) and pa- bpm) was significantly shorter than for controls tients (Figure 2), Increasing interbeat interval—that (835,19 ms; 71.84 bpm) (F\ 1,36)= 10,54, /? = ,0025), is, cardiac deceleration—is indicated by upward de- The faster heart rate of the patients is consistent flections in these figures. Several trends can be seen with the majority of repwrts in the literature. in the data for the controls (Figure 1). Prior to the Significant main effects were also obtained for two stimulus events that were not predictable (the probability (/•T(2,64)'=3,39. /)=',O45) and over suc- T(,33) and Nn(,67) conditions), cardiac decelera- cessive interbeat intervals (/•X3,102) —3,95, p— tion was observed. Following stimulus presenta- ,012), with a Probability X Interbeat Interval in- tion, greater cardiac acceleration occurred following teraction (/•R132) =4,17,/?-,0042) also observed, the T(,33) condition than the Nn(,67) condition. A three-way Probability X Interbeat Interval X Di- The response in the Nt(I,O) condition showed little agnosis interaction (/^4.132)-=2,96, ^-",026) re- flects significantly greater changes among the con- Controls trol subjects across both probability and interbeat eao T fp. 33) interval as compared to patients. »•« Nn |p*eT] »-• Nt ip-l OJ Given the multiple interactions of interbeat in- terval with probability in the control data set, fur- ther analyses were carried out in which prestimulus, poststimulus, and immediate stimulus reception could be focused upon. For this purpjose, scores were develop)ed to capture the prestimulus related changes (i,e,, anticipatory deceleration), poststim- ulus changes (i,e., pmststimulus acceleration), and changes occurring at the time of stimulus delivery. The analyses focusing on the time of stimulus re- ception took into account stimulus timing at dif- -10 12 3 4 ferent intervals with respect to the preceding R- Successive Interbeat Intervals Fi^arc I. Interbeat Interval data for control subjects wave in the investigation of primary bradycardia. (n— 18) for each conditional probability event. Increasing interbeat interval indicates cardiac deceleration pnor to Anticipatory Deceleration stimulus delivery in tbe "T" and "Nn" conditions, fol- The extent of anticipatory deceleration prior to lowed by cardiac acceleration, which is greatest for tbe stimulus onset was calculated as the difference be- least probable event ("T~), tween Interbeat Interval —2 and Interbeat Interval March, 1992 Cardiac Activity in Schizophrenia 227 Schizophrenic! tervals -2 and 0 for controls (/1(1,I7)- 15.58, p= Conlrol* ,001), but not for patients. a Anllcipalory Acceleration Deceleratton Poststimulus Acceleration I- Poststimulus acceleratory activity was calculat- & ed as the difference between Interbeat Interval +3 ^ a and Interbeat Interval 0 (Figure 3). In parallel with the above findings, there was greater acceleration .5 0 among controls than p}atients (diagnosis: fl[l,36) = ac 5.97. /7=,O2). increasing acceleration for more rare c events (probability; fT2,54)-7.15. /7 = ,0041), with 5 a Probability X Diagnosis interaction (F\2,54) = -1- 7,80, p=',OO27) also observed. Simple effects indi- T Nn Nl T Nn Nt .33 .67 1 0 33 67 1.0 cated significantly greater acceleration in the infre- quent T(,33) condition than in both ofthe frequent Conditional Probability conditions for controls only, Covarying prestimulus Finure 3. Prestimulus cardiac deceleration (Intcrtxjat level eliminated only the main effect for diagnosis. Intervals - 2 to 0) and poststimulus cardiac acceleration In separate analyses by subject group, control sub- (Interbeat IntervalsOto +3) for control and schizophrenic groups as a function of conditional probability of stimulus jects showed a significant overall increase in rate events. In this figure, deceleration is indicated by negative from Interbeat Interval 0 to +3 (fll,17)= 16.19, values, acceleration by positive values. p=.OOO9). For patients, a trend for acceleration (/•T[ 1.19)-= 3.62. p-,072) seemed related to accel- 0 for each probability condition (Figure 3). A main eration only in the T(,33) and Nt(l,O) conditions. effect for probability (fT2.68) = 8,23. p'-.OOOS) and This paralleled findings reported previously for a Probability X Diagnosis interaction {F\2,67)'= other psychophysiological measures. Planned anal- 5.82. /7=,(X)56) reflect greater deceleration in the yses to evaluate acceleration between Interbeat In- T(,33) and Nn(,67) conditions and acceleration in terval 0 and Interbeat Interval -1-3 only for the T the Nt(I,O) condition for controls, but lack of and Nt conditions in the patients indicated a sig- change for the patients. nificant acceleration (/•\1.19)-4,68, />-,043), Because the acceleration in the Nt( 1,0) condition which was not reduced by covarying age or pre- appxiared to be indicative of a continuing acceler- stimulus level, ation to the immediately preceding target, a more spweifie analysis was limited to anticipatory decel- Anticholinergic Medication Effects eration in only the unpredictable T(,33) and To evaluate the contribution of anti-Parkinson- Nn(,67) conditions. There was a significantly great- ian agents, comparisons between the 7 p>atients re- er increase in interbeat interval for controls than ceiving anticholinergics and the other 13 patients for patients (/•T1.36)-7,34. /J=,01). but no differ- were carried out. There were no significant differ- ences between the conditions, indicating similar de- ences in prestimulus level or acceleration, but a eeleration prior to unpredictable events,' Signifi- marginal trend (/-111. 18)=3,17, ;>-=,092) was ob- eant deceleration occurred between Interbeat In- served for greater average decelerations in the pa- tients receiving anticholinergics than for those with- out anticholinergics. Thus, use of anticholinergics 'Bccausc ofthe complicated assumptions required in analysis of covariance. all analyses were run with as well is unlikely to account for the smaller decelerations as without covariates (age and neuroleptic medication). seen in general among patients as compared to con- In addition, subsamples were examined to ensure that trols. these covariates were not causing the primary results de- scribed. In the case of age. subsamples of equal age were Primary Bradycardia compared across conditions and interbeat intervals. Em- To test for the occurrence of primary bradycar- ploying C PZeq to describe dosage is problematic because dia in relation to cardiac cycle time, individual of differences in therapeutic effectiveness of different trials at each probability were first sorted according drugs, and these were not systematically assessed for all patients. Differences between patients on and off anti- to the delay between stimulus onset and the pre- cholincrgics are reported below. The only analysis affected ceding R-wave, using increments of 250-ms inter- by covarying age, neuroleptic, and prestimulus level was vals (e.g.. 0-250 ms. 251-5(X) ms. etc). Because the this difference in deceleration (/>—,01). which was reduced number of trials in each condition is reduced by to a marginal level (/)—,054) when all covariates were this sorting, only three intervals provided sufficient included; only prestimulus level was a significant covar- numbers of trials for cardiac cycle analysis (0-250, iate. 251-500, and 501-750), Data for one patient and 228 Steinhauer, Jennings, van Kammen, and Zubin Vol. 29, No. 2 two control subjects were excluded due to fewer Although there was no significant interaction be- than five trials in at least one cell for each subject. tween cardiac cycle time and diagnosis, it was of In this analysis, changes were of interest only across interest to examine effects during the pwriod of stim- the three interbeat intervals proximal to stimulus ulus reception separately for each ofthe two groups. onset—Interbeat Interval — I. Interbeat Interval 0. Controls exhibited the phenomena of primary and Interbeat Interval +1. These intervals, sepa- bradycardia as described by the Laceys (Lacey & rated by probability and by beat-to-stimulus delay, Lacey. 1980). with differences due both to changes are plotted in Figure 4 (controls) and Figure 5 across interbeat interval as well as among proba- (schizophrenics). In addition to findings reflecting bilities. The effect for primary bradycardia occurred the analyses described previously, there were sig- in the T(,33) and Nn(,67) conditions; greater de- nificant interactions with cardiac cycle time (Inter- celeration was observed for the interbeat interval beat Interval X Cardiac Cycle: fl(3,98)-=3,38. /;= of stimulus occurrence when there was less delay ,022, and Probability X Interbeat Interval X Car- between the stimulus and the preceding R-wave, diac Cyele: fT5,150)-3,38, p-.0QS7). For the schizophrenic patients, there were no significant differences related to the delay between the previous beat and the time of stimulus onset— Controls that is, no differential effect of primary bradycardia 860 occurred in relation to the interval between the R- Beal-to-Stlmulu9 Dolay (ms) •-• O-Z&O wave and stimulus. However, a significant main o-o 2S1 -500 effect among the three interbeat intervals was found 1 850 » -« 501 -750 (/•(2,35)=3.48. /7=,O43), reflecting deceleration from Interbeat Interval - 1 to Interbeat Interval 0, t 840 V Discussion _c The greatest difference observed b>etween schiz- - 830 d •• V ,0 ophrenics and controls was in average heart rate; u patients had an average rate that was approximately T p- 33 Nn. p- 67 Nl. P-! 