International Journal of Psychophysiology, 12 (1992) 41-51 0 1992 Elsevier Science Publishers B.V. All rights reserved 016%8760/92/$05.00

41

INTPSY 00369

tensity an cardiac an

e elicitation of unconditio ermal responses

Gerhard Vossell and Heinz Zimmer Psychologisches Institut der lJnil?ersitiirMaim, Maint (F.R. G.)

(Accepted 21 August 1991)

Key words: Rise time; Stimulus intensity: Startle response; Orienting response; Defense response; Heart rate; Skin conductance

In recent discussions on the differentiation of orienting, startle and defense responses, the importance of stimulus rise time for the elicitation of different cardiac response patterns was re-emphasized. Especially, it has been claimed that phasic accelerative heart rate (HR) responses-interpreted as indicators of startle-might not only be evoked by auditory stimuli with instantaneous rise times and high intensities, but also by low to caoderate stimulus intensities with sudden onsets. The present study examined this question by manipulating rise time (instantaneous vs. 50 ms) and intensity (60 vs. 95 dB). Subjects (N= 120) were randomly assigned to one of the four independent groups. They performed a habituation experiment in which 12 tones of 1000 Hz with a constant interstimulus interval of 90 s were presented. On trial 13, a change in rise time was introduced by reversing the rise time condition in each group (i.e., from instantaneous to 50 ms and vice versa). Examination of HR changes across 4 poststimulus seconds, of maximal HR deceleration and acceleration, as well as examination of magnitude of skin conductance responses (SCRs) indicated clear intensity effects. Rise time, on the other hand, had no effects on HR and SC that could be interpreted as indicative of startle. Thus, it could neither be observed that instantaneous rise time led to stronger accelerative HR responses nor that the combination of instantaneous rise time and high intensity elicited anything but a dominant accelerative response pattern. These findings were also obtained when the first four trials were analyzed separately. As a consequence of stimulus change, larger SCRs as well as larger decelerative HR responses were observed without, however, being affected by the direction of the change in rise time. In sum, the present‘study suggests that the role attributed to rise time with respect to eliciting qualitatively different cardiac response patterns has been exaggerated. The consequences for the differentiation of different unconditioned responses are briefly addressed.

INTROIXJCTION In recent contributions by Barry (1987) and Turpin (1986) concerned with the problem of identifying and differentiating unconditioned responses as ‘orienting’, ‘startle’, and ‘defense’ on the basis of autonomic nervous system reactivity, especially on the basis of heart rate (HR), a number of contradictory conclusions can be

Cbrwpondeme:

G. Vmsel, Psychologisches la%itut der Johannes Gutenberg-Universitiit Mainz. Staudingelweg 9. Postfach 3980, D-6500 Mainz. F.R.G.

found. They have largely to do with stimulus features being sufficient or necessary for the elicitation of these responses and their manifestation in different patterns which should permit unequivocal inferences about the relevant theoretical constructs. One major problem with classical resesrch on HR deceleration as an index of the orienting response (OR) is the prominent assumption that an ‘instantaneous’ rise time for auditory stimuli may distort HR data insofar as a rapid rise could elicit accelerative responses or response tendencies (a common index for startle) that interfere with decelerative response components, espe-

42

cially as both types of HR change are dependent on vagal mediation (cf. Levy 8r Martin, 1979). This competition expected particularly in early trials could prevent a regular decline in event-related HR deceleration, an essential prerequisite for infering habituation (cf. Barry, 1982; Vessel & Zimmer, 1989). If it is now argued that not only a combination of high intensity and instantaneous rise time may evoke HR startle (cf. Graham, 1979), but that startle responses in HR can also be evoked by low to moderately intense stimuli with sudden onsets (cf. Simons, 19891, then results and conclusions of a number of studies in the context of the OR would require reappraisal. An examination of the literature, however, reveals that the assumption that accelerative HR responses are evoked by low to moderately intense stimu!i with sudden onsets is hardly supported by empirical data. In fact, only one published report (Hatton, Berg & Graham, 1970) systematically investigated the effects of different rise times and different intensities on HR *. Hatton et al. reported two experiments. In the first, stimulus intensities of 50 and 75 dB, and rise times of 300 and 3 ms were compared. Under all four conditions the initial HR response was decelerative. The only rise time effect observed was that the 3 ms rise slightly shortened the duration of initial deceleration independently of stimulus intensity. The second experiment compared intensities of 50 and 90 dB, and rise times of 300 ms and less than 5 hs. Results showed an immediate HR deceleraticq to the 50 dB stimu!us, whether or not the rise time was gradual. On the other hand, the 90 dB stimulus evoked immediate deceleration when rate of rise was gradual, and an immediate HR acceleration when rate of rise was instantaneous. This study, therefore, appears not to support the assumption that accelerative HR responses indicative of startle are evoked by low to moder-

* A second study by Berg, Jacksnn and Graham investigated the effects of different

