Australian Journal of Psychology Vol. 27, No. I. 1975. pp. 61-83
AN INTEGRATIVE PROGRESS REPORT ON
INFORMATIONAL CONTROL IN HUMANS: SOME
LABORATORY FINDINGS AND METHODOLOGICAL
CLAIMS
JOHN J. FUREDY' University of Toronto
When noxious unavoidable events (e.g., shocks) are signalled (e.g., by a 5-sec. tone), it is widely believed that there is informational control (1C) in the sense that signalling reduces perceived event noxiousness; the IC mechanism arises through signal-elicited preparatory anticipatory responses which reduce event noxiousness; there is a preference for signalled over unsignalled events (preference for signalling). Since these beliefs, though related, are different, the relevant evidence from the Toronto laboratory is presented separately for each belief. That evidence, comprising studies published from 1970-1974 and some additional unpublished experiments, indicates: (a) no support for an IC mechanism in the elcctrodermal, plethysmographic, and cardiac response systems; (b) no support for IC itself with shocks and loud noises as noxious events; (c) no general prefer-ence-for-signalling (PFS) phenomenon (assertion based on total of over 570 5s); but (d) a recent but consistently emerging specific PFS which appears based on 5s' (false) beliefs in the benefits of signalling. There follow some methodological claims which partially integrate the Toronto studies with apparently conflicting evidence from other laboratories. Complete integration involving a statement of the conditions under which signalling has beneficial, detrimental and no effects is not yet possible because of the empirical complexity of the problem, a complexity which requires an investigative rather than a demonstrational experimental approach. To illustrate the apparent strength of the latter approach, some examples are giver, of the current strong but unjustifiable Zeitgeist in favour of such beliefs as the generality of the PFS phenomenon in humans.
' Most of the material of this articlestudents at the University of Toronto: G.
(written while on sabbatical leave at theB. Biederman, C. D. Creelman, A. N.
University of Sydney) was presented at aDoob, M. Katie, F. Klajner, S. Ginsberg,
working conference supported by Canadaand C. X. Poulos. The empirical work
Council and the Department of Psy-reported in this paper was supported dur-
chology, University of Toronto held ating the period of 1968-1973 by grants
the University of Toronto, September,from Canada Council and from the
1973 with P. Badia, D. E. Berlyne, G. B.National Research Council of Canada.
Biederman, H. Murray, C. C. Perkins, Jr,For critical advice (of which only some
M. E. P. Seligman, M. D. Suboski as thewas accepted) on an earlier draft of the
other participants to whom I am indebtedpresent article I am especially indebted
lor comments on that conference versionto R. A. Champion with whom I have
which was prepared with the aid of C. X.profitably been agreeing to disagree for
Poulos. The evidence and ideas presentedmore than a dozen years. For stylistic
both in the conference version and in theimprovements in the writeup of the present
present paper were developed in collabor-version, I owe much to the advice of
ation with the following colleagues andtwo anonymous referees and of C. P.
------Furedy. Finally, the investigative approach
Manuscript received 27 February 1974. which I have tried to take and stress
Address for reprint requests:throughout the paper was indirectly, but
John J. Furedy,powerfully, influenced by the realist-em-
Department of Psychology,piricist philosophy of science taught at the
University of Toronto,University of Sydney by the late John
Toronto 181, Ontario, CanadaAnderson and his students.
61
62
John J. Furedy
What happens when the human organism receives a signal about a noxious event which, however, is physically unmodifiable? The concern of this paper will be to examine arrangements in which shocks or loud noises are signalled by weak tones or lights of about -5 to 10 sec. in duration, and to compare these arrangements with those where the same events are not so signalled. Related but different questions which involve comparing delayed with immediate shocks (e.g., Maltzman & Wolff, 1970), or even more closely related questions such as those involving regularly presented, but not explicitly signalled, shocks (e.g., Lovibond, 1968) will not be taken up here. These questions are excluded not because of irrelevance, but rather because they complicate a picture which is already sufficiently complex. Moreover, since the animal literature on the question of signalling has specific and complex problems of its own (cf., e.g., Biederman & Furedy, 1973), this paper will be confined to data gathered only from humans.