0 12 bpm faster than the rate for controls. This rep- licates the major findings on overall heart rate level BIO in schizophrenia (Zahn. 1986), In addition, patients 0 1 -tot -1 1 exhibited an absence of anticipatory deceleration Interbeat Interval x Event and little p)oststimulus acceleration. In general, ) 4. Cardiac cycle time effects (Beat-to-Stimulus there was a pattern of less cardiac respxjnsivity dur- delay) for control subjects (n'-l6). For each event, the ing this task for the patients. interbeat intervals preceding, dunng. and following stim- ulus onset are plotted as a function ofthe latency between Reduced phasic cardiac responsivity has been stimulus onset and the preceding R-wave. grouped by 250- reported in several other paradigms employing both ms intervals. drug-free and medicated patients, Zahn, Schooler, and van Kammen (1985) reported less anticipatory Schizophrenics deceleration among patients during performance of a recognition task. These effects were present even D«lsy though all patients were medication free. Increases -• o-ZfiO in heart rate have been seen for patients receiving E 251 500 e 740 H - 50i-7S0 phenothiazines, with a correlation between dosage and increased rates for those patients with the low- L 730 est heart rates (Spohn. Thetford, & Cancro. 1971), « For patients administered chlorpromazine. de- creased decelerative responses were seen over an eight-week period as compared to patients reeeiving placebo (Spohn. Lacoursiere. Thompson, & Coyne, 1. 1977), However, other data strongly indicate that T, p- 33 Nn. p- 67 Nt. p-1 0 neuroleptics by themselves are not likely to be pri- marily responsible for either the greatly elevated 700 -t 0 I -tot -10 1 heart rates or the reduced cardiac responsivity in Interbeat Interval x Event schizophrenics during tasks. Patients tested first Figiirc 5. Cardiac cycle time effects for schizophrenic while drug-free and later after either haloperidol or patients (n— 19). scaled to same range as m F-'igurc 4. chlorpromazine treatment showed higher rates than March. 1992 Cardiac Activity in Schizophrenia 229 controls regardless of medication status (Castcllani been presented. At the other extreme, they may et al,. 1982), In addition, going from a supine to have been attempting to px;rfonn the task as well standing position resulted in a minimum mean ac- as possible (as patients often report), but could not celeration of at least 15 bpm. even for patients with effectively focus their attention. Zahn (1970) ob- extremely high initial heart rates (Castellani et al,. served that schizophrenics were unable to over- 1982), come effects of "minor set" during a reaction time In the present data, control subjects showed a task even when preparatory intervals were indicat- clear anticipatory deceleration when the succeeding ed to the patients. stimulus could not be predicted—i,e,, in the T(,33) Other data suggest that differences in autonomic and Nn(.67) conditions. For these conditions, stim- responsivity may be associated with an ability to ulus delivery eventually resulted in increased ac- utilize information. In an initial conditioning task, celeration, with greater acceleration associated with schizophrenics as a group showed less skin resist- the T(,33)condition. Given that the rare T(.33) tone ance and cardiac (deceleration) change than con- was also the relevant event associated with counting trols, and only controls benefitted from information behavior, these data cannot dissociate the effects of on the conditioning contingency (Rist. Baumann. the rareness of the event from the relevance of the & Cohen. 