(1975) also

rise times and intensities

in detail, as the effects were studied during sleep, and it is known that

on HR. This study, however, will not be considered

HR responses in sleep differ from those in the waking state.

ately intense sensory stimuli with instantaneous rise times. It only demonstrated that immediate HR accelerations can be evoked by a combination of relatively high stimulus intensity with a rapid rise. There are, however, several problems with the Hatton et al. study that limit the generalization of their findings: (a) the presentation of the four types of stimuli within subjects (balanced across blocks) may have induced systematic effects in some groups (e.g., due to sensitization or changes in adaptation level, cf. Turpin & Siddle, 1978) that may not be present when subjects are repeatedly exposed to the same stimulus; (b) Hatton et al. presented their HR data as an average of the four stimulus presentations without analyzing single trials. This might have obscured effects only present in early trials; (cl a controlled rise time of 300 ms is unusually long. In more recent experiments (e.g., Simons, Rockstroh, Elbert, Fiorito, Lutzenberger & Birbaumer, 1987; Turpin & Siddle, 1978, 1979, 1983), rise times in the range of 25 to 50 ms were considered as sufficiently sli;~ to avoid the elicitation of startle responding; (d) Hatton et al. analyzed only poststimulus HR on a second-by-second basis. As noted by Turpin and Siddle (1983). such a procedure may obscure the detection of startle responding. According to their view, only the use of uni-directional analyses of decelerative and accelerative components can demonstrate the presence or absence of startle; and (c) in the meantime, alternative criteria for the identification of startle responses in HR have been formulated. While Hatton et al. interpreted a cardiac acceleration with an onset latency of less than 2 s as reflecting startle (see also Graham, 1973, 19791, Turpin (1986; Turpin & Siddle, 1983) assumed that startle is reflected by a cardiac acceleration with a peak laten cy of approx. 4 s. As noted by Turpin (1983). these two terms are not equivalent and might !ead to different conclusions. The aim of the present study was, therefore, to present a constructive extension of the Hatton et al. study taking into account the critical points raised. Two levels of stimulus intensity were factorially combined with two rise time conditions. Nonsignal auditory stimuli of 60 or 95 dB, with rise times of 50 ms or 1 ms were used. A be-

43

tween-subjects design was er,lployed with subjects receiving 12 identical stimulus presentations. Furthermore, the effects of a change in rise time were investigated by introducing a test trial in which the rise time conditions were reversed in each of the four groups. In addition to KR, skin uctance responses (SCRs) were measured to examine whether rise time effects are manifested in this frequently-used variable. A preliminary report of this paper was presented at the annual meeting of the ‘Deutsche Gesellschaft fiir Psychophysiologie und ihre Anwendung’, Giessen, in June 1990.

1000 Hz -

1000

Hz -

50 ms Rise Time

Instantaneous

Rise Time

METHOD Subjects Subjects were 120 right-handed female and male students of the University of Mainz aged 19-33 years (mean age = 24.42 years) who reported no history of hearing loss or hearing difficulties. The paid volunteers were randomly assigned to one of the four experimental groups with the restriction that 15 female and 15 male subjects were in each group. Stimuli, apparatus, and procedure The stimulus series consisted of 12 presentatie--, or a 1000 Hz pure sine tone with a constant interstimulus interval of 90 s (onset to onset). Under low intensity conditions, tone intensity was 60 dB; under high intensity conditions, 95 dB (re: 0.0002 dynes/cm2) as measured by a sound level meter (Briiel & Kjaer, Tyoe 2203, A-Scale) at the headrest of the subject’s chair. Tones were presented against a background of 30 dB white noise produced by a Lafayette (15011) generator. The tulle signals were generated by a Eurocom II microcomputer (ELTEC) and amplified by a Technics (SU-VSX) amplifier. They were presented through a speaker (WKD, Type PM1625) in front of the subject at a distance of approx. 190 cm. Rise time was controlled by a computer program developed by Bergmann and Kuras (University of Mainz). Essentially, in this program, the

Fig. I. Onset &raGeristics of the 50 ms rise time and instantaneous rise time stimuli (Note: The small irregularities that can be seen are neither due to the computer program generating the signals nor to the D/A converter of the Eurocom II, but to small errors in plotting the tone signals from a storage oscilloscope).

amplitude of each single wave of a sine tonestarting at zero voltage-that corresponds to a given linear rise and decay function is calculated and transmitted to a D/A converter. Under controlled rise time conditions (CRT), rise time was set at 50 ms; under instantaneous rise time conditions (IRT), at 1 ms (see Fig. 1). Decay times under both conditions were 50 ms. Equivalent duration of the stimuli under both rise time conditions was 117 ms as calculated by a formula provided by Dallas and Olsen (1964). On a final test trial (trial 13) a change in rise time was introduced by reversing the rise time condition in each group (i.e., from CRT to IRT and vice versa). An electrocardiogram (ECGJ was obtained from two Keliige Ag-AgCl electrodes filled with Kellige electrode paste and attached to the manubrium of the sternum and the left lower rib cage. For amplifying the ECG signal and defining the peak of the R-wave an electronic device based on the principle of a turning trigger developed by Shimizu (1978) was used, R-R intervals in ms