These restrictions do not imply that the problem of concern is viewed as an esoteric one devoid of any practical significance, or as a purely empirical question which is unrelated to theories of behaviour. On the contrary, the psychological effects of signalling on physically unmodifiable noxious events are quite important in various areas of medicine and dentistry. However, that importance lies not so much in how we can "come up" with signalling methods that will reduce perceived noxiousness or pain (which is to put the question in a way that at least partly begs it), but in determining the conditions under which signal-
ling intensifies, reduces, or has no effect on pain. Again, from a theoretical point of view, signalling effects are not only relevant to, but are predictable by, various more general theories of behaviour. Most current theories (e.g., Berlyne, 1960; Lykken, 1962, Perkins, 1968; Selig-man, 1968) predict that signalling has beneficial or alleviating effects on pain. However, the situation becomes complex once it is recognised that it is easy to deduce conclusions of the opposite nature from other theories which take the motivating effects of conditioned fear into account (Hull, 1943; cf. Furedy & Doob, 1971a, p. 263). This complexity, indeed, can also be recognised by listening to subjects' reasons for preferring the signalled and unsignalled situations, respectively. These reasons are actually formulations of the two basically opposing theoretical viewpoints in man-in-the-street terminology, i.e., "the shock hurts less when signalled because I can prepare myself for it", vs. "the shock hurts more when signalled because I tense up". Accordingly, as in the case of applications, so in the case of basic theoretical research: what should be sought from experiments are not demonstrations of the beneficial effects of signalling but specification of the conditions under which the effects of signalling are beneficial, null, and detrimental.
In support of preferring an investigative over a demonstrational approach to the problem, this paper will first present evidence on the effects of signalling that has emerged from the Toronto laboratory from 1969-1973. Much of this evidence contradicts the current view that these effects are beneficial and that signalling is preferred. Where this
An Integrative Progress Report on Informational Control in Humans 63
evidence has been published in detail elsewhere, the presentation will be very brief, while a small proportion of data which are not currently available for critical scrutiny in published form will be presented in more detail. The second purpose of this paper is to put forward certain methodological claims which have arisen in the course of attempting to reconcile the Toronto results with those reported in the literature from other laboratories. Finally, based on these considerations and on the available theories, an integration will be suggested which, however, will be of a limited nature for reasons to be made clear later.
EVIDENCE FROM THE TORONTO LABORATORY
Central to most current beliefs and theories concerned with the effects of signalling is the notion that has been called the phenomenon of "informational cognitive control", or, more precisely, "informational control" (IC).2 The IC phenomenon may be defined as occurring whenever merely signalling a noxious and physically unmodifiable event reduces the perceived noxiousness of that event (Furedy & Doob, 1972). The first set of evidence presented is that concerned with a mechanism whereby IC might operate: the signal-elicited preparatory (adaptive) response as conceived by Perkins (1968). The evidence gathered from the Toronto laboratory concerning the IC phenomenon itself will then be considered with the understanding
- As noted most forcefully by D. E. Berlyne during the conference presentation of this material, the use of the term "cognitive" is inadvisable in this context, because it presupposes a particular sort (i.e.. cognitive) of mechanism for the informational control phenomenon.
that, as detailed elsewhere (e.g., Furedy & Doob, 1971a, p. 258), the issues of mechanism and phenomenon, though related, are readily distinguishable. Finally, the evidence relevant to the preference-for-signal-ling (PFS) phenomenon will be considered. The PFS phenomenon is exemplified by subjects preferring signalled over unsignalled inescapable, unavoidable shocks of the same intensity. There are obvious relationships between the IC mechanism, IC phenomenon, and PFS phenomenon, but, as detailed elsewhere (Furedy & Doob. 1971a, pp. 258-259), these notions cannot be run together.