1981), Rist and colleagues then selected target on acceleratory activity. Prestimulus accel- only patients who were electrodermal respwnders. eration in the Nt(l,O) condition app)ears to be a Cardiac responding was observed prior to and dur- continuation ofthe large acceleration generated by ing the conditioning phase in all groups. During the immediately preceding target stimulus. These conditioning, greater deceleration to the CS-t- pre- results replicate the primary cardiac findings of La- ceding a noxious loud tone, and greater acceleration cey and L,acey (1980). who employed a similar par- to an innocuous imperative stimulus as UCS. were adigm differing primarily in the difficulty ofthe au- seen among those patients who were informed ditory discrimination (target and nontarget tones about the CS-UCS contingencies than among those were, respectively. 1000 and 1050 Hz), and who did patients who were uninformed. Thus, utilization of not inform their subjects that targets never (Kcurred information corresp>onded with cardiac changes in in succession. these autonomically resp)onsive patients. Schizophrenic subjects differed from controls in The pattern of pxiststimulus acceleration in the both anticipatory deceleration and pwststimulus ac- schizophrenic patient data (Figure 3) is of particular eeleration. Several possibilities are suggested by interest. In the two extreme probability condi- these findings. That medications were responsible tions—T(,33) and Nt( 1,0)—there was mild though for floor and ceiling effects, restricting possible var- significant acceleration, but little change tn the iations in heart rate, is unlikely given the report of Nn(,67) condition. In our earlier evaluation of this Castellani et al. (1982), although it would have been paradigm (Steinhauer & Zubin. 1982). involving dif- useful to have obtained similar data resulting from ferent samples of schizophrenics and controls, we changes in bodily p)osition for the current subjects. repwrted patterns for both pupillary dilation and the As indicated by the present findings, administration P300 compwnent ofthe event-related potential that of anticholinergic agents also are not sufficient to are similar to those observed in the present cardiac explain the reduced reactivity ofthe patients, acceleration data. The patients were similarly char- A second alternative is that patients were not acterized by greater amplitudes in the T(,33) and aware ofthe fact that some stimuli were predictable Nt( 1,0) conditions, in which the stimulus was phys- and others were not. The instructions to subjects ically changed from the previous trial, as compared only required counting ofthe infrequent target once to the Nn(,67) condition, which was associated with it had occurred, so patients may have ignored the stimulus repKtition, Controls showed increasing fact that only single targets occurred. We note, how- amplitudes of pupillary dilation and P300 with de- ever, that patients were able to report, even at fol- creasing event proliability. Thus, the trends ob- low-up discussions several weeks after testing, the served in these cardiac data are similar to those contingency of no more than one target occurring obtained with other psychophysiological measures in a row. If patients were actively treating all events in schizophrenics. The status of such cardiac as unpredictable, then clear decelerations should changes as either indicators of episodes or p)ersistent have been generated prior to all stimuli, which was markers of vulnerability to schizophrenia (Zubin & not clearly demonstrated. Steinhauer. 1981) remains a further question to be Two other alternatives are quite pwssible. It is clarified. conceivable that patients were not actively trying There are several ways in which the p»aradigm to employ the information regarding predictability Utilized in this study differs from previous evalu- of subsequent stimuli after a target or nontarget had ations of cardiac responses during task performance 230 Steinhauer, Jennings, van Kammen, and Zubin Vol. 29, No. 2 in schizophrenics. In the present study, testing was the analysis indicated a significant deceleration be- prolonged over at least 320 trials, with a short 3-s tween Interbeat Intervals — I and 0 for the patients. interval between trials. In previous studies, the This immediate deceleration in the schizo- number of experimental trials has ranged from as phrenic sample is surprising, because it was not ob- few as 5 stimuli plus 2 catch trials (Waddington et served in the initial analyses across all seven inter- al., 1978) to 20 warned reaction time trials (Zahn beat intervals (presumably because of greater var- et al,, 1981a). with intervals between trials of no iability when all intervals were considered at once) less than 24 s before a warning signal (Gray. 1975) or even in the anticipatory deceleration response. to as much as 60 s (Gruzelier. 