44

were measured and registered by a second Eurocorn Ii. Bipolar recording of skin conductance was accomplished using Heliige Ag-AgCI electrodes (surface area = 1 cm’) filled with 0.05 M NaCl electrolyte. The electrodes were placed on the thenar and hypothenar surfaces of the subject’s left hand using Hellige electrode collars. Skin conductance was detected by a constant voltage (0.5 VI skin conductance coupler (cf. Fowles, Christie, Edeiberg, Grings. iykken & Venabies, 1981). Conductance data were frequency-modulated by means of a transducer (C-EDA 5, Juris & Wiilk, 1988) and registered by a counter of the I/O card of the second Eurocom II with a sampling rate of 10 Hz. Subsequcnt!y. the data were converted into PSiemens (PS) and stored on disk. Prior to having the electrodes attached, subjects were requested to wash their hands; subsequently, electrode sites for the measurement of the ECG and SC were prepared by cleaning the skin with ethyl alcohol (70%). The recording took place in a soundproof chamber with subjects seated in a semi-reclining chair. The temperature was between 21 and 23°C. Ail recording and programming equipment was iocatcd outside the chamber. Subjects were informed that after a 5 min rest period :hey would hear a series of tones. They wcrc rcqucstcd to sit quietly during the whole mcasurcmemt period.

Scoring

R-R intervals in ms were converted into HR (bpm), and were initially sampled second-by-sccond for 80 poststimulus seconds according to the formula provided by Velden and Wiiik (1987; cf. also Veldcn & Graham, 1988). Prestimuius HR activity was defined as the average HR for 4 s prior to stimulus onset. These averages were then used to derive poststimulus diffcrcncc sccrcs. Additionally. for each subject and each trial, a rcsponsc parameter analysis was performed by scparatcly determining maximal HR dccclcration and acceleration that occurred over the initiai 4 poststimulus seconds (cf. Turpin & Siddlc, 1983). Non-criterion responses were set at zero.

Artifact-free SCRs occurring from 1.0 to 3.0 s after stimulus onset 4> 0.02 ,&I were scored as stimulus-evoked responses. They were converted to log pets according to the formuia log(X + 1).

RESULTS Data were analyzed using repeated measures ANOVAs, incorporating the Greenhouse-Geisser correction for non-homogeneity o? covariance. Factors were Blocks or Trials, Rise Time, Intensity and for poststimulus HR difference scores, Seconds. The factor Gender was not included as there were no theoretical reasons to expect different physiological rer;ponses to rise time and intensity manipulations in females and males.

HR difference scores for 10 poststimulus s and three trial blocks (average of four trials each) are displayed in Fig. 2. This presentation of HR data is similar to Hatton’s et al. (1970) presentation. An inspection of Fig. 2 indicates a deceleratory R response peaking 2 s after stimulus onset for the 60 dB conditions. For the 95 dB conditions, weak decelerations were observed during the first 2 s that grew stronger over blocks of trials. and that were followed by acceleratory responses peaking primarily at second 4. As the criteria for identifying startlc responding in HR formulated by Graham ( 1979) and Turpin ( 1986) emphasize changes within 4 poststimulus seconds, and as the present data confirm that the most pronounced changes occurred within this time period, statistical analyses of HR difference scores were confined to lircsc 4 s. ANOVA results with the repeated measures factor Blocks are presented in TabEe I. From Table I it is evident that the manipuiation of rise time did not significantly influence poststimulus HR. On the other hand, significant main effects of Intensity, Trial Blocks, and Seconds, various two-factor interactions among them, as well as one triple interaction were obtained. The triple interaction involving Intensity, Blocks and Seconds is displayed in Fig. 3.

45

7-

Block

1

1 (f. Trials

2,--

Block

TABLE I

l(LTriols)

1

Analysis of t-ariance of poststimulus trial block::

HR difference scores for

Factor

df

F

P

Main effects Rise Time (R) Intensity (I) Blocks (B) Seconds (S)

1.112 1.112 2.0, 223.6 1.9. 211.8

1.59 13.87 4.98 32.00

ns < 0.01 < 0.01 E:0.01

Interactions Rxi BxR BxI BxS SxR SXI BxRxl SxRxl BxSxR BxSxI BxSxRxI

1.112 2.0, 223.6 2.0, 223.6 3.8,431.5 1.9,211.8 1.9, 211.8 2.0. 223.6 1.9, 211.8 3.8,43 1.5 3.8.431.5 3.8, 431.5

0.03 1.46 15.22 2.78 0.25 5.14 2.98 0.17 0.74 3.53 0.54

ns ns < 0.01 < 0.05 ns < 0.01 ns ns ns < 0.01 ns

0 -1

-1 E

0” -2 z

-2 -3

-3

0 (L = -4

X 60dB-IAT #i 60dB-CRT

i

-4

c

-5L’

-51’“’ 123L5678910

1 2

X 95 dB-SRT e495 dB- CRT 3 C 5

StXOllClS

‘I-

Block

7 8 9 10

Seconds

2 (4 Trials

2r

1

Block2

1

;i

6

(4 Trials

1

-1

CL

”

-2

.G

-3

x COdB-IRT

a: I -4

x 95 dB- IRT

@60 dB-CRT

I;[

,

,

:‘,Sd,B-,CR,T

, c

ns. nonsignificant.