The IC Mechanism
When Perkins (1968) published a paper which still remains as the most explicit and testable version of how an IC mechanism might operate, he used the galvanic skin response (GSR) as a specific example of adaptive preparation (Perkins, 1968, p. 163). Consistent with Perkins' notions, later psychophysiological reviews concerned with the GSR have referred to its role in serving "the adaptive notion of allowing the organism to prevent signalled injury to the skin" (Dengerink & Taylor, 1971, p. 358). However, this role requires a negative relationship to hold between conditioned GSR and UCS aversiveness ratings (for derivation of this "regression" hypothesis, cf., e.g., Furedy, 1970, p. 301). a relationship which failed to emerge in a number of studies (Furedy, 1970, 1973; Furedy & Doob, 1971a: Furedy, Katie, Klajner & Poulos, 1973). Moreover, this search for an electrodermal form of the IC mechanism has not been just a matter of simple replications of the same study. For example, because
64
John J. Furedy
UCS intensity was varied within 5s in Furedy (1970), a feature which could have obscured the operation of the electrodermal IC mechanism, Furedy & Doob (1971a) kept UCS intensity constant within 5s. Again, because of the possibility (to be discussed in the methodological claims section in more detail) of an "attentional" confounding as raised by Obrist, Webb and Sutterer (1969) in studies using the rated-aversive-ness methodology, Furedy et al. (1973) manipulated these "attentional" factors, but still failed to find evidence for an electrodermal IC preparatory-response mechanism under any of the conditions. Finally, because of suggestions (e.g., Denger-ink & Taylor, 1971, p. 358) that the clearest electrodermal preparatory response would be discovered only if GSRs occurring before the UCS were separated into those occurring closer to CS onset and those occurring later in the interval (closer to UCS onset), Furedy (1973) extended the CS-UCS interval from 5 to 8 sec. to facilitate separate observation of these later-occurring GSRs. However, as before, no evidence for any electrodermal IC preparatory-response mechanism was found.
Another potential autonomic preparatory response is the vasomotor plethysmographic digital volume-pulse change (VPC), but in studies where the CS-UCS interval was long enough (8 sec.) to permit adequate observation of this anticipatory vasomotor CR, no evidence for any IC preparatory-response mechanism was obtained (Furedy & Doob, 1971b: Furedy, 1973). Both the electro-derma! and plethysmographic response systems, however, are relatively peripheral in nature, and, if one were to search for the most
likely system in which to find an instance of the IC mechanism, the more pervasive and organismically important cardiac response system could be a more appropriate place to look. More recently Ginsberg and Furedy looked for evidence for such a cardiac IC preparatory-response mechanism in the data from 30 5s whose heart rates were continuously recorded during a signalled loud-noise experiment, the conditions for which have been reported elsewhere (Furedy, Fainstat, Kulin, Lasko & Nichols, 1972). Very briefly, the study was a conditioning experiment with a delayed CS-UCS interval of 8 sec, -3-sec. loud white noise (varied from 80-120 dB within 5s) as the UCS, with both signalled (CS-UCS) and unsignalled (UCS alone) trials being delivered to all 5s. In such an extended CS-UCS interval conditioning arrangement, a heart-rate deceleration is typically observed prior to UCS onset and, in the case of a noxious UCS, it has been suggested that this deceleratory component of the heart-rate response serves to attenuate the aversiveness of the UCS (Obrist, Webb & Sutterer, 1969). Preparatory-response notions would seem to require that in this cardiac IC preparatory-response, deceleration magnitude prior to UCS onset be negatively related to the rated intensity of the immediately following UCS. For defining this deceleration on a given trial, the score was zero if the heart was not decelerating immediately before UCS onset (UCSO); if there was deceleration during this period, then we worked back from UCSO beat by beat to the point where the deceleration began, i.e. response onset (RO); deceleration magnitude was then expressed as a percentage of the
An Integrative Progress Report on Informational Control in Humans 65
algebraic difference in rates measured at RO and UCSO divided by the rate at RO. However, the predicted negative relationship between magnitude of cardiac deceleration immediately before the UCS onset and rated intensity failed to emerge, since the number of obtained negative regression coefficients (15) equalled the number of positive coefficients, and the mean of the coefficients did not differ significantly from zero, f = -08.
In summary, then, the Toronto laboratory has examined a number of potential preparatory responses ranging from peripheral responses like the GSR to a more central response like cardiac deceleration. Also the search for evidence that would confirm the operation of an IC mechanism has ranged over a number of possibly relevant variables (e.g. variable vs. constant UCS intensity within 5s), although, quite obviously, the range of variables has by no means been exhausted. Nevertheless, it bears emphasis that in none of these studies was there any support for the notion of an IC preparatory-response mechanism.3 On the other hand, the lack of support for this preparatory-response type of IC mechanism is not inconsistent with the existence of the TC phenomenon.
The IC Phenomenon
The evidence regarding IC that emerged between 1968 to 1971 has been summarised previously (Furedy & Doob, 1972) to indicate that a variety of conditions which employed shock as the noxious unavoidable event presented to over 150 5s failed to reveal any significant tendency for signalled shocks to be rated as less aversive than unsignalled shocks.