1975), Thus, in the It suggests that there is indeed some cardiac re- present study, more consistent attention was re- sp>onse in the patient group, which may be related quired to a considerably greater number of trials, to effects of stimulus reception. However, while with a faster rate of presentation, and over longer .stimulus reception occurs, the effects of further periods of time, than for previous expenments. evaluation, normally exhibited by cardiac acceler- Although the studies of Gray (1975), Gruzelier ation in relation to the information provided (Lacey (1975), Waddington et al, (1978), and Zahn and & Laeey. 1974), are not observed in the patients. colleagues (198 la; Zahn. Rumsey. & van Kammen, This situation has an analog in the event-related 1987) all involved an overt motor response by the potential literature; brainstem auditory potentials, subjects, in the present task, a coven behavior which are related primarily to sensory processing, (counting) was employed. However, the counting are generally reported to be intact in schizophrenics task was sufficient to elicit vigorous anticipatory (Brecher & Begleiter. 1985; Pfefferbaum, Horvath, deceleration and p)oststimulus acceleratory activity Roth, Tinklenberg. & Kopell. 1980). whereas the in control subjects, A separate group of 19 control longer-latency P300 component, associated with subjects has also been tested using the same para- cognitive processing, is typically attenuated in these digm involving the counting task on some blocks, patients compared to controls (Roth, Teece, Pfef- and a discriminative button press to different tones ferbaum. Rosenbloom. & Callaway. 1984; Zubin. on other blocks. The same pattern of cardiac Kietzman. & Steinhauer. 1985), changes reported here is seen across the two tasks, The increased average heart rate for the schiz- with the exception that deceleration is also seen ophrenic patients compared with controls is con- prior to the 1,0 condition when a response must be sistent with the notion of maintained high levels of executed. Thus, the nature of the task itself does sympathetic arousal (Zahn. 1986) in schizophrenia. not seem to preclude the development of cardiac What is less clear is why there should be so little activity. In current studies, both the counting and change in heart rate, espxxially the lack of slowing discrimination tasks are being employed with pa- of an already quickly beating heart. Slowing of heart tients. rate is primarily due to vagal inhibition (cardiac Time-dependent primary bradycardia was ob- deceleration following increased vagal activity). served across different probabilities among control This would suggest decreased responsiveness of subjects, as previously described by Lacey and La- central piarasympathetic mechanisms, but could cey (1980), Schizophrenics did not exhibit a differ- also be due to central inhibition of parasympathetic ential respK>nse on the basis of the delay between activity by the ongoing high levels of sympatheti- the preceding R-wave and the stimulus. However, cally mediated processes. REFERENCES Bartfai, A,. Levander. S,. Edman. G,, Schalling. D,. & Coles, M,G.H,. & Duncan-Johnson. CC, (1975), Cardiac Sedvall. G. (1983), Skin conductance responses in un- activity and information processing: The effects of medicated recently admitted schizophrenic patients. stimulus significance, and detection and response re- Psychophysiology. 20. 180-187. quirements. Journal of Experimental Psychology: Hu- Brecber. M.. &. Begleiter. H, (1985), Brain stem auditory man Perception and Performance, I, 418-428. evoked potentials in unmedicated schizophrenic pa- Graham. F.K,(1979), Distinguishing among orienting, de- tienu. Biological Psychiatry. 20, 199-202. fense, and startle reflexes. In H.D. Kimmol, E,H, van Castellani. S., Ziegler, M.G,, van Kammen. DP,. Alex- Olst. & J,F, Orlebeke (Eds), The orienting reflex in ander. P,E,. Sins. S,G,. & Lake. CR, (1982), Plasma humans (pp, 137-197), Hillsdale, NJ: Lawrence Earl- norepinepbnne and dopamine-bcta-hydroxylasc activ- baum. ity in schizophrenia. Archives of General Psychiatry, Gray. A.L, (1975), Autonomic correlates of chronic schiz- 39, 1145-1149. ophrenia: A reaction time paradigm. Journal of Ab- Oark. W,R,. Niemi. CA,, & Balogh. D.W, (1985), Task- normal Psychology, 84, 189-196. evoked pupillary response to a conditional probability Gru7x:lier. J.H, (1975). The cardiac responses of schizo- counting Usk, Psychophysiotogy, 22, 586-587, phrenics to orienting, signal and non-signal tones, Bi- March, 1992 Cardiae Activity in Schizophrenia 231 ological Psychology: 3, 143-155. Steinhauer. S.R.. & Zubin. J. (1982). Vulnerability to Hel/cr, J.E.. Clayton. P.J.. Pambakian. R.. Reich. T. schizophrenia: Information proccsstngin the pupil and Woodruff. R,A,, Jr,, & Reveley. MA, (1977), Reli- event-related potential. In E. Usdin & I, Hanin (Eds.). ability of psychiatnc diagnosis: II, The test/retest re- Biological markers in psychiatry and neurology- (pp, liability of diagnostic classification. .