-5 !??‘S6’99lCl

123s567a910

s

Second

Block 3tLTrials)

x 60dB-

2r

IRT

2

3 4

5

6

Seconds

7

a 9

Block

-3

10

3 (4 Trials)

x95dB-IRT

t

‘--I, , ,

@60dB-CRT I

third and fourth second. In the second trial block, the response under the 95 dB conditions did not appear to change, while under the 60 dB conditions the deceleration was diminished, reaching prestimulus level in the fourth second. Finally, in the third trial block, HR responses for both intensity conditions were nearly identical, demonstrating an initial deceleration peaking at second 2, that was followed by a return to prestimulus

Seconds

1

2

~S~d~-~R~

3 L

5 6

2

, ,

1

Block

Block 2

Block 3 F

1

7 8 910

c

Seconds

Fig. 2. HR difference scores for 10 poststimulus seconds and three trial blocks (IRT = Instantawws Rise Time; CRT = Controlled Rise Time).

Fig. 3 shows that HR responses for the 60 and 95 dB conditions changed across trial blocks. In the first block a pronounced deceleratory response was observed for the 40 dB conditions that did not reach prestimulus level within 4 s, while for the 95 da conditions a weak deceleration was followed by HR accelerations in the

-+

-4

I

12

w

I

I

I

3

4

12

I

,

3 Seconds

95dB

I

I

c

12

I

,

I

3

L

Fig. 3. HR poststimulus difference scores 1s 1 fx?r!ion of intensity, trial blocks and seconds.

4h Trml

TABLE II ,&&;s;s

of r-ariance of p~ststitdits

Trml

2

HR differel1ce SCOWS for

Trml

3

L

b ,

trials 1 to 4 Fuctar

df

F

P

Main effects Rise Time (R) Intensity (1) Trials (T) Seconds(S)

1,112 1.112 ’ 9._a. 3’9 7 _. I .9. 220.2

0.04 31.21 I.0 9.x3

ns < 0.01 ns < 0.01

Interactions Rx1 TxR Txl TxS SxR SXl TxRxl SxRxl TxSxR TxSxi TxSxRxI

1.112 ‘0 ,*_ 3’97. “. 7 9..-379.7 L. 4.x. 537.8 1.9 , __ “0.2

0.96 I .4h 8.10 Oh0 0.49

ns ils < 0.01 ns ns

10.08 0.78

ns

I .9. “0.’

0. I5

ns

4.8. 537.x 4.x. 537.x 3.8. 537.8

1.hH 1.37 2.4x

n\ ns < 0.0s

60dB-IRT

e

95dB--IAT

--$I---

60d%-CRT

__*__

95dB-CRT

23&123&

23C123Cl

Seconds

Fig. 4. MR poststimulus difference intensity.

< 0.01

1.9.220.2 2.9. 3’9.7

-

rise time. trials.

scores as a f~~~ti~fl

and seconds (IRT

=

Rise Time: CRT = C*wrolteei Rise Time).

Seconds X Rise Time x Intensity apparent in Fig. 4, which indicates various more or less unsystematic changes in response morphology across seconds and trials dependent on intensity and rise time.

ns. nonsignificant.

level in second 3, and an acceleration in the fourth second. In order to examine the possibility that rise time efrccts on HR were only presigr.t during the beginning of the habituation procedure, the first four trials were analyzed separately. Results are shown in Table 11. The only significant effect involving rise time is the 4th order interaction between Trials x

pcmmem analyses ANOVA results with the Trial Block factor for maxima! magnitudes of HR deceleration ( DECl 2nd acceleration WR-ACC) occurrmg within the four initial poststimulus seconds are shown in Table Ill. R-DEC. significant main effects for Intensity and Bbcks, as well as a significant Blocks

Resporuc

TABLE l;l

hmr

Main effects Rise Time (R) Intensity ,lJ Blocks(B)

HR-IXC 4

r

1.117 I.112 1.x. IYKY

_.__ - .$ 11.98

..

1.112 1.x. 198.9 1.8, 19x.9 1.8. 198.Y

HR-A CC

.P

4

F

P

1.112 1.11’1

0.07 0.96

lw

< 0.01

S..W

< 0.0

‘11 1. ,,

2.48

RS

0.w 1.36 X.13 2.32

ns ns < 0.01

FIS

I

-l,L, I .I,.-

la-3

Interactions Rxi BxR Bxl BxRxl ns.

nonsignificant.

of

lnStafltanCOMS

ItS

1.112 1.Y. 36.2 1.9.21h.2 1.9, 21h.2

0.87 03 IO.58 l.6?