Indeed, in one reported experiment (Furedy & Doob, 1972, Exp. Ill), there was a slight but significant tendency in the opposite direction, with signalled shock being rated as more unpleasant than unsignalled shock. Also, the same tendency which did not reach significance in the 16 5s of Experiment He as reported by Furedy & Doob (1972) did reach the -01 level of significance when all 32 5s were run (cf. Furedy & Doob, 1971a, Exp. Ill), with signalled shocks being rated as more intense than unsignalled shocks. In addition, in a later series of experiments (Furedy et al., 1973), data was obtained from 120 5s to indicate that, again, signalling did not reduce the rated intensity of shocks. However, while based on many
3 As noted by R. A. Champion, this is not to say that the possibilities for verifying the presence of an IC preparatory-response mechanism even in the electrodermal GSR system have been exhausted. One possibility which we have not investigated is that with shock durations which exceed the latency of the fully-developed GSR. there, may be a negative correlation between the GSR elicited by shock onset and the rated aversiveness of the shock, and, more particularly, the rated aversiveness of that part of the shock which occurs after the onset itself. Another possible source of verification is the electrodermal resistance level at the point of shock onset in a signal-shock situation, where these levels should be positively correlated with the rated aversiveness of the shock if lower skin resistance does, indeed, lessen the perceived noxiousness of the shock. We are currently starting to investigate these possibilities, but it bears emphasis that even if the results do agree completely with the preparatory-response mechanism predictions, it is nevertheless the case that the body of data reviewed in the text of this paper are uniformly disconfirmatory, and. to that extent, the empirical status of the preparatory-response notion in the autonomic response systems examined remains at present in serious doubt.
66
John J. Furedy
Ss, these studies did not exhaust either all the potentially relevant independent variables, nor even all the important dependent variables. To take an example of the former, although the previous Toronto studies had explored signal-shock intervals of 5 as well as 8 sec, to workers like Suboski and his associates, both those intervals would have seemed to be far too long for producing an IC effect, on the grounds that it was only at intervals approximating -5 sec. that conditioning (of the eyelid preparatory response) is optimal (Suboski, Brace, Jarrold, Teller & Dieter, 1972). However, in two more recently reported studies (Furedy & Klajner, 1972a: Furedy & Ginsberg, 1973) based on two sets of 48 5s, the signal-shock interval was -5 sec, and no difference was obtained between the rated intensity of signalled and unsignalled shocks. Nor, to take an example of a different dependent variable, did autonomic indices of aversiveness yield a difference between signalled and unsignalled shocks in these studies, although various forms of these indices (Furedy & Klajner. 1972a; Furedy, 1972; Furedy & Ginsberg. 1973) were validated in the sense of being demonstrably sensitive to shock intensity differences.
On the other hand, the experiments with Klajner and Ginsberg were quite complex in design, since they included "secondary" as well as "primary" signals. However, in two recent more simple studies to be reported in more detail in the next section, shocks signalled at from -5 to 1-0 sec. could be compared to shocks signalled at longer intervals at 35 and 5 sec. Despite the interesting arguments and the evidence
of Suboski et al. (1972), shortening the signal-shock interval in our studies did not produce a reliable IC phenomenon.
In the continuing search for the IC phenomenon, it was suggested that IC might be more likely to occur with loud noise rather than shock at the noxious event. The reason for this supposition was the known fact that the middle-ear reflex attenuates noise intensity (Fletcher, 1962), so that a conditioned form of this reflex could act as a signal-elicited IC preparatory-response mechanism. The study which reports relevant data from the largest sample of (56) 5s is that by Furedy, Fainstat, Kulin, Lasko and Nichols (1972), where, despite that fact that there was significant preference for signalled over unsignalled noise, the former was not rated as less intense than the latter. Another more recent study, which has been reported only in conference form (Furedy & Klajner, 1972c), involved presenting both shocks (1-5 to 2-5 mA) and noises (80 to 120 dB) to the same 5s. This experiment will be detailed in the next section, but it is relevant here to note the middle panels of Fig. 2 which show the mean rated unpleasantness over 40 5s of the various noxious inescapable events. The notion that IC will emerge more clearly with noise than with shock implies an interaction between type of event (shock vs. noise) and signalling (signalled vs. unsignalled), with signalling reducing rated unpleasantness more in the case of the noises than the shocks. However, as suggested by inspection of the middle panel of Fig. 2 and as confirmed by statistics, no such interaction occurred.