-Irc/i/j'wo/'f/rnpra/ 371-385), Oxford: Pergamon Press. Psychiatry: i< 136-141, van der Molen. M.W., Somsen, R,J.. & Orlebeke. J.F, Jennings, J.R,. & Wood. CC, (1976), The epsilon-adjust- (1985), The rhythm ofthe heart beat in information mcnt procedure for repeated-measures analyses in hu- processing. In P.K. Ackles. JR. Jennings. & M.G.H. mans, Psychophysiology: 13. 277-278. Coles (Eds.). Advances in p.sychophysiotogy (Vol. 1, pp. Kiet/man, ML, Spring, B.. & Zubin. J. (1980). Percep- 1-88). Greenwich, CT: JAI Pres,s. tion, cognition, and attention. In H.I. Kaplan, A.M. Waddington, J.L, Maccolloch, M,J,, Schalken, ML., & Freedman. & B,W. Sadock (Eds), Comprehensive text Sambrooks, J,E. (1978) Absence of cardiac decelera- Imok ofpsychtatry/lll (pp, 334-370), Baltimore: Wil- tion m a signaled escape paradigm: A psychophysio- liams & Wilkins, logical deficit in chronic schizophrenics. Psychological l^ccy. B.C .,& Lacey, J,I. (1974). Studiesof heart rate and Medicine. 8, 157-162. other bodily processes in sensorimotor behavior. In Warner. D.E, (1973) Cardiac rate responses in two groups PA, Obrist, AH. Black, J. Brener, & L,V. DiCara of schizophrenics during the performance of a size es- (Eds.). Cardiova.scular psychophysiology (pp, 538- timation task. Unpublished master's thesis. University 564). Chicago: Aldine. of Wisconsin, Madison, (cited by Spohn & Patterson, Lacey. B.C. & Lacey. J.I. (1980). Cognitive modulation 1979), of time dependent primary bradycardia. Psychophys- Zahn, T,P, (1970), Effects of reductions in uncertainty on iotogy: 17. 209-221, reaction time in schizophrenic and normal subjects. McCormick. D,J,, & Broekema, V,J. (1978), Size esti- Journal of Experimental Research in Personality. 4, mation, perceptual recognition and cardiac rate re- 135-143, sponse in acute paranoid and non-p>aranoid schizo- phrenics. Journal of Abnormal Psychology. 87, 385- 2^hn, TP, (1986), Psychophysiological approaches to psy- ,398. chopathology. In M,G,H, Coles. E. Donchin. & S.W, Porges (Eds.), P.sychophysiotogy: Systems, proces.ses. I'fcffcrhaum, A,. Horvath. T,B,. Roth. W,T,. Tinklenberg, and applications (pp, 508-610), New York: Guilfortl JR., & Kopell, B.S, (1980), Auditory brainstem and Press, cortical evoked potentials in schizophrenia, Btolotfical Psychiatry: 15, 209-222. Zahn, T,P,. Carpenter. W.T,. & McGlashan. T,H, (1981a), Rist. F., Baumann, W,, & Cohen, R. (1981). Effects of Autonomic nervous system activity in acute schizo- awareness and motor involvement on autonomic con- phrenia: I, Method and comparison with normal sub- ditioning in chronic schizophrenics, Pavlovian Journal jects. Archives of General Psychiatry, 38, 251-258, of Biological Science, 16, 8-17, Zahn. TP,, Carpenter, W.T., & McGlashan. T,H, (198 lb). Roth, W,T.. Teece. J,J,. Pfefferbaum. A,. Rosenbloom, M,, Autonomic nervous system activity in acute schizo- & Callaway. E, (1984). ERPs and psychopathology: I. phrenia: II, Relationships to short-term prognosis and Behavioral process issues. In R, Karrer. J, Cohen. & clinical state. Archives of General Pswhiatry: 38, 260- P, Tueting (Eds.). Brain and information: Event-relat- 266, ed potentials. Annals ofthe New York Academy of Sci- Zahn, T,P., Rosenthal, D,. & Lawlor. W.G. (1968). Elec- ences (Vol. 425. pp. 496-522), New York: New York trodermal and henn rate orienting reactions in chronic Academy of Sciences, schi7QphTtnia. Journal of Psychiatric Research. 6.117- Shimizu. H, (1978), Reliable and precise identification of 134, R-waves in the EKG with a simple peak detector, Psy- Zahn. T.P,, Rumsey, J,M,. & van Kammen. D.P, (1987). chophy.siotogy. /.5. 499-501. Autonomic nervous system activity in autistic, schiz- Spohn, J.E,, Lacoursiere. R,B,. Thompson. K,, & Coyne. ophrenic, and normal men: Effects of stimulus signif- L, (1977), Phenothiazine effects on psychological and icance. i (Manuscript received January 26, 1990; accepted for publication February 12. 1991)