BE

ns < 0.01 flS

47

X Intensity interaction were obtained, while for i-IR-ACC only the Blocks x Intensity mteraction reached significance. In both variables no significant rise time effects occurred. The decelerative component was stronger in blocks 1 and 2 for the 60 dB conditions than for the 95 dB conditions, e in the third trial block the differences between intensities had disappeared. Maximal HRDEC did not change for the 95 dB groups over blocks, while it became weaker for the 60 dB groups. For HR-ACC, intensity groups differed only in the first trial block where the accelerative component for the 95 dB conditions was stronger than that for the 60 dB conditions. The 95 dB conditions showed a decrease in HR-ACC from the first to the second block and the 60 dB conditions an increase. In neither condition did HR-ACC change from the second to the third block of trials. The results for the analyses of the first four trials are displayed in Table IV. For bDth variables, significant intensity effects as well as significant Trials x Intensity interactions were obtained. Again no significant effects of rise time were found. For both HR-DEC and HR-ACC, the significant interactions can be attributed to differences between intensities during the first two trials, where the 60 dB conditions displayed stronger decelerations and the 95 dB conditions stronger accelerations. These differences disappeared in trials 3 and 4.

-6OdB-IRT -0-95 dB-

IRT

---x----4--

CRT CRT

60 de95 dB-

06 -‘: In 9 $ &

O.L-

0.21’

1

’

2

’

3

b

4

’

’

5

6

’

7 Trials

’

6

’

’

9

10

Skin conductance SCR magnitudes across 12 trials for the four experimental groups are shown in Fig. 5. A three-factor ANOVA revealed significant effects for Intensity, F(1, 112) = 12.19, P < 0.01, and Trials, F(6.37, 713.52) = 35.25, P < 0.01, whereas the Trials x Intensity interaction was only marginally significant, F(6.37, 713.52) = 1.85, P = 0.08. Signifilrant effects for rise time were not obtained. As can be seen in Fig. 5, the marginally Significant interaction was due to a strongei Jecline in SCR-magnitudes for the 60 dB conditions, especially during initial trials. A separate

Analyses of variance of magnitude of HR deceleration (HR-DEC) and acceleration (HR-ACC) for trials 1 to 4 HR-ACC

HR-DEC df

F

P

df

F

P

Main effects Rise Time (RI Intensity (I) Trials (T)

1,112 1,112 2.9, 331.5

0.3 18.95 0.66

ns < 0.01 ns

1,112 1,112 2.8, 317.4

0.3 15.57 0.96

ns < 0.01 ns

Interactions Rx1 TxR TxI TxRxI

1,112 2.9, 331.5 2.9,33 1.5 2.9, 331.5

3.06 2.16 3.95 0.87

ns ns < 0.01 ns

1,112 2.8, 317.4 2.8,317.4 2.8,317.4

0.42 0.86 7.72 0.23

ns ns < 0.01 ns

ns, nonsignificant.

’

12

Fig. 5. SCR magnitude as a function of intensity, rise time, and trials (IRT = Instantaneous Rise Time; CRT = Controlled Rise Time).

TABLE IV

Factor

’

11

3

T 0”

0

g 0 cc Ib

-1

r

__--x ,a*x_--

-2

_____-r----K+ --a--__ --

-3 -4

t

I 3

I 2

1

60 da-H12 95dB-HlZ 60dB-Cl 95d0-Cl I b

Seconds

Fig. h. HR

poststimulus

to the last stimulus

difference scores between responses

of the habituation change stimulus (C

series (H

12) and the

1).

analysis of the first four trials confirmed this. A significant Trials x Intensity interaction was now obtained, F(2.5, 282.3) = 4.99, P < 0.05, indicating different &.c!~.nesbetween the intensity groups during these initial stimulus presentations. Effects of rise time change

The effects of a change in rise time (i.e., from IRT to CRT and vice versa) were assessed by comparing responses to the 12th stimulus of the habituation series with those to the change stimulus by means of four- or three-factor APJc’OVAs. For poststimulus HR, difference scores beese trials showed significant effects for Habituation vs. Change Stimuius, Ffl, 112) = 9.28, P < 0.01, and for Seconds, F(1.88, 210.08) = 9.71, P < 0.01, as well as a triple interaction between Stimulus, Seconds and Intensity, F(1.92, 214.88) = 3.49, P < 0.05 The triple interaction is shown in Fig. 6. It suggests that the changs stimulus induced a stronger and longerlasting deceleratory response t!:an the last habituation stimulus, and that the change stimulus presented with 95 dB induced a return to prestimulus level in second 4 while the change stimulus presented with 60 dB did not. For HR-ACC no significant effects were found, for HR-DEC a significant effect was obtained only for Stimulus, Ffl, 112) = 10.16, I’ K 0.01, in-

dicating a stronger deceleratory response to the change stimulus. The ANGVA for SCRs revealed significant effects for Intensity, F(1, 112) = 10.15, P < 0.01, and Stimulus, F(1, 112) = 4.00, F < 0.05. No other effects were found to be significant. SCR magnitudes for the 95 dB conditions were greater than for the 60 dB conditions, and SCR magnitude for the rise-time-change stimulus was greater than for the last habituation stimulus.

DISCUSSION The results of the present study revealed a number of significant effects of stimulus intensity on cardiac and electrodermal responding. Stimulus intensity had clear effects on response morpholoev, indicating that, with increasing stimulus intensity, deceleration was reduced and acceleration became prominent. This finding is in general accordance with previous reports (cf. Graham, 1979, for a summary). With regard to the habituation of cardiac responses, especially to the high intensity stimuli, the present data do not support Graham’s (1979) conclusions. According to her view (Graham, 1979, p. 143). repeated presentations of high intensity stimuli should result in a decrease in decelerative phases and an increase in accelerative ones. The findings for poststimulus HR difference scores, however, revealed that under the 95 dB conditions, the initial deceleration grew stronger over blocks of trials, while there were no substantial changes in the acceleratory response seen during the third and fourth second. Results of the response parameter analyses also do not support Graham’s position, although they were not completely congruent with those for poststimulus HR. While no changes over blocks were observed for HR-DEC under the 95 dB conditions, HR-ACC showed an initial decrease and then remained constant. It should be emphasized that these findings replicate previous ones: they correspond largely to the results of the response parameter analyses reported by Turpin and Siddle (1983) for their high intensity conditions (i.e., 90 dB and 105 dB). For the 60 dB conditions the repetition effects indicated a de-

crease of the deceleratory response over blocks for poststimulus HR and HR-DEC. Finally, significant stimulus repetition effects were obtained for skin conductance, and it could furthermore be shown that, especially during early trials, the higher stimulus intensity led to a slower decline CR magnitude. These latter findings are in general accord with previous ones (Turpin & Siddie, 1979) and will not be discussed further. Compared to the effects of stimulus intensity, the effects of stimulus rise time on cardiac and electrodermal responding were clearly negative. Actually, in all analyses based on blocks neither significant main effects for rise time nor significant interactions between rise time and any of the remaining factors were obtained. This was true for poststimulus HR difference scores, for cardiac response parameters, as well as for skin conductance. Statistical analyses based on the first four trials also did not reveal any substantial impact of the manipulation of rise time, with the exception of a four-factor interaction involving rise time, intensity, trials, and seconds for HR difference scores. This complex interaction indicated that differences among the four experimental groups across trials and seconds were rather inconsistent. As Fig. 4 indicates, in the first trial, instantaneous rise time led to an acceleratory response at 95 dB with an onset latency less than 2 s and a peak latency of 3 s. While this corresponds to the criteria for HR startle as formulated by Graham (1979) and Turpin (1986), the 95 dB stimulus with controlled rise time did not elicit a deceleratory response here, as expected from the study by Hatton et al. (19701, but an immdiate __-.-..__.___

arreleratinn ____._._..V..

nnlv . . . ..J cliahtlv “..0..__,

weaker

than

that for the stimulus -with instantaneous rise time. The inconsistency of the rise time effect becomes espeMly evident in the second trial of the 95 dB conditions. In that trial controlled rise time resulted in an immediate HR acceleration peaking at the fourth second, thus fulfilling the criteria for startle responding, while instantaneous rise time elicited an apparent deceleratory response. These findings are obviously at variance with the role attributed to instantaneous and controlled rise time. The fmdings for the third and fourth trial add to the inconsistency of rise time effects,

as in both trials completely different responses were observed that were also not comparable to the responses seen in the first two trials. In sum, the four-factor interaction including rise time cannot easily be reconciled with the assumption that controlled and instantaneous rise times elicit qualitatively different cardiac responses. Furthermore, this interaction can hardly be considered as supporting the position that a combination of instantaneous rise time and high intensity evokes startle responding in HR (Graham, 1979), nor can it be taken as supporting the assumption that an instantaneous rise time-independent of stimulus intensity-is sufficient to evoke HR responses that can be interpreted as indicating startle (cf. Simons, 1989). The failure to obtain the predicted rise time effects makes it impossible to examine furttcr criteria for a startle-response in HR, as for example the proposed rapid habituation of that response (cf. Graham, 1979) or the proposed dependence of habituation rate upon intensity (cf. Turpin, 1986). The presentation of a change stimulus produced by reversing the rise time condition in each group, had significant effects on poststimulus HR, HR-DEC, and SCR magnitude. These effects were independent of the direction of the change in rise time (i.e, from CRT to IRT and vice versa!, confirming the negative results for the rise time manipulation reported above. For HR-DEC and skin conductance, significant maili effects were found indicating a greater maximal deceleration and a greater SCR magnitude to the change stimulus. For poststimulus HR, the rise time change induceu a greater and longer-lasting HR deceleration that returned to prestimulus level in the fourth second under the 95 dB conditions, while under the 60 dB conditions it did not. The change effects found for SCRs and HR are in gcncral agreement with a number of previous studies in which mainly the effects of changes in tone frequency were investigated (e.g., Graham, 1979, p. 144; Siddle, O’Gorman & Wood, 1979; Vessel & Zimmer, 1989), demonstrating that both variables are sensitive to changes in various physical stimulus attributes. What are the main conclusions to be drawn from the present data? First, they suggest that

50

the role attributed to rise time as eliciting oualitatively different cardiac response patterns has been overestimated. Instead, the HR findings revealed that instantaneous rise time did not lead to consistently greater accelerative responses, and that (in contrast to Hatton et al., 1970; EXP. 2) a cqmbination of instantaneous rise time and high stimulus intensity did not elicit a dominant accelerative response pattern. It cannot be easily decided why Hatton et al. and the present study produced such discrepant findings. Likely causes of the discrepancies ars the different designs used (e.g., within vs. between subjects designs) and the different controlled rise time conditions investigated (e.g., 300 ms vs. 50 ms). Any definite answer, of course, would reqluire further experiments. At the moment, we can only point out that the conditions realized in the present study with regard to the experimental design and rise time duration are in better agreement with most of the studies in this area than were the conditions realized by Hatton et al. Therefore, our conclusions concerning the role of rise time should be relevant for the discussion of whether a startle pattern in HR can be reliably demonstrated. In this context it must be emphasized that the failure to observe rise time effects on HR (and SCRs as well) does not imply that rise time may not influence other physiological variables. For example, Blumenthal (1988) and Blumenthal and Berg (1986) have demonstrated that rather subtle manipulations of the rise ti es of broadband noise and of tones influence periorbital EMG responses, with increasing rire times rc;zrl:ing in a lower response amplitude and/or a lower response probability. Second, the HR data of the present study are interesting in the context of recent accounts of differentiating defense and orienting responses (cf. Tiirpin, 1986, for a summary), especially as clear startle respoiise patterns in HR did not occur. Although the effects of high intensity on response morphology were in general accord with Graham’s (1979) HR criteria for a defense response, the stimulus repetition effects were not. It could neither be shown that decelerative components decrease with stimulus repetition nor that accelerative components increase. in general, our findings suggested an in-

crease in poststimulus decelerations (see Figs. 2 and 3) and-as indicated by response parameter analyses-a decrease in accelerative components. These results add to the confusion in this area of research and make it difficult to interpret this HR acceleration straightforwardly as an indicator of the defense response (cf. also Barry & Maltzman, 1985, and Turpin, 1986, for similar conclusions), especially as accelerative responses also occurred under the 60 dB conditions (see Figures 2 and 3) *. Under low intensity conditions the dominant HR response was a deceleration that showed a marked and relatively regular decline across trial blocks. Actually, these are the first HR findings that are in accordance with the operational criteria as formulated for autonomic indicators of the orienting response (cf. Barry, 1987; Vossel & Zimmer, 1989). An unambiguous interpretation of HR deceleration as an OR index (cf. Simons et al., 1987; Turpin, 19861, however. seems to be premature as these data are in part at variance with previous ones (e.g., Simons et al., 1987; Turpin & Siddle, 1983: Vessel & Zimmer, 1989)-especially with regard to changes in HR deceleration across repeated stimulus presentations-and as the discrepancies cannot be attributed to rise time. Therefore, further research is clearly needed to determine the necessary conditions under which HR deceleration fulfills the criteria for an OR in order to decide between (or integrate) different theoretical interpretations of HR deceleration, e.g., classical OR theory and Barry’s (1987) theory of preliminav processes in OR elicitation. In sum, the present findings point to the difficulty in identifying different unconditioned responses on the basis of HR changes. As summarized by Turpin (1986), this conclusion holds not only for HR, but pertains to other physiological

* As we initiallysampled HR For 80 postslimulus seconds, we were able to examine whether a long-latency HR responseidentified by Turpin (198h) as an indicator of the defense response-occurred. Repeated measures ANOVAs of the firs: four trials involving the factors Trials. Rise Time, Intensity, and Blocks of HR over 5 s revealed no occurrence of significant accelerative responseswith peak latencies in the rarge of 30 s for either condition.

variables (especially vasomotor changes) as well. We therefore tend to agree with Turpin that the use of single autonomic measures for the differentiation of orienting, defenJz and startle requires reappraisai. Further research should especially concentrate on the functional significance ic interactions of these proposed response complexes and, consequently, must broaden the psychophysiological data-base in order to integrate various physiological and behavioral measures. In this context, further research might profit by additionally considering the measurement of facial muscle activity as demonstrated by Ekman, Friesen and Simons 11985) in their investigation of the uniformity of the facial startle expression in man.

REFERENCES Barry. R.J. (1982) Novelty and significance effects in the fractionation of phasic OR measures: A synthesis with traditional OR theory. Psychophysiology, 19: 28-35. Barry. R.J. (1987) Preliminary processes in orienting response elicitation. In P.K. Ackles, J.R. Jennings and M.G.H. Coles (Eds.1. Adrances in Psychophysiology (Vol. 2). Greenwich. CT: JAI Press, pp. 131-195. Barv, R.J. and Maltzman. 1. (1585) Heart rate deceleration is not an orienting reflex: heart rate acceleration is not a defensive reflex. Parlorian J. Biol. Sci.. 20: 15-28. Berg. W.K.. Jackson, J.C. and Graham, F.K. (1975) Tone intensity and rise-decay time effects on cardiac responses during sleep. Psychophysiology. 12: 254-261. Blumenthal, T.D. (1988) The startle response to acoustic stimuli near startle threshold: Effects of stimulus rise and fall time. duration, and intensity. Psychophysiology, 25: 607-611.

Blumenthal, T.D. and Berg, W.K. (1986) Stimulus rise time, intensity, and bandwidth effects on acoustic startle amplitude and probability. Psychophysiology, 23: 635-641. Dallos. P.J. and Olsen, W.O. (1964) Integration of energy at threshold with gradual rise-fall tone pips. J. Acoust. Sot. .Am., 36: 743-751. Ekman, P., Friesen, W.V. and Simons, R.C. (1985) Is the startle reaction an emotion? JOWWU/of Personality und Serial PsycholoD, 49:1416-1426. Fowles, D.C.. Christie, M.J.. Edelberg. R.. Grings. W. W, Lykken, D.T. and Venables. P.H. (1981) Publication recommendations for electrodermal measurements. Psychophysiology, 18: 232-239. Graham, F.K. (1973) Habituation and dishabituation of responses innervated by ih: autonomic nervous system. In H.V.S. Peeke and M.J. Helz (Eds.). ffubituarion. Vol. I:

Behuciorul studies. New York: Academic Press, pp. 163-

218. Graham, F.K. (1979) Distinguishing among orienting, defense, and startle reflexes. In H.D. Kimmel, E.H. van Olst and J.F. Orlebeke (Eds.), 77zeorienting rej7ex in hxnuns. Hillsdale, NJ: Lawrence Erlbaum, pp. 137-167. Hatton, H.M., Berg, WK. and Graham, F.K. (1970) Effects of acoustic rise time on heart rate response. Psychonom. Sci., 19: 101-103. Juris. M. and W(ilk, C. (1988) A low-cost measuring system for recording of the electrodermal response without splitting into tonic and phasic parts. Paper presented at the 17th Meeting of the German Psychophysiology Society, Munich, IWSLu’erman;.

Levy, M.N. and Martin, P.J. (1979) Neural control of the heart. In L.M. Beme, N. Sperelakis and S.R. Geiger (Eds.), Handbook of physiology: The curdiorascular system (Section 2, Vol. 1). Bethesda, MD: American Physiological Society, pp. 581-620. Shim& H. (1978) Reliable and precise identification of Rwaves in the EKG with a simple peak detector. Psychophysiology, 15: 499-511. Siddle, D.A.T., O’Gorman, J.G. arrd Wood, L. (1979) Effects of electrodermal lability and stimulus significance on electrodermal response amplitude to stimulus change. Psychophysiology, 16: 520-527.

Simons, R.F. (1989) “A rose by any other name”: A comment oc Vossel and Zimmer. J. Psychophysiology, 3: 125-127. Simons, R.F.. Rockstroh, B., Elbert, T., Fiorito, E., Lutzenberger, W. and Birbaumer, N. (1987) Evocation and habituation of autonomic and event-related potential responses in a nonsignal environment. J. Psychophysiol., 1: 45-59. Turpin. G. (1983) Unconditioned reflexes and the au!onomic nervous system. In D.A.T. Siddle (Ed.), Orienting and habituation: Perspectives in human resenrch. Chichester: Wiley, pp. I-70. Turpin, G. (1586) Effects of stimuius intensity on autonomic responding: The problem of differentiating orienting and defense reflexes. Psychophysiofogy, 23: 1- 14. Turpin, G. and Siddle, D.A.T. (1978) Cardiac and forearm plethysmographic resuonses to high intensity auditory stimulation,. Siol. Pnnchol., 6: 267-281. Turpin, G. and Siddle, D.A.T. (1979) Effects of stimulus intensity on electrodermal activity. Psychophysio/ogv, 16: 582-591.

Turpin, G. and Siddle, D.A.T. (1983) Effects of stimulus intensity on cardiovascular activity. Psychophysiofogy, 20: 61 l-624.

Velden, M. and Graham, F.K. (1988) Depicting heart i:te over real time: two procedures that are mathematically identical. J. Psychophysiol., 2: 291-292. V&en, M. and Wiilk, C. (1987) Depicting cardiac activity over real time: A proposal for standardization. J. Pgjchophysiol., I: 173-175. Vossel, G. and Zimmer, H. (1989) Heart rate deceleration as an index of the orienting response? J. P~chophysio~.. 3: